ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS

NATALIE STEWART

Abstract. We affirm and generalize a conjecture of Blumberg and Hill: unital weak N∞ -operads are closed
under ∞-categorical Boardman-Vogt tensor products and the resulting tensor products correspond with joins
of weak indexing systems; in particular, we acquire a natural G-symmetric monoidal equivalence
arXiv:2504.02143v1 [math.AT] 2 Apr 2025

CAlg⊗ CAlg⊗ C ≃ CAlg⊗ C.

I J I∨J
⊗
is ⊗ -idempotent and O ⊗ is local for the corresponding smashing
BV
We accomplish this by showing that NI∞
localization if and only if O-monoid G-spaces satisfy I-indexed Wirthmüller isomorphisms.
Ultimately, we accomplish this by advancing the equivariant higher algebra of cartesian and cocartesian
I-symmetric monoidal ∞-categories. Additionally, we acquire a number of structural results concerning
BV
G-operads, including a canonical lift of ⊗ to a presentably symmetric monoidal structure and a general
disintegration & assembly procedure for computing tensor products of non-reduced unital G-operads. All
such results are proved in the generality of atomic orbital ∞-categories.
We also achieve the expected corollaries for (iterated) Real topological Hochschild and cyclic homology
and construct a natural I-symmetric monoidal structure on right modules over an NI∞ -algebra.

Contents
Introduction 2
1 I-symmetric monoidal categories and I-operads 11
1.1 Recollections on I-commutative monoids and I-symmetric monoidal ∞-categories . . . . . 11
1.2 Recollections on T -operads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 Restriction and arity-borelification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.4 (Co)cartesian I-symmetric monoidal ∞-categories . . . . . . . . . . . . . . . . . . . . . . . 21
2 I-commutative algebras 25
2.1 Indexed tensor products of I-commutative algebras . . . . . . . . . . . . . . . . . . . . . . 25
⊗
2.2 The smashing localization for NI∞ and Blumberg-Hill’s conjecture . . . . . . . . . . . . . 27
2.3 Failure of the nonunital equivariant Eckmann-Hilton argument . . . . . . . . . . . . . . . 30
3 Corollaries in higher algebra 33
3.1 Coherences and restrictions of equivariant Boardman-Vogt tensor products . . . . . . . . . 33
3.2 Disintegration and equivariant Boardman-Vogt tensor products . . . . . . . . . . . . . . . 35
3.3 Norms of right-modules over I-commutative algebras . . . . . . . . . . . . . . . . . . . . . 36
3.4 Equivariant infinitary Dunn additivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.5 Norms on Real topological Hochschild and cyclic homology . . . . . . . . . . . . . . . . . 38
Appendix A Technicalities on (co)cartesian I-symmetric monoidal ∞-categories 44
Appendix B I-operadic disintegration and assembly 51
Appendix C Algebraic patterns and the I-symmetric monoidal structure on overcategories 54
References 55

Date: April 4, 2025.

1
2 NATALIE STEWART

Introduction
We’re concerned with the relationship between homotopy-coherent interchange and equivariant com-
mutative algebras, incarnated via NI∞ -algebras (henceforth I-commutative algebras) in the sense of [BH15;
Ste25]. In particular, in [Ste25] we constructed a natural “pointwise” G-symmetric monoidal structure Alg⊗ (C)
O
on AlgO (C). We hope to answer the following questions, where CAlg⊗ (C) B Alg⊗ (C).
I NI∞
Questions. Let O ⊗ be a unital G-operad and I, J ⊂ FG a pair of unital weak indexing categories.
(I) When is the forgetful natural transformation AlgO CAlg⊗ (−) =⇒ CAlgI (−) an equivalence?
I
(II) When is the forgetful natural transformation CAlgI Alg⊗ (−) =⇒ AlgO (−) an equivalence?
O
(III) What is the (unique) G-operad O ⊗ with natural equivalence CAlgI CAlg⊗ (−) ≃ AlgO (−)? ◁
J
Each of the left hand sides of these proposed equivalences are corepresented by Boardman-Vogt tensor
products of G-operads, so these Questions (I) and (II) are equivalent to the question of when distinguished
⊗ BV
⊗ ⊗ BV ⊗
maps NI∞ → O ⊗ ⊗ NI∞ and O ⊗ → NI∞ ⊗ O are equivalences; moreover Question (III) asks the value of the
⊗ BV
⊗
tensor product NI∞ ⊗ NJ∞ . In this form, [BH15, Conj 6.27] conjectured an answer.
⊗ ⊗BV
⊗
Conjecture (Blumberg-Hill). If I and J are indexing categories then NI∞ ⊗ NJ∞ ≃ NI∨J∞ . ◁
We begin by completely characterizing G-operad algebras in (co)cartesian I-symmetric monoidal ∞-
categories: cocartesian I-symmetric monoidal structures are characterized by the property that their G-objects
canonically lift to O-algebras for any reduced I-operad O ⊗ , and cartesian I-symmetric monoidal structures are
characterized by an O-monoid formula generalizing [HA, Prop 2.4.2.5]. Using this, we show that Question (I)
is true precisely when O ⊗ is reduced and I-commutative algebras admit underlying O-algebra structures.
We conclude that the unique map E⊗ ⊗ ⊗
0 → NI∞ witnesses NI∞ as an idempotent algebra in OpG ;
uni

Question (II) asks to classify the associated smashing localization. Indeed, we show that the equivalence
holds whenever O-algebra G-spaces satisfy I-indexed Wirthmüller isomorphisms.
Since Alg (SG )-ambidextrous arities form a weak indexing category, we find that the intersection of the
O
⊗ ⊗ ⊗
NI∞ - and NJ∞ -smashing local categories is the NI∨J∞ -smashing local category, constructing an equivalence
⊗ BV ⊗ ⊗
NI∞ ⊗ NJ∞ ≃ NI∨J∞ in full generality. This answers Question (III) by affirming the evident unital extension
of Blumberg-Hill’s conjecture, constructing a (unique) natural equivalence
CAlgI CAlg⊗ (C) ≃ CAlgI∨J (C).
J
This is the third part of an ongoing project to develop the parameterized and equivariant higher algebra
predicted in [BDGNS16; NS22] into simply usable foundations for equivariant homotopy theory and K-theory
[Ste24; Ste25]; as such, we spend the last third of the paper fleshing out higher algebraic corollaries.
These corollaries fall into two classes: the first class gives Comm⊗
G ∈ OpG a unique idempotent algebra
structure, which determines a unique compatible idempotent algebra structure on its G-symmetric monoidal
envelope, enabling symmetric monoidality of the equivariant equifibered perspective of [BHS22; BS24b;
HK24]. From this, we lift OpG with the Boardman-Vogt tensor product to a canonical presentably symmetric
monoidal G-∞-category; as an application, we develop equivariant operadic disintegration and assembly, and
BV
the associated distributivity of ⊗ allows us to compute tensor products of unital G-operads whose underlying
G-∞-categories are G-spaces in terms of tensor products of reduced G-operads.
⊗
The second class simply applies Questions (I) and (II): by answering Question (II) for NI∞ ≃ E⊗∞ , we get
⊗
an O-symmetric monoidal structure on left modules over an O-algebra; for instance, specializing to O ⊗ = NJ∞
confirms a hypothesis of Hill [Hil17, Rmk 3.15].
By answering Question (I) for O ⊗ ≃ E⊗ V , we acquire an I-commutative algebra structure on (lax)
I-symmetric monoidal EV -algebra invariants of I-commutative algebras; for instance, this constructs an I-
commutative algebra structure on Real topological Hochschild homology and Real topological cyclic homology
of an I-commutative algebra whenever I-commutative algebras have underlying Eσ -algebras.
We now move to a more careful account of the background, motivation, and main results of this paper.
Background and motivation. Let C be a 1-category with finite products. Recall that a commutative monoid
in C is the data
A ∈ Ob(C), multiplication µ : A × A → A, and unit η : ∗ → A,
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 3

subject to the usual unitality, associativity, and commutativity assumptions; more generally, if (C, ⊗, 1) is a
symmetric monoidal 1-category, a commutative algebra in C is the data of
R ∈ Ob(C), multiplication µ : R ⊗ R → R, and unit η : 1 → R,
satisfying analogous conditions. When C = Set, this recovers the traditional theory of commutative monoids,
and when C = Modk with the tensor product of k-modules, this recovers the traditional theory of commutative
k-algebras. These have been the subject of a great deal of homotopy theory in three guises:
(i) We may define the (2, 1)-category Span(F) to have objects the finite sets, morphisms from X to Y
the spans of finite sets X ← R → Y , 2-cells the isomorphisms of spans
R
X ∼ Y,
R′
and composition the pullback of spans
RXZ

⌟
RXY RY Z
X Y Z.
If C is an ∞-category, then we define the ∞-category of commutative monoids in C as the C-valued
models of the associated Lawvere theory; that is, we define the product-preserving functor category
CMon(C) B Fun× (Span(F), C),
noting that products in Span(F) correspond with disjoint unions of finite sets. Indeed, if C is a
1-category and A a commutative monoid in C, we flesh this out with the dictionary
([2] = [2] → [1]) 7−→ µ : A×2 → A;
(∅ = ∅ → [1]) 7−→ η : ∗ ≃ A×0 → A;
([1] ← [2] = [2]) 7−→ ∆ : A → A×2
([1] ← ∅ = ∅) 7−→ ! : A → A×0 ≃ ∗.
Unitality, associativity, and commutativity are conveniently packaged by functoriality. This turns out
to be equivalent to Graeme Segal’s special Γ spaces [Seg74] when C = S, and for general C, it recovers
the anologously defined theory in C (see [BHS22, Ex 3.1.6, Prop 3.1.16, Prop 5.2.14]).
(ii) We say that a pointed ∞-category is semiadditive if it has finite
` productsQand coproducts and for
all finite sets S, the “identity matrix” natural transformation s∈S (−) =⇒ s∈S (−) is an equivalence.
The full subcategory PrL,⊕ ⊂ PrL of semiadditive presentable ∞-categories possesses a localization
functor L⊕ : PrL → PrL,⊕ , which we study.
(iii) Let Op denote the ∞-category of operads.1 Then, there is a terminal operad Comm⊗ ≃ E⊗ ∞ ; given C
a symmetric monoidal ∞-category, we may form the ∞-category of commutative algebra objects
CAlg(C) B AlgComm (C) ≃ AlgE∞ (C).
We study this and its specialization to the cartesian symmetric monoidal structure.
These three perspectives each present the same ∞-category, i.e. [Cra11; GGN15] show that
(1) CMon(C) ≃ CAlg(C × ) ≃ L⊕ C.
As a result, translating between these perspectives has proved invaluable; for instance, [GGN15] uses
Perspectives (ii) and (iii) to construct an essentially unique symmetric monoidal structure on CMon(C) and
[CHLL24a] uses Perspectives (i) and (iii) to model commutative algebras in CMon(C)⊗ as models for the
Lawvere theory of commutative semirings.
Crucially, Perspectives (i) and (iii) may be used to construct homotopical lifts of the Eckmann-Hilton
argument; for instance, in [SY19], it is shown that for any reduced operad O ⊗ , the forgetful functors
CAlgAlg⊗ ⊗
O (C) → CAlg(C) ← AlgO CAlg (C),

1 This is unambiguous [HM23], but we will tend to model these as ∞-operads in the sense of [HA].
4 NATALIE STEWART

are equivalences for the “pointwise” symmetric monoidal structure on algebras. Such an equivalence may
be exhibited by recognizing the far left and far right side each as algebras over the Boardman-Vogt tensor
product O ⊗ ⊗ Comm⊗ and each arrow as pullback along the canonical map
BV

can⊗id
Comm⊗ ≃ triv⊗ ⊗ Comm⊗ −−−−−−→ O ⊗ ⊗ Comm⊗ ;
BV BV

that Eq. (1) consists of equivalences reduces to the well-known fact that O ⊗ ⊗ Comm⊗ ∈ Op is terminal,
BV

which one can quickly prove via Perspectives (ii) and (iii).
This result is used ubiquitously to replace (lax) symmetric monoidal functors Alg⊗ ⊗
O (C) → C with (lax)
symmetric monodial endofunctors
CAlg⊗ (C) ≃ CAlg⊗ Alg⊗ ⊗
O (C) → CAlg (C);

for instance, this underlies the symmetric monoidal structure on left-modules [HA] and the multiplicative
structure on factorization homology [HA, Thm 5.5.3.2], TC [NS18, § IV.2], and algebraic K-theory [Bar15].
This paper concerns the analogs of Perspectives (i) to (iii) in the equivariant theory of algebra stemming
from Hill-Hopkins-Ravanel’s use of norms of G-spectra on the Kervarire invariant one problem, as well as the
resulting theory of indexed tensor products and (co)products (see [BDGNS16; HH16; NS22]).
For the rest of this introduction, fix G a finite group. In G-equivariant homotopy theory, the point is
replaced with elements of the orbit category OG ⊂ SetG , whose objects are homogeneous
 op G-sets
 [G/H]; indeed,
Elmendorf’s theorem [Elm83] realizes G-spaces as coefficient systems SG ≃ Fun OG , S .2 In G-equivariant
higher category theory, ∞-categories are thus replaced with G-∞-categories
 op 
CatG B Fun OG , Cat .

In G-equivariant higher algebra, following Perspective (i), we may form the effective Burnside 2-category
Span(FG ) whose objects are finite G-sets, whose morphisms are spans, whose 2-cells are isomorphisms of
spans, and whose composition is pullback; the following central definition is the heart of this subject.
Definition. The ∞-category of G-commutative monoids in C is the product-preserving functor ∞-category
CMonG (C) B Fun× (Span(FG ), C);
the ∞-category of small G-symmetric monoidal ∞-categories is
Cat⊗
G B CMonG (Cat). ◁
These are a homotopical lift of Dress’ semi-Mackey functors [Dre71] (c.f. [Lin76]). Indeed, given
C ⊗ ∈ Cat⊗
G a G-symmetric monoidal ∞-category, pullback along the product-preserving functor
∗7→G/H
ιH : Span(F) −−−−−−−→ Span(FG )
⊗
constructs a symmetric monoidal ∞-category CH B ι∗H C ⊗ whose underlying ∞-category CH is the value of C ⊗
on the orbit [G/H]. For all subgroups K ⊂ H ⊂ G, the covariant and contravariant functoriality of C ⊗ then
yield symmetric monoidal restriction and norm functors
⊗ ⊗
ResH
K : CH → C K ,
NKH : CK⊗ → CH
⊗
,
which satisfy a form of Mackey’s double coset formula.
Example ([BH21; CHLL24b]). There is a presentably G-symmetric monoidal ∞-category Sp⊗ with:
G
 
⊗ ⊗
• H-value given by the symmetric monoidal ∞-category Sp ≃ SpH of genuine H-spectra,
G H
⊗ ⊗
• restriction functors ResH
K : SpH → SpK given by the usual restriction functors, and
H ⊗ ⊗
• norm functors NK : SpK → SpH given by the HHR norm of [HHR16].
In fact, this structure is completely determined by its unit object SG ∈ Sp⊗
G. ◁

2 Maps [G/K] → [G/H] may equivalently be presented as elements of g such that gKg −1 ⊂ H , modulo K ; see e.g. [Die09] for details.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 5

Fix C ⊗ ∈ Cat⊗
G . If H ⊂ G is a subgroup and S ∈ FH a finite H-set, we may form the induced G-set
IndG
HS → [G/H], and the covariant and contravariant functoriality then yield an S-indexed tensor product
and S-indexed diagonal
OS
: CS → CH , ∆S : CH → CS .
K
Note that NHK is the [H/K]-indexed tensor product and ResH
Q
where CS B CK . K the [H/K]-indexed
[H/K]∈Orb(S)
diagonal. As explained in [Ste25], the “orbit collapse” factorization IndG
`
HS → [H/K]∈Orb(S) [G/H] → [G/H]
yields natural equivalences
S
O O  
XK ≃ NKH XK , ∆S (X) = ResH
K X ,
[H/K]∈Orb(S)
K [H/K]∈Orb(S)

so we may often reduce arguments about S-indexed tensor products to to binary tensor products and norms.
Similarly, we define the S-indexed tensor power
S 
O  OS O
⊗S
XH B ∆S X H ≃ ResK
H XH ≃ NKH ResH
K XH .
K K [H/K]∈Orb(S)

If it exists, the pointwise left-adjoint to ∆S is the indexed coproduct

S
a a
XK ≃ IndH
K S,
K [H/K]∈Orb(S)

where IndH
K is the left adjoint to the restriction map CH → CK . The indexed products are defined analogously.
Given H ⊂ G a subgroup, we say that C is H-pointed if CK is pointed for all (K) ⊂ (H). Given S ∈ FH ,
we say that S is C-ambidextrous if C is H-pointed, C admits S-indexed products and coproducts, and the
Wirthmüller natural transformation
aS S
Y
WS : (−) =⇒ (−)
K K
(called the norm in [Nar16, § 5]) is an equivalence. We say that C is G-semiadditive if S is C-ambidextrous
for all S ∈ FH and H ⊂ G. More generally, if FI ⊂ FG is a weak indexing system corresponding with the weak
indexing category I ⊂ FG (see [Ste24] or our review in Section 1.2), we say that C is I-semiadditive if S is
C-ambidextrous whenever S ∈ FI,H .
In this level of generality, Perspectives (i) and (ii) are known to present equivalent ∞-categories of
I-commutative monoids; indeed, the semiadditive closure theorem of [CLL24, Thm B] demonstrates that
PrL,I−⊕
G ⊂ PrLG is a smashing localization implemented by
 
LI−⊕ (C) ≃ CMonI (C) B Fun×G SpanI (FG ), C ,
and in particular, when C is a G-∞-category of coefficient systems
 op 
CoeffG (D)H B Fun OH , D ,
[CLL24, Thm C] yields the formula
 
CMonI CoeffG (D) ≃ Fun× (SpanI (FH ), D) ,
H
where SpanI (FH ) ⊂ Span(FH ) is the wide subcategory of spans whose forward maps lie in the restriction of
G
I to FH . Thus, we set the notation CMonI (D) B CMonI Coeff (D) ≃ Fun× (SpanI (FG ), D) and make the
G
following definition.
Definition. For I is a weak indexing category, the ∞-category of small I-symmetric monoidal ∞-categories is
Cat⊗ ×
I B Fun (SpanI (FG ), Cat) . ◁
6 NATALIE STEWART

Following through on Perspective (iii), algebraic objects X• in a G-symmetric monoidal ∞-category

⊗S
should possess collections of S-ary operations XH → XH subject to various coherences, controlled by a theory
of genuine equivariant operads; we use Nardin-Shah’s ∞-category OpG , whose objects we call G-operads.
Given O ⊗ ∈ OpG a G-operad, K ⊂ H ⊂ G a pair of subgroups, S ∈ FH a finite H-set, and Ti a finite Ki -set for
all orbits [H/Ki ] ⊂ S, in [Ste25] we constructed a space of S-ary operations O(S), operadic composition maps
 
O  a 
(2) γ : O(S) ⊗ O(Ti ) → O  IndH T  ,
 
 Ki i 
[H/Ki ]∈Orb(S) [H/Ki ]∈Orb(S)

operadic restriction maps

 
(3) Res : O(S) → O ResH
KS ,

and equivariant symmetric group action

(4) ρ : AutH (S) × O(S) → O(S).

We made the following simplifying definition.

Definition. A O ⊗ ∈ OpG has one color if O(∗H ) = ∗ for all H ⊂ G; these span a full subcategory Opoc
G ⊂ OpG . ◁
We showed in [Ste25, Thm A] that Eqs. (3) and (4) lift to a monadic functor Opoc G → Fun(TotΣG , S),
i.e. one color G-operads are monadic over G-symmetric sequences; in particular, (S 7→ O(S) | H ⊂ G, S ∈ FH )
are jointly conservative.
When O ⊗ has ⊗
 one color, an O-algebra in a G-symmetric monoidal ∞-category C can intuitively be
BWG (H)
viewed as a tuple XH ∈ CH with XK ≃ ResH
K XH for all K ⊂ H ⊂ G, together with O(S)-actions
G/H∈OG
 
⊗S
(5) µS : O(S) → MapCH XH , XH

for all H ⊂ G and S ∈ FH , homotopy-coherently compatible with Eqs. (2) to (4).3 We are concerned with the
following examples.
Example. There exists a terminal G-operad Comm⊗ G , which is characterized up to (unique) equivalence by
the property that CommG (S) is contractible for all S ∈ FH ; its algebras are endowed with contractible spaces
⊗S
of maps XH → XH for all S ∈ FH , as well as coherent homotopies witnessing their compatibility. We call
these G-commutative algberas.
On one hand, we saw in [Ste25, § 2.7] that CommG -algebras present a homotopical lift of Hill-Hopkins’
G-commutative monoids [HH16, § 4], though we prefer to reserve this name for the Cartesian case, following the
convention of [HA]. On the other hand, our model agrees with that of [CHLL24b], so the recent homotopical
Tambara functor theorem of Cnossen, Lenz, and Linskens [CHLL24b, Thm B] presents G-commutative algebra
objects in Sp⊗ (i.e. G-commutative ring spectra) as a form of homotopical G-Tambara functors.
G
Additionally, the recent rectification theorem of Lenz, Linskens, and Pützstück [LLP25] establishes
G-commutative ring spectra as a Dwyer-Kan localization of strict commutative algberas in symmetric (or
orthogonal) G-spectra at the weak equivalences transferred from a “positive stable” model structure. ◁
Example. Let V be a real orthogonal G-representation. There is a little V -disks G-operad E⊗
V whose structure
spaces are spaces of equivariant configurations:

EV (S) ≃ ConfH
S (V )

 This is modelled by the Steiner graph G-operad, so e.g. pointed G-spaces of the form
(see [Hil22; Hor19]).
X = ΩV Y B Map∗ S V , Y lift to EV -spaces by composition of loops [GM11; HHKWZ24]; moreover, many
EV -ring spectra may be constructed as Thom G-spectra of V -fold loop maps [HHKWZ24]. ◁

3 Here, W (H) = N (H)/H is the Weyl group of H ⊂ G, i.e. the automorphism group of the homogeneous G-set [G/H]. The
G G
restriction-compatible data specified above may be more familiarly referenced as a G-object ; it’s canonically extended from a choice
XG ∈ C G .
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 7

⊗
Example. Given I ⊂ FG a weak indexing category, in [Ste25] we constructed a weak N∞ G-operad NI∞ which
is characterized up to (unique) equivalence by its structure spaces

∗ S ∈ FI


(6) NI∞ (S) ≃ 
∅ S < FI


These recover the N∞ -operads of [BH15] when I is an indexing category, i.e. NI∞ (n · ∗G ) ≃ ∗ for n ∈ N; in
general, they are identified as the sub-terminal objects of OpG [Ste25, Thm C]. ◁
For instance, we verify in Corollary 3.15 that the condition V ⊕V ≃ V for an orthogonal G-representation
V implies that EV is a weak N∞ -operad, which is an N∞ -operad precisely when V G is positive-dimensional;
In particular, Comm⊗ ⊗ ⊗ ⊗
G ≃ E∞ρ ≃ NFT ∞ . Moreover, E∞ presents the initial N∞ -operad, and its algebras are
naive commutative algebra objects [Ste25, § 3.3]:
AlgE∞ (C) ≃ CAlg(CG ).
If I is an indexing category, the structure of an NI∞ -ring spectrum is intuitively viewed as commutative ring
structures on each spectrum XH , connected by multiplicative I-indexed norms, suitably compatible with the
restriction and (additive) transfer structures inherent to G-spectra. We refer to NI∞ -algebras in general as
I-commutative algebras and NI∞ -ring spectra as I-commutative ring spectra, writing
CAlgI (C) B AlgNI∞ (C).
In this paper, we are primarily concerned with homotopy coherently interchanging O− and P -algebra
BV
structures, which are implemented as algebras over Boardman-Vogt tensor product O ⊗ ⊗ P ⊗ of [Ste25]; in
⊗ BV
⊗
particular, we are concerned with computing NI∞ ⊗ NJ∞ , which corepresents pairs of interchanging I- and
J-commutative algebra structures.
⊗
To start, in [Ste25, § 2.6] we characterized I 7→ NI∞ as right adjoint to the arity support construction
 

 Y 

 
AO B  T →S O (T ×S [H/K]) , ∅ ⊂ FG ;
 

 

 [H/K]∈Orb(S) 

when O ⊗ , P ⊗ have one object, we will show that A (O ⊗ P ) = AO ∨ AP , the latter denoting the join in the
⊗ BV ⊗ ⊗
poset of weak indexing category. This constructs a unique pairing NI∞ ⊗ NJ∞ → NI∨J∞ .
Intuitively, given an algebra with I ∨ J-indexed norms, we may separate these into I- and J-indexed
norms together with coherent homotopies witnessing interchange between the two. Now, the transfer system
for I ∨ J consists of those inclusions K ⊂ H which can be factored as
K ⊂ KI1 ⊂ KJ1 ⊂ KI2 ⊂ · · · ⊂ KJn ⊂ H
where KIℓ ⊂ KJℓ is in I and KJℓ ⊂ KI(ℓ+1) is in J (see [Rub21, Prop 3.1]); intuition suggests that we may
combine interchanging I- and J-commutative algebra structures to construct an I ∨ J-commutative algebra
⊗ BV ⊗ ⊗
structure. Indeed, Blumberg and Hill conjectured that there is an equivalence NI∞ ⊗ NJ∞ ≃ NI∨J∞ [BH15,
Conj 6.27]; the main theorem of this paper confirms their conjecture in OpG , as well as characterizing exactly
how far we may weaken I and J.

Summary of main results. Recall that a weak indexing category I ⊂ FG is almost essentially unital if whenever
a non-isomorphism T ⊔ T ′ → S lies in I, the factor map T → S lies in I, and almost-unital if additionally
∗G ∈ I. We begin with a rigidity result for (co)cartesian I-symmetric ∞-categories under almost-unitality.
Theorem A. When I is almost-unital, there are fully faithful embeddings (−)I−⊔ and (−)I−× making the
following commute:
(−)I−⊔ (−)I−×
Cat⊔
I Cat⊗
I Cat×I
U
U U
CatG
The essential image of (−)I−⊔ is spanned by the I-symmetric monoidal ∞-categories whose I-indexed tensor
products are indexed coproducts, and (−)I−× by those whose I-indexed tensor products are indexed products.
8 NATALIE STEWART

Remark. After this introduction, we replace OG with an atomic orbital ∞-category T ; we prove Theorem A
as well as the other theorems in this introduction in this setting, greatly generalizing the stated results at the
cost of ease of exposition. ◁
We refer to I-symmetric monoidal ∞-categories of the form C I−× as cartesian, and C I−⊔ cocartesian.
Remark. Given I-symmetric monoidal ∞-categories C ⊗ and D⊗ and an I-product-preserving functor F : C → D
between their underlying G-∞-categories, we may define the ∞-category of I-symmetric monoidal lifts

Fun⊗,F ⊗ ⊗
G (C , D ) Fun⊗ ⊗ ⊗
G (C , D )
⌟

{F} FunG (C, D)

To interpret Theorem A as a rigidity theorem, note that it directly implies that whenever C ⊗ and D⊗ are
cartesian (resp. cocartesian), the core space Fun⊗,F ⊗ ⊗ ≃
G (C , D ) is contractible if F is I-product preserving
(I-coproduct preserving) and empty otherwise. Moreover, we confirm this fact without taking cores in
Proposition A.21 and Corollary A.23. ◁
To state our remaining theorems, we need the following definition.
Definition. An I-operad O ⊗ is unital if the unique map f : O ⊗ → NI∞ induces an equivalence
O(∅H ) ≃ NI∞ (∅H )
for all H ⊂ G (c.f. Eq. (6)); an I-operad is reduced if additionally f induces an equivalence
O(∗H ) ≃ NI∞ (∗H ).
A G-operad O ⊗ is almost essentially unital (resp almost essentially reduced ) if it’s unital (reduced) as an
AO-operad and AO is almost essentially unital. ◁
Algebraically, we identify cartesian I-commutative algebras with I-commutative monoids and cocartesian
(unital) I-commutative algebras with G-objects, identifying Perspectives (i) to (iii).
Theorem B. If I is almost-unital, C ⊗ is a cartesian I-symmetric monoidal ∞-category, and O ⊗ is an I-operad,
then the forgetful functor
U : AlgO (C) −→ FunG (TotG O ⊗ , C)
is fully faithful with image spanned by the G-functors TotG O ⊗ → C sending S-indexed tuples to S-indexed
products; in particular, this specializes to an equivalence
  ∼  
CAlgI C I−× −−−→ FunI−⊕
G FI,∗ , C .
In particular, in the case of coefficient systems, we acquire an equivalence
 
AlgO CoeffG DI−× ≃ SegTotTotG O (D) ≃ SegTotO (D),
where Seg(−) (−) refers to Segal objects in the sense of [CH21]. Hence there is an additional equivalence
 
CAlgI CoeffG DI−× ≃ CMonI (D).
Moreover, for all unital I-operads O ⊗ , the forgetful functor yields an equivalence
  ∼
AlgO C I−⊔ −−−−−→ FunG (U O, C).
References. This is Propositions 1.51 and 1.61 and Corollaries 1.52 to 1.54. □
In this theorem, TotTotG O is the total ∞-category of the fibration over FG,∗ and TotO is the total
∞-category of the fibration over Span(FG ).
 
Remark. The composed equivalence FunI−⊕ G F I,∗ , CoeffG
D ≃ Fun× (SpanI (FG ), D) is not new; indeed, it was
claimed for the complete weak indexing system as far back as [Nar16], it was proved in greater generality
than this article in [CLL24], and we verified in [Ste25, § A] that it also follows from [BHS22], as well as the
more general comparison between the two Segal object models for cartesian algebras. The new content is the
identification of these notions with G-operad algebras.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 9

Moreover, in the case that D = S and that I is an indexing category, this is a direct analog to [Mar24,
Thm A] in the ∞-categorical setting; the reader should interpret this relationship as a lift of Pavlov-Scholbach’s
comparison result [PS18, Thm 1.3] for a particularly nice choice of G-operad and value category. ◁
In Section 1.4 we verify that Alg⊗ (C) is cartesian when C is. Following this, in Section 2.1 we show
O
that I-indexed tensor products in CAlg⊗ C are indexed coproducts (i.e. its underlying I-symmetric monoidal
I
⊗
∞-category is cocartesian) and that this completely characterizes NI∞ . The heart of our strategy uses the
explicit monadic description of [Ste25, § 2.4] to reduce to the case of G-spaces C ⊗ ≃ S G−×
G ; in this case, we may
⊗ G−×
easily see that the cartesian I-symmetric monoidal ∞-category CAlg (S G ) ≃ CMonI (S G )I−× is cocartesian,
I
as its underlying G-∞-category is I-semiadditive by [CLL24, Thm B-C]. We conclude the following.
Theorem C. Let O ⊗ be an almost essentially reduced G-operad. Then, the following conditions are equivalent.
(a) The G-∞-category Alg S G is AO-semiadditive.
O
⊗
(b) The unique map O ⊗ → NAO∞ is an equivalence.
Moreover, for all almost essentially unital weak indexing categories I and I-symmetric monoidal ∞-categories
C ⊗ , the I-symmetric monoidal ∞-category CAlg⊗ C is cocartesian.
I

Theorems B and C together with conservativity of Alg (S G ) (as in [Ste25, § 2.4]) yields the following.
(−)
⊗ BV
⊗
Corollary D. NI∞ ⊗ NI∞ is a weak N∞ -operad if and only if I is almost essentially unital. In this case, if
⊗ ⊗
O is a reduced I-operad, then the unique map O ⊗ ⊗ NI∞
⊗ → NI∞ is an equivalence.
In particular, whenever I is almost unital, there exists a map triv⊗ ⊗ ⊗
G → NI∞ witnessing NI∞ as an
idempotent object in OpG . We verified in [Ste25, Thm D] that Env : OpG → Cat⊗ G is compatible with the
unit and tensor products under the mode symmetric monoidal structure on Cat⊗G ; this yields a ⊛-idempotent
algebra structure on FG = Env(CommG ) ∈ CatG , and hence a symmetric monoidal structure on Cat⊗ G−⊔ .
G−⊔ ⊗
G,/FG
We acquire an equivariantization of a modification of [BS24a, Thm E].
Corollary E. There exists a unique symmetric monoidal structure Op⊗ on Op attaining a (necessarily
G G
unique) symmetric monoidal structure on the fully faithful G-functor
G−⊔
Env/FG : Op⊗ −→ Cat⊗
G G,/FG−⊔
G
BV
of [BHS22; NS22] with respect to ⊛; the tensor product of this structure is ⊗ .
Idempotent objects correspond with smashing localizations, i.e. they classify particular properties [HA,
⊗
§ 4.8.2]; in Theorem 2.6, we conclude that the smashing localization corresponding with NI∞ ∈ Opred
J classifies
the property of having I-indexed Wirthmüller isomorphisms
S S
∼
a O
⊗
O ⊗ ⊗ NI∞ ≃ O⊗ ∀C ⊗ ∈ Cat⊗
BV
⇐⇒ J , ∀S ∈ FI,V , ∀(XU )S ∈ Alg (C)S WS :
O
XU −−−→ XU
U U
⇐⇒ Alg (S G ) is I-semiadditive.
O
Recall that tensor products of idempotents algebras are idempotent algebras, classifying the intersection of
the associated smashing localizations [CSY20, Prop 5.1.8]; conveniently, indexed semiadditivity is classified
by a weak indexing category [Ste25, § 1.2], so Alg (S G ) is I ∨ J-semiadditive if and only if it is I-semiadditive
O BV
and J-semiadditive. This allows us to affirm Blumberg and Hill’s conjecture with respect to ⊗ .
⊗
Theorem F. N(−)∞ : wIndexG → OpG restricts to a fully faithful symmetric monoidal G-right adjoint
A

wIndexaEuni OpaEuni
⊣

G G

⊗
N(−)∞
10 NATALIE STEWART

Furthermore, the resulting tensor product of weak N∞ -operads is computed by the Borelified join
⊗ ⊗
≃N⊗
BV
NI∞ ⊗ NJ∞ .
BorG
cI∩cJ (I∨J)∞

Hence when I, J are almost-unital weak indexing categories and C ⊗ is an I ∨ J-symmetric monoidal ∞-category,
there is a canonical equivalence of I ∨ J-symmetric monoidal ∞-categories

CAlg⊗ CAlg⊗ (C) ≃ CAlg⊗ (C).

I J I∨J

For instance, using [CHLL24b, Thm 4.3.6] to identify I-Tambara functors in an ∞-category C with
I-commutative algebras in Mackey functors, this confirms that I ∨ J-Tambara functors are equivalent to
I-commutative algebras in J-Tambara functors with respect to the box product.
Remark. The reader interested in computing tensor products of G-operads may benefit from reading the
combinatorial characterization of joins of weak indexing systems in terms of closures in [Ste24, § 2.3]; there,
we prove that the join of weak indexing systems FI ∨ FJ is computed by closing the union FI ∪ FJ under
iterated I and J-indexed coproducts. ◁

Relationship to the literature. There are three main bodies of literature which present results in homotopy-
coherently equivariant algebra: the model categorical, the atomic orbital, and the global. We now attempt to
give a bit of a Rosetta stone to connect our definitions to the model categorical and global settings.
We established in [Ste24] that our weak indexing categories specialize to Blumberg-Hill’s indexing
categories [BH18] in the case T = OG and n · ∗G → ∗G lies in I for all n ∈ N, and our weak indexing systems to
the indexing systems of [BH15] when n · ∗G ∈ FI,G for all n ∈ N; moreover, this was shown to be compatible
with Bonventre’s nerve in [Bon19; Ste25], which is intertwines with the underlying G-symmetric sequence and
restricting to an equivalence on at-most-one-color G-operads with 0-truncated structure spaces, showing that
our weak N∞ -operads specialize to those of [BH18; BP21; GW18; Rub21]. Additionally, we saw in [Ste24]
that our weak indexing categories specialize the weakly extensive span pairs of [CHLL24b] to the case that
the larger category is FT .
We saw in [Ste25] that our algebras agree with Blumberg-Hill’s in the discrete setting, and combining
Corollary 1.54 with the main result of [Mar24] identifies a Dwyer-Kan localization of the latter with the
former in the case S G−×
G . Of course, the recent results of Lenz, Linskens, and Pütszstück [LLP25, Thm A]
have established rectification to G-commutative ring spectra, establishing them as presented by Hammock
localization of a right-transferred structure on commutative algebras, CAlg(SpΣ G ), with respect to the positive
stable model structure on symmetric G-spectra (or equivalently, on orthogonal G-spectra). The author is not
aware of any study into rectification for the incomplete case.
Our I-symmetric monoidal ∞-categories and I-operads generalize [NS22] and specialize [LLP25] by
[Ste25, § A] and by definition, respectively. Moreover, there is a homotopical operadic nerve construction
mapping the (equivalent) settings of [BH15; BP21; Per18] to OpI and Cat⊗ I by [Bon19; Ste25].
BV
The author is not aware of a comparison result between ⊗ and the point-set Boardman-Vogt tensor
product appearing in [BH15]; moreover, the “derived” tensor product appearing in [Rub21] is only defined on
an ∞-category which is equivalent to IndexG , so it’s not clear that it makes sense to ask for a comparison to
BV BV
⊗ other than confirming that ⊗ confirms Blumberg-Hill’s conjecture, as demanded by the results of [Rub21,
Thm A] (after which Rubin explicitly claimed that the conjecture remained open).
For (co)cartesian I-symmetric monoidal ∞-categories, we show in Section 1.4 that our definitions
generalize [NS22] and agree with [CHLL24b] when both are defined; in particular, we recover a non-(∞, 2)-
categorical reproof of the identification theorem of the two, as conjectured in [CHLL24b] and verified in
[CLR25]. The author is not aware of a definition to be compared with in the model-categorical setting, but
comparisons with such constructions will be as easy as verifying that those structures have indexed tensor
products which present (derived) indexed products.
Some versions of our results on cocartesianness and algebras are proved independently in the literature;
though it is not clear that Nardin-Shah’s T -operad of algebras [NS22] agrees with ours, they confirmed that
their version of CAlg⊗ (C) is cocartesian and claimed that CAlg⊗ (C) is I-cocartesian when I is an indexing
T I
category. Moreover, it is shown in [LLP25, Prop 2.26] that (−)I−⊔ admits a left adjoint, but this left adjoint
is not computed therein (whereas we confirm it to be U in the almost-unital case as Corollary 2.16).
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 11

Notation and conventions. We assume that the reader is familiar with the technology of higher category
theory and higher algebra as developed in [HTT] and [HA, § 2-3], though we encourage the reader to engage
with such technologies via a “big picture” perspective akin to that of [Gep19, § 1-2] and [Hau23, § 1-3]. In
particular, our treatment is almost entirely model agnostic–we only pierce the veil in Appendix A.1 and use
quasicategorical language in order to verify that a few functors are exponentiable.
We additionally assume that the reader is familiar with parameterized higher category theory over an
∞-category as developed in [Sha22; Sha23]; the material reviewed in the prequel [Ste25, § 1] will be enough.
In particular,
• T will always be an atomic orbital ∞-category in the sense of [NS22], FT its corresponding ∞-category
of finite T -sets, and FT its corresponding T -1-category of finite T -sets.
• F ⊂ T will always be a T -family in the sense of [Ste24].
• I ⊂ FT and FI will always be a weak indexing category and corresponding weak indexing system in
the sense of [Ste24]. c(I) will be its color family and υ(I) its unit family.
• Cat will always be the ∞-category of small ∞-categories, CatT of small T -∞-categories, and Cat⊗ I
of small I-symmetric monoidal ∞-categories. All ∞-categories will be assumed to be small unless
otherwise mentioned.
• T -operad will always mean T -∞-operad in the sense of [NS22] and OpT the ∞-category of T -operads.
Acknowledgements. I’m indebted to Maxime Ramzi for disillusioning me to a fatally flawed strategy on
work related to this paper, leading me to the drawing board; it wasn’t so clear at the time, but it was in
subsequent conversation with him that the main idea of this paper emerged. I am additionally grateful to Piotr
Pstrągowski, who pointed out a mistake in my early strategy in this paper, leading to the condition of almost
essential unitality on the main theorem. Also, I owe Mike Hill for pointing out to me that pullback-stable
subcategories are replete, obviating one of the assumptions on weak indexing categories.
Additionally, I would like to thank Andy Senger, Clark Barwick, and Dhilan Lahoti, with whom I had
enlightening (to me) conversations about the topic of this paper. Of course, none of this work would be
possible without the help of my advisor, Mike Hopkins, who I’d like to thank for many helpful conversations.
While developing this material, the author was supported by the NSF Grant No. DGE 2140743.

1. I-symmetric monoidal categories and I-operads

We begin in Section 1.1 by recalling results of [CLL24; Nar16; NS22; Ste24; Ste25] concerning the
theory of I-commutative monoids and I-symmetric monoidal ∞-categories. Moving on, in Section 1.2 we
recall results of [NS22; Ste25] concerning T -operads; in either case, all reviewed information was used in
the preceding article [Ste25]. We then go on in Section 1.3 to begin to carefully study the interactions of
restriction, arity-borelification, arity-support, and Boardman-Vogt tensor products. We finish the section
in Section 1.4, where we develop a number of foundational results on (co)cartesian I-symmetric monoidal
∞-categories, ultimately elaborating on the technical minutiae of Appendix A.
1.1. Recollections on I-commutative monoids and I-symmetric monoidal ∞-categories. For the rest of this
paper, we fix T an atomic orbital ∞-category.
1.1.1. Weak indexing systems and semiadditivity. We will use the following machinery of [Ste24].
Definition 1.1. A T -weak indexing category is a subcategory I ⊂ FT satisfying the following conditions:
(IC-a) (restrictions) I is stable under arbitrary pullbacks in FT , and
(IC-b) (segal condition) T → S and T ′ → S are both in I if and only if T ⊔ T ′ → S ⊔ S ′ is in I.
A T -weak indexing system is a full T -subcategory FI ⊂ FT satisfying the following conditions:
(IS-a) whenever the V -value FI,V B (FI )V is nonempty, we have ∗V ∈ FI,V , and
(IS-b) FI ⊂ FT is closed under FI -indexed coproducts. ◁
We say that a T -weak indexing system FI :
(i) has one color if for all V ∈ T , we have FI,V , ∅,
(ii) is almost essentially unital (or aE-unital) if whenever FI has a non-contractible V -set, ∅V ∈ FI,V ,
(iii) is almost-unital (or a-unital) it’s almost essentially unital and has one color,
(iv) is unital if ∅V ∈ FI,V for all V ∈ T , and
(v) is an indexing system if the subcategory FI,V ⊂ FV is closed under finite coproducts for all V ∈ T .
12 NATALIE STEWART

These occupy embedded sub-posets

IndexT ⊂ wIndexuni auni
T wIndexT ⊂ wIndexaEuni
T ⊂ wIndexT .
Given a weak indexing category I ⊂ FT , we denote the I-admissible V -sets by
n o
FI,V B S ∈ FI,V | IndTV S → V ∈ I ⊂ FV ;
Condition (IC-a) guarantees that these assemble into a full T -subcategory FI ⊂ FT , which contains all of
the information of I by Condition (IC-b). In [Ste24, Thm A] we proved the following and expressed the
conditions of Definition 1.1 in the language of weak indexing categories.
Proposition 1.2 (Generalized [BH18, Thm 1.4]). The assignment I 7→ FI implements an equivalence between
the posets of T -weak indexing categories and T -weak indexing systems.
Intuitively, Condition (IS-a) corresponds with identity arrows in I and Condition (IS-b) with composition.
We will need the following invariants of weak indexing systems.
Construction 1.3. Given FI a weak indexing system, we define the color and unit families

c(I) B V ∈ T | ∗V ∈ FI,V ⊂ T ;

υ(I) B V ∈ T | ∅V ∈ FI,V ⊂ T .
Indeed, we saw that these are families in [Ste25]. ◁
One reason to study this is indexed semiadditivity in the sense of the following definitions of Nardin.
Definition 1.4. A T -∞-category C is said to be V -pointed if CU is a pointed ∞-category for all U → V . Given
S ∈ FV is a finite V -set, C V -pointed admitting S-indexed products and coproducts, and (XU )S ∈ CS an
S-tuple, we define the S-indexed Wirthmuller map WS,(XU ) : SU XU → SU XU to extend the following maps
` Q
via the universal property for S-indexed coproducts:
V S−W
ResY V S−W
ResY
W U S
(id;!) Y
WS,(XU ),W : XW ≃ XW × ∗U ′ −−−−−−−→ XW × ResVU ′ Xo(U ′ ) ≃ ResVW XU
U′ U′ U
where o(U ′ ) ∈ Orb(S) is the orbit whose restriction contains U ′.
We say that S is C-ambidextrous if C is
V -pointed and WS,(XU ) is an equivalence for all (XU ) ∈ CS ; given FI a weak indexing system, we say that C is
I-semiadditive if S is C-ambidextrous for all S ∈ FI . ◁
Remark 1.5. The map WS,(XU ),W is determined via the universal property for S-indexed products by its
projections WS,(XU ),W ,W ′ : XW → ResVW CoIndVW ′ XW ′ , which are zero when W , W ′ , and otherwise they are
the map induced under functoriality of products by the dashed arrow

IndVW ∗W

IndVW ResVW IndVW ∗W IndVW ∗W

⌟

IndVW ∗W V
In particular, they match the norms constructed in [Nar16]. ◁
In [Ste25] we proved that the collection of C-ambidextrous finite V -sets form a weak indexing system
and concluded the following important observation.
Proposition 1.6 ([Ste25, § 1.2]). Let ∨ denote the join in wIndexCatT . Then, C is I-semiadditive and
J-semiadditive if and only if C is I ∨ J-semiadditive.
1.1.2. I-commutative monoids. In [Bar14], the notion of adequate triple was defined, consisting of triples
(C, Cb , Cf ) with Cf , CB ⊂ C a pair of wide subcategories satisfying pullback-stability and distributivity conditions;
if I is a weak indexing category, then (Fc(I) , Fc(I) , I) is an adequate triple.
Adequate triples form a full subcategory TripAdeq ⊂ Fun(• → • ← •, Cat); [Bar14] constructed a functor
Span−,− (−) : TripAdeq → Cat,
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 13

called the effective Burnside category. In the case that c(I) is a 1-category (e.g. T has a terminal object, see
[NS22, Prop 2.5.1]), Fc(I) is a 1-category, so the effective Burnside category
SpanI (FT ) B SpanFc(I) ,I (Fc(I) )
f
is a (2, 1)-category with objects agreeing with Fc(I) , morphisms the spans X ← R −
→ Y with f in I, 2-cells the
isomorphisms of spans, and composition of morphisms computed by pullbacks in Fc(I) (which are guaranteed
to be morphisms in SpanI (FT ) by pullback-stability of I).
Much of the technical work of [Bar14; BGS20] has been extended by [HHLN23], so we generally refer
the reader there. At any rate, we recall this in order to define homotopical incomplete semi-Mackey functors
for I, which we call I-commutative monoids.
Definition 1.7. If C is an ∞-category with finite products, then an I-commutative monoid in C is a product-
preserving functor SpanI (FT ) → C. More generally, if D is a T -∞-category with I-indexed products, then an
I-commutative monoid in D is an I-product-preserving T -functor SpanI (FT ) → D. We write
CMon(D) B FunI−×
T (SpanI (FT ), D)
CMon(D) B Γ T CMon(D)
 
CMon(C) B CMon CoeffT C
 
CMon(C) B CMon CoeffT C . ◁
An important result of Cnossen-Lenz-Linskens resolves the notational clash.
Proposition 1.8 ([CLL24, Thm C]). When C is an ∞-category, restriction furnishes an equivalence
CMon(C) ≃ Fun× (SpanI (FT ), C) ,
and more generally, we have CMon(C)V ≃ Fun×V (SpanI (FV ), C) with restriction given by pullback along
SpanI (FV ) → SpanI (FW ).
Let I be a one-object weak indexing category and let CatI−×
T ⊂ CatT be the (non-full) subcategory whose
objects are T -∞-categories admitting I-indexed products and functors preserving I-indexed products. Let
CatI−⊕
I ⊂ CatI−×
T be the full subcategory spanned by I-semiadditive T -∞-categories. The following result is
fundamental in the theory of equivariant semiadditivity and equivariant higher algebra.
Theorem 1.9 ([CLL24, Thm B]). The inclusion CatI−⊕
T ⊂ CatI−×
T has left adjoint CMon(−).
1.1.3. I-symmetric monoidal ∞-categories. The following definition is central to equivariant higher algebra.
Definition 1.10. The ∞-category of small I-symmetric monoidal ∞-categories is Cat⊗
I B CMonI (Cat). ◁
We refer to maps in Cat⊗
I as I-symmetric monoidal functors. An important lemma is the following.
Lemma 1.11 ([CH21, Cor 8.2]). If C is an ∞-category and I a one-object weak indexing category, then the
underlying coefficient system functor CMonI (C) → CoeffT C is conservative; in particular, if a I-symmetric
monoidal functor’s underlying T -functor is an equivalence, then it is a T -symmetric monoidal equivalence.
Now, these are defined for the following notation’s sense.
Notation 1.12. Given an I-symmetric monoidal ∞-category and an I-map IndTV S → V with V ∈ T , we denote
NS S
the covariant functorialty of C ⊗ by U : CS → CV and the contravariant functoriality by ∆ : CV → CS . ◁
We will also need presentability. In his thesis, Nardin defined a T -symmetric monoidal ∞-category
PrL,⊗
T of presentable T -∞-categories, whose S-ary tensor products are characterized by mapping ∞-categories
 S   S 
O  Y 
L  S−∂
FunT 
 CU , D B FunT 
  CU , D
U U

where FunS−∂
T consists of the “S-distributive T -functors.” We will not care too much about the details of this
in general, and instead shunt the interested the reader to [QS22b, Def 5.14]. Nevertheless, we care about the
following case.
14 NATALIE STEWART

2∗ −∂
Example 1.13. FunV V (C × D, E) ⊂ Fun (C × D, E) is the full subcategory of T -functors whose fibers C ×
{D} , {C} × D → E all strongly preserve T -colimits. ◁
Now, we make the following definition.
Definition 1.14. A distributive I-symmetric monoidal ∞-category is an I-symmetric monoidal ∞-category C ⊗
whose S-tensor functors ⊗SU : CS → CV are S-distributive. A presentably I-symmetric monoidal ∞-category is
a distributive I-symmetric monoidal ∞-category whose underlying c(I)-∞-category is presentable. ◁
We will also need the following construction.
Construction 1.15. The “opposite category” construction op : Cat → Cat is an equivalence, so in particular it
is product-preserving. Hence postcomposition with (−)op yields fiberwise opposite functor
(−)vop : Cat⊗ ⊗
I → CatI .
vop
Note that the underlying category (C ⊗ ) is the traditional vertical opposite T -∞-category C vop . ◁

1.2. Recollections on T -operads.

1.2.1. I-operads. In [Ste25], we made the following definition.

Definition 1.16. An I-operad is a functor π : TotO ⊗ → SpanI (FT ) satisfying the following conditions.
(a) Tot‘O ⊗ has π-cocartesian lifts for backwards maps in SpanI (FT );
(b) (Segal condition for colors) writing OS ≃ π−1 (S), for every S ∈ FT , cocartesian transport along the
π-cocartesian lifts lying over the inclusions (S ← U = U | U ∈ Orb(S)) together induce an equivalence
Y
OS ≃ OU ;
U ∈Orb(S)

(c) (Segal condition for multimorphisms) for every map of orbits T → S in I and pair of objects
(C, D) ∈ OT × OU , postcomposition with the π-cocartesian lifts D → DU lying over the inclusions
(S ← U = U | U ∈ Orb(S)) induces an equivalence
Y
T ←T →U
MapTTotO
→S
⊗ (C, D) ≃ MapTotO⊗U (C, DU ).
U ∈Orb(S)

where TU B T ×S U .
The ∞-category of I-operads is defined to be a localizing subcategory
LOpI

OpI Catint−cocart
/ Span (F ) ;
I T

that is, a morphism of I-operads is a functor TotO ⊗ → TotP ⊗ over SpanI (FT )sending πO -cocartesian mor-
phisms to πP -cocartesian morphisms. We also call these O-algebras in P and we let
AlgO (P ) B Funint−cocart ⊗ ⊗ ⊗ ⊗
/ Span (F ) (O , P ) ⊂ Fun/ SpanI (FT ) (O , P )
I T

be the full subcategory spanned by O-algebras in P . ◁

This doesn’t obviously recover the notion of [NS22]. To discuss the comparison, we temporarily assume
the reader is familiar with fibrous patterns in the sense of [BHS22] (which are essentially weak Segal fibrations
in the sense of [CH21]).
Construction 1.17. Let TotF∨
T be the cartesian unstraightening of the functor V 7→ FV , so that its objects are
T -arrows S → U with V ∈ T and its morphisms f : (T → V ) −→ (S → U ) are commutative diagrams between
arrows, i.e.
fs

T f◦ S ×U V S
⌟

ft
V U
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 15

We say that f is s.i. if f ◦ is a summand inclusion and I-tdeg if ft is an identity arrow and fs lies in I. Then,
we define the algebraic pattern  
TotFI,∗ B Spans.i.,I−tdeg TotF∨T .
The map of triples (TotF∨
T , s.i., I − tdeg) → (FT , all, I) induces a Segal morphism s : TotFI,∗ → SpanI (FT ). ◁
We recover Nardin-Shah’s notion of T -∞-operads by the following result.
Proposition 1.18 ([BHS22; Ste25]). The ∞-category OpT ,∞ of [NS22] is equivalent to fibrous TotFT ,∗ -patterns,
and s induces an equivalence
∼  
OpI ≃ Fbrs (SpanI (FT )) −−−−−→ Fbrs TotFI,∗ .
In particular, we get a composite functor
s∗ U
TotT : OpI ,→ Catint−cocart −−→ CatT ,/FT ,I −−−→ CatT
/ Span (F ) − I T

The following observation about this composite functor is key. We greatly strengthen it in Appendix A.6.
Observation 1.19. TotT : OpI → CatT is conservative, since each of the component arrows are conservative. ◁
Now, it follows by unwinding definitions that a cocartesian fibration π : TotT → Span(FT ) is an O⊗
I-operad if and only if its unstraightening SpanI (FT ) → Cat is an I-symmetric monoidal category. [BHS22]
and [NS22] thus independently construct an adjunction
EnvI

OpI Cat⊗
I.
⊣

U
⊗ ⊗
Now, OpI has a terminal object and in [Ste25] we computed EnvI NI∞
NI∞ , ≃ FI−⊔
I , i.e. it is the weak
indexing system for I with indexed tensor products given by indexed coproducts; [BHS22, Prop 4.21] then
/FI−⊔
verifies that the sliced left adjoint EnvI I
: OpI → Cat⊗ is fully faithful and identifies its image, i.e. OpI
I,/FI−⊔
I
is a colocalizing subcategory of I-symmetric monoidal ∞-categories over FI−⊔
I consisting of the equifibrations.
Now, the following construction is very occasionally important.
Construction 1.20. Given I a weak indexing category and V ∈ T , define the T/V -weak indexing category IV to
consist of those maps over V which lie in I. Define
SpanI (FV ) B SpanIV (FV ).
This is evidently functorial under unslicing functors; in particular, pullback along SpanI (FV ) → SpanI (FW )
yields a functor
ResW
V : OpIW → OpIV .
We refer to the associated T -∞-category as Op . ◁
I

1.2.2. The underlying T -symmetric sequence.

Definition 1.21. The underlying T -∞-category U O of an I-operad O ⊗ is the straightening of the pullback
TotU O O⊗
⌟

T op SpanI (FT )

A T -operad has at most one color if each value U OV is either empty or contractible, has at least one color
if U OV is nonempty for each V , has has one color if U O ≃ ∗T . These occupy full subcategories
≤oc ≥oc
Opoc
I ⊂ OpI , OpI ⊂ OpI . ◁
In [Ste25, § 2.3] we defined an underlying T -symmetric sequence functor and proved the following.
Theorem 1.22 ([Ste25, Thm A]). The underlying T -symmetric sequence functor sseq : Op≤oc
T → Fun(TotΣT , S)
is monadic; in particular, it is conservative.
16 NATALIE STEWART

The V -objects in ΣT ≃ F≃ ≃ ⊗
T are finite V -sets; given S ∈ ΣV ≃ FV , writing O(S) for sseq O (S), we
remember this as saying that at-most-one-color T -operads are identified conservatively by their S-ary
structure spaces. Using this, we defined the full subcategory of T -d-operads as those with (d − 1)-truncated
structure spaces: n o
OpT ,d B O ⊗ | ∀S, O(S) ∈ S≤(d−1) ⊂ OpT
In [Ste25, § 2.5], we verified the following.
Proposition 1.23 ([Ste25]). The inclusion OpT ,d ⊂ OpT has a left adjoint hd : OpT → OpT ,d , and given
P ⊗ ∈ OpT ,d , the ∞-category AlgO (P ) is a d-category; moreover, if P ⊗ ∈ OpT ,0 , then AlgO (P ) is either empty
or contractible. In particular, OpT ,d is a (d + 1)-category and OpT ,0 is a poset.
We call hd O ⊗ the homotopy T -d-operad of O ⊗ . We went on to compute the free O-algebra monad ; for
algebras in a cartesian structure on coefficient systems in a cocomplete cartesian closed ∞-category C, this
sends X ∈ CoeffT C to the coefficient system TO X with
 
a  Y 
V U
(TO X) ≃ FrC O(S) × X  ,

 
S∈FV U ∈Orb(S) h AutV (S)

where FrC : S → C is the unique symmetric monoidal left adjoint. In particular, given S ∈ FV , in [Ste25]
we found a natural splitting FrC O(S) ⊕ J ≃ (TO S)V . A multiple-color version of this argument yielded the
following.
Proposition 1.24 ([Ste25, § 2.4]). A map of T -operads ϕ : O ⊗ → P ⊗ is an hd -equivalence if and only if:
(a) the underlying T -functor U ϕ : U O → U P is essentially
 surjective,
 and
(b) the pullback functor ϕ ∗ : AlgP (S T ,≤(d−1) ) → AlgO S T ,≤(d−1) is an equivalence.
In particular, ϕ is an equivalence if and only if it is U -essentially surjective and induces an equivalence on
algebras in S T .
1.2.3. Rudiments of weak N∞ -operads. In [Ste25, § 2.2], we constructed a family of T -operads:
Proposition 1.25 ([Ste25]). Let I ⊂ J ⊂ FT be pullback-stable subcategories. Then, SpanI (FT ) → SpanJ (FT )
presents a J-operad if and only if I is a weak indexing category.
These are called weak N∞ -operads; in the case that I is an indexing category, these are called N∞ -operads.
To state their universal property, we defined the arity support subcategory
 

 Y 

 
AO B  T →S O(T ×S U ) , ∅ ⊂ FT ,
 

 

 U ∈Orb(S) 

Theorem 1.26 ([Ste25, § 2.6]). The arity support of a T -operad is a weak indexing category, and the associated
essential surjection admits a fully faithful right adjoint
A

OpT wIndexCatT
⊣

N(−)∞

The essential image of N(−)∞ is spanned by T -operads O ⊗ satisfying the following equivalent conditions.
(a) O ⊗ is a weak N∞ -operad.
(b) O ⊗ is a T -0-operad.
(c) The map of T -operads O ⊗ → Comm⊗
T is a monomorphism.

In particular, this isolates the weak N∞ -operads as those possessing a fully faithful unslicing functor
OpT ,/NI∞ ,→ OpT .
This yields functors for change of weak indexing systems, which we use to generically specialize to I = T .
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 17

Proposition 1.27 ([Ste25]). Postcomposition along SpanI (FT ) → SpanJ (FT ) yields a fully faithful embedding

OpI ≃ OpJ,/N ⊗ ,→ OpJ .

I∞

We denote the associated push-pull adjunction as

J
EI

(7) OpI OpJ ;

⊣
J
BorI

J J
it follows from Theorem 1.26 that ABorI O ≃ I ∩ AO and AEI O ≃ AO. We have several examples:
Example 1.28. Let I triv be the initial one-color weak indexing category. The corresponding weak N∞ -operad
triv⊗ ⊗
T B NI triv ∞ is called the trivial T -operad, and it is characterized by its algebras [NS22; Ste25]

CAlg (O) ≃ U O;
I triv

in particular, the restriction of the underlying T -∞-category construction to I triv -operads yields an equivalence
∼ J
OpI triv −
→ CatT and EI triv is compatible with U [NS22]; that is, Eq. (7) takes the form of an adjunction

triv(−)⊗

CatT OpT ,
⊣

i.e. given a T -∞-category C, we acquire a T -operad characterized by its algebras [NS22]

Algtriv(C) (O) ≃ FunT (C, U O);

this formula allows for an alternative construction for triv(C) is as the operadic localization [Ste25]

triv(C)⊗ ≃ LOpT (C → T op → Span(FT )) . ◁

Example 1.29. Given F ⊂ T a family, define


{∅V , ∗V } V ∈ F

F0T ,F ,V

B
{∗V }
 V <F.

Let IT0 ,F be the associated weak indexing category; this is the initial one-color weak indexing category with
υ(I) ⊃ F . We define E⊗ ⊗
0,F B N 0 ; in particular, we write E⊗ ⊗
0 B E0,T .
IT ,F ∞
It was shown in [NS22, Thm 5.2.10] (and re-shown in [Ste25, § 3.3]) that there is an equivalence
 ⊗
Γ T Alg⊗ (C) ≃ Γ T C . ◁
E0 1/

Example 1.30. Let F∞ be the initial indexing system, i.e.

F∞
V B {n · ∗V | n ∈ N} .

⊗
Let I ∞ be its indexing category. We write E⊗
∞ B NI ∞ ∞ . In [Ste25, § 3.3] we constructed an equivalence
 
Γ T Alg⊗ (C) ≃ CAlg⊗ Γ T C . ◁
E∞

Example 1.31. The terminal T -operad Comm⊗

T = (Span(FT ) = Span(FT )) is the N∞ -operad for the terminal
indexing category FT = FT . ◁
18 NATALIE STEWART

1.2.4. The Boardman-Vogt tensor product. In [Ste25, Thm D], in the case that T has a terminal object, we
equipped OpT with a closed Boardman-Vogt tensor product
 
∧
O ⊗ ⊗ P ⊗ B LOp O ⊗ × P ⊗ −→ Span(FT ) × Span(FT ) −−−→ Span(FT ) ,
BV

Its internal hom is denoted Alg⊗ (P ); its underlying T -∞-category is denoted Alg (P ), and it has values
O O

Alg (P )V ≃ AlgResT O (ResTV P ).

O V

We verified several properties in [Ste25]; for instance, Alg (C) is an I-symmetric monoidal ∞-category when C
P
is, functorially for I-symmetric monoidal maps in C ⊗ and T -operad maps in P ⊗ . We interpret O ⊗ P -algebras
as homotopy-coherently interchanging pairs of O-algebra and P -algebra structures via the following.
Recollection 1.32. Suppose C ⊗ is an I-symmetric monoidal 1-category and O ⊗ , P ⊗ are one-color T -operads.
BV
We saw in [Ste25] that an O ⊗ ⊗ P ⊗ -algebra structure on a T -object X ∈ Γ T C is equivalently viewed as a
pair of O-algebra and P -algebra structures subject to the interchange relation that, for all µS ∈ O(S) and
µT ∈ P (T ), the following diagram commutes.

S T (ResVW µS )
⊗ ResV
UT ⊗ ResV
WS
XV⊗S×T XV⊗T
N N
XV ≃ ≃ XV
U W
µT
(ResVU µT )
µS
XV⊗S XV
BV
A morphism of O ⊗ ⊗ P ⊗ -algebras is simply a morphism of T -objects which is simultaneously an O-algebra
map and a P -algebra map. ◁
The following proposition exhibited a key role played by triv⊗
T.

Proposition 1.33 ([Ste25, Thm D.(3)]). triv⊗

BV

T is the ⊗ -unit; hence there is an equivalence of T -operads

Alg (O) ≃ O ⊗
trivT
BV
We also saw that ⊗ is compatible with the Mode (i.e. Day coonvolution or box product) structure.

Proposition 1.34 ([Ste25, Thm D.(7)]). The T -symmetric monoidal envelope intertwines the mode symmetric
monoidal structure on Cat⊗ T with Boardman-Vogt tensor products, i.e.
 BV     
Env O ⊗ ⊗ P ⊗ ≃ Env O ⊗ ⊛ Env P ⊗ .
 
Furthermore, Env triv⊗
T is the ⊛-unit.

To use all of these results, for the remainder of Sections 1 and 2 we will make the following assumption.
In Corollary 3.1, we will establish Assumption (b) in full generality, so the results of Sections 1 and 2 will
apply for arbitrary T after that point.
Assumption 1.35. We assume one of the following things is true.
(a) T has a terminal object.
BV BV
(b) ⊗ : Op × Op → Op is a T -bifunctor whose restriction OpV × OpV → OpV is ⊗ over T/V . ◁
T T T

J
1.3. Restriction and arity-borelification. We now expand on ResVU and BorI .

1.3.1. Operadic restriction and (co)induction. Recall from [Ste25, § 2.3] that the underlying T -symmetric
sequence forms a T -functor sseq : Opred → FunT (ΣT , S T ); in particular, restriction of V -operads lies over
T
restriction of V -symmetric sequences. This upgrades Theorem 1.26 to an adjunction of T -∞-categories.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 19

⊗ ⊗ ⊗
Proposition 1.36. ResW
V NI∞ ≃ N ; more generally, A ⊣ N(−)∞ lifts to a T -adjunction
ResW
V I∞

Op wIndexT

⊣
T
⊗
N(−)∞

Proof. Restriction compatiblility of the underlying symmetric sequence implies that ResW W
V AO = A ResV O,
⊗
lifting A to a T -functor Op → wIndexT whose V -value is A : OpV → wIndexV . The right adjoints N(−)∞
T
⊗
uniquely lift to a right T -adjoint to N(−)∞ by [HA, Prop 7.3.2.6], completing the proposition. □

Since A is a T -left adjoint, it is compatible with T -colimits. Applying this for indexed coproducts, we
immediately acquire the following convenient properties of A.
Corollary 1.37. If O, P are T -operads, then we have
A(O ⊔ P ) = AO ∨ AP .
If Q is a V -operad, then we have
AIndW W
V Q = IndV AQ.

We may use an analogous argument to that of [BHS22, Lem 4.1.13] to show that Op strongly
T
admits T -limits; since the fully faithful T -functor Op → Catint−cocart
possesses pointwise left adjoints
/ Span(FT )
T
(given by LFbrs ), it possesses a T -left adjoint; in particular, we may compute T -limits of T -operads in
Catint−cocart int−cocart
/ Span(FT ) . Then, an analogous argument using [BHS22, Prop 2.3.7] constructs T -limits in Cat/ Span(FT )
in FunT (Span(FT ), CatT )/FT −⊔ , which strongly admits T -limits, as its a slice T -∞-category of a functor
T
T -∞-category into a T -∞-category which strongly admits T -limits. In particular, this constructs a right
W
adjoint to ResW V : OpW → OpV , which we call CoIndV .

Proposition 1.38. If O ⊗ is a V -d-operad, then CoIndW ⊗

V O is a W -d-operad.

Proof. This follows simply by taking right adjoints within the following diagram
ResW
V
OpW OpV
(8)
ResW
V
OpW ,d OpV ,d
□
Corollary 1.39. There exist equivalences
sseq CoIndW ⊗ W ⊗
V O ≃ CoIndV sseq O ;
ACoIndW W
V O = CoIndV AO.

Proof. The first statement equivalence by noting that Fr ResW W∗

V = ιV Fr and taking right adjoints. The second
follows by taking right adjoints of Eq. (8) in the case d = 0. □
We care about CoIndW ⊗
V O because it is a structure borne by norms of algebras as follows.
Construction 1.40. Let I be a W -weak indexing category containing the map V → W , let P ⊗ → CoIndW
V O
⊗
⊗
be a functor of one-object I-operads, and let C be a I-symmetric monoidal ∞-category. Then, the adjunct
map ϕ : ResW ⊗
V P → O participates in a commutative diagram of symmetric monoidal functors

ϕ∗ NVW
AlgO (ResW
V C) AlgResW P (ResW
V C) AlgP (C)
V

UV UV UW
NVW
CV CV CW

Intuitively, CoIndW ⊗ W
V O bears the universal natural structure on NV X for all X ∈ AlgO (C). ◁
20 NATALIE STEWART

1.3.2. Color and arity Borelification. Let F ⊂ T be a T -family. There is a terminal F -colored weak indexing
category FF ; we refer to FF -Borelification as F -Borelification and write BorTF B BorTFF . Note that
OpF ≃ OpIF .

Let triv⊗ ⊗ triv

F B triv(∗F ) ; this is the weak N∞ -operad for the initial F -colored weak indexing category IF =
≃ T triv
FF = EF I , and in particular, there is an equivalence
(9) triv⊗ T ⊗
F ≃ EF trivF .
BV
This is our first nontrivial example of a ⊗ -idempotent T -operad.
Proposition 1.41 (Color-borelification). Given F ∈ FamT a T -family, there is a natural equivalence
AlgtrivF (O) ≃ Γ F O;
hence there is a natural equivalence
triv⊗ ⊗ T T ⊗
BV

F ⊗ O ≃ EF BorF O .

Proof. The first statement follows by using Eq. (9) and Proposition 1.33 to construct equivalences
AlgtrivF (O) ≃ AlgtrivF (BorTF (O)) ≃ Γ F O.
The second statement then follows by Yoneda’s lemma, noting that
AlgtrivF ⊗O (P ) ≃ AlgtrivF Alg⊗ (P )
O
≃ Γ F AlgO (P )
≃ AlgBorT O (BorTF P )
F
≃ AlgE T BorT O (P ). □
F F

Given O ⊗ ∈ OpT , we define the color family c(O) B c(AO) = {V | OV , ∅}.

n o
Remark 1.42. Proposition 1.41 exhibits ImEFT = O ⊗ ∈ OpT | c(O) ⊂ F as a ⊗-ideal, i.e. if c(O) ⊂ F , and P ⊗
 BV 
is arbitrary, then c O ⊗ P ⊂ F . ◁
BV
This is important in part because it reduced ⊗ computations to the at-least one color case.
Observation 1.43. There is a string of natural equivalences
O ⊗ ⊗ P ⊗ ≃ O ⊗ ⊗ triv⊗ ⊗ ⊗
BV BV BV BV

cO ⊗ trivcP ⊗ P ,
≃ O ⊗ ⊗ triv⊗ ⊗
BV BV

cO∩cP ⊗ P ,
≃ O ⊗ ⊗ triv⊗ ⊗ ⊗
BV BV BV

cO∩cP ⊗ trivcO∩cP ⊗ P ,
  BV  
T
≃ EcO∩cP BorTcO∩cP O ⊗ ⊗ EcO∩cP
T
BorTcO∩cP P ⊗ ,
   BV  
T
≃ EcO∩cP BorTcO∩cP O ⊗ ⊗ BorTcO∩cP P ⊗ .

Moreover, the cO ∩ cP -operads BorTcO∩cP (O ⊗ ) and BorTcO∩cP (P ⊗ ) both have at least one color. ◁
Having done this, we may compute arity-supports of arbitrary tensor products of T -operads.
Proposition 1.44. Suppose O ⊗ , P ⊗ are T -operads. Then,
 BV 
T
A O ⊗ P = EcO∩cP BorTcO∩cP (AO ∨ AP ) .
Proof. By Observation 1.43, we have equivalences
     BV  
A O ⊗ ⊗ P ⊗ ≃ EcO∩cP
T
A BorTcO∩cP O ⊗ ⊗ BorTcO∩cP P ⊗ ,
so it suffices to prove the proposition in the case that O ⊗ and P ⊗ have at least one color.
In this case, first note that there exist maps
O ⊗ ⊗ triv⊗ ⊗ ⊗ ⊗ ⊗
T , trivT ⊗ P → O ⊗ P ,
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 21

and applying A together with the universal property for joins yields an inequality
AO ∨ AP ≤ A(O ∨ P ).
To provide the inequality in the other direction, by Proposition 1.36, in the case of Assumption (b) we may
pass to a restriction and assume that T has a terminal object; in this case, there exists a composite map
⊗ ⊗ ⊗ ⊗ ⊗
O ⊗ ⊗ P ⊗ → NAO∞ ⊗ NAP ∞ → NAO∨AP ∞ ⊗ NAO∨AP ∞ → NAO∨AP ∞ ,

whose last map is presented by the bifunctor

∧
SpanI (FT ) × SpanI (FT ) SpanI (FT )

∧
Span(FT ) × Span(FT ) Span(FT );

here, the top map is defined canonically by the fact that weak indexing categories I ⊂ FT are closed under
cartesian products [Ste24]. Applying A to this map yields A(O ∨ P ) ≤ AO ∨ AP , as desired. □

We immediately acquire the following corollary.

Corollary 1.45. OpI ⊂ OpT is closed under binary tensor products; if I has one color, then triv⊗
T ∈ OpI .

1.4. (Co)cartesian I-symmetric monoidal ∞-categories. Fix I an almost-unital weak indexing system. In this
section, we characterize cartesian and cocartesian I-symmetric monoidal ∞-categories, in part as examples of
interest and in part as universal construction.
We defer the minutiae of these to Appendix A, where we construct ∞-categories C I−× , C I−⊔ over TotFT ,∗ ,
verifying that C I−× is an I-symmetric monoidal ∞-category precisely when C has I-indexed products, that
C I−⊔ is always an I-operad, and that that C I−⊔ is an I-symmetric monoidal ∞-category precisely when C
has I-indexed coproducts. Most of this abuts to unenlightening technicalities about parameterized higher
category theory, which we defer to Appendix A, summarizing the outcomes as they become relevant.

1.4.1. (Co)cartesian rigidity. Denote by CatI−⊔

I , CatI
I−×
⊂ CatT the replete subcategories with objects given
by T -∞-categories attaining I-indexed coproducts (resp. products) and with morphisms given by T -functors
which preserve I-indexed coproducts (products). In Appendix A, we prove the following.
Theorem A’. There are fully faithful embeddings (−)I−⊔ , (−)I−× making the following commute:
(−)I−⊔ (−)I−×
CatI−⊔
I Cat⊗
I CatI−×
I

U
U U
CatT

The image of (−)I−⊔ is spanned by the I-symmetric monoidal ∞-categories whose I-admissible indexed tensor
functors ⊗S : CS → CV are left adjoint to the indexed diagonal ∆S : CV → CS (i.e. whose indexed tensor
products are are indexed coproducts), and the image of (−)I−× is spanned by those whose I-admissible indexed
tensor functors ⊗S are right adjoint to ∆S .
We call I-symmetric monoidal ∞-categories of the form C I−⊔ cocartesian, and C I−× cartesian.
Philosophical remark 1.46. In higher category theory, a fundamental rigidity result is that of adjoints; by
[HTT, Prop 5.2.6.2] the full subcategory of Fun(D, C)op spanned by functors right adoint to a fixed functor
L : C → D is contractible. Moreover, adjointness itself as a property requires only finitely data to test
per-object, separately (c.f. [HTT, Prop 5.2.2.9, 5.2.2.12]).
It is this rigidity which we are leveraging in Theorem A’: in essence, the coherent data witnessing
I-symmetric monoidality of a functor C I−× → DI−× is constructed (up to contractible ambiguity) by the
NS
property that the underlying T -functor C → D is compatible with the adjunctions ∆S ⊣ . ◁
Many similar definitions have been made in the literature. Luckily, they all agree.
22 NATALIE STEWART

` Q
Remark 1.47. [NS22] constructed a pair of structures C , C which, after unwinding definitions, satisfy the
conditions
` of Theorem A’ inQthe case I = T . In particular, there are unique I-symmetric monoidal equivalences
BorTI C ≃ C I−⊔ and BorTI C ≃ C I−× lying over the identity whenever C admits finite indexed (co)products.
Moreover, [CHLL24b] introduced another structure, specifically a cartesian I-symmetric monoidal
structure on FT , and conjectured it to be equivalent to Nardin-Shah’s construction; this conjecture was
recently verified in [CLR25] by verifying universality of the unfurling construction used in Nardin-Shah’s
construction. We acquire an independent proof of this result without serious (∞, 2)-category theory: Cnossen-
Haugseng-Lenz-Linsken’s construction satisfiesQ the condition of Theorem A’, so there is a unique pair of
I-symmetric monoidal equivalences BorTI FT ≃ FI−× T ≃ FT ,× lying over the identity.
Moreover, after drafts of this article were made public, [LLP25, Def` 2.15] constructed another structure,
which specializes to a cocartesian I-symmetric monoidal structure C I− ; by [LLP25, Lem 2.16(i)], in the case
`
that C has I-indexed coproducts,
` C I− satisfies the conditions of Theorem A’, so there is a unique I-symmetric
monoidal equivalence C I− ≃ C I−⊔ lying over the identity. ◁
We prove the following in Appendix A as a precursor to Theorem A’, though it also follows from it.
 vop
Proposition 1.48. There is a unique equivalence C I−× ≃ (C vop )I−⊔ lying over the identity.

Before characterizing the algebras in C I−⊔ and C I−× , we point out that they are often presentable.
Proposition 1.49. Suppose C is a presentable ∞-category with I-indexed products and coproducts.
(1) CoeffT C I−⊔ is presentably I-symmetric monoidal.
(2) If finite products in C commute with colimits separately in each variable (i.e. it is Cartesian closed),
then CoeffT C I−× is presentably I-symmetric monoidal,
Proof. It follows from Hilman’s characterization of parameterized presentability [Hil24, Thm 6.1.2] that
CoeffT is presentable, so we’re tasked with proving that the T -symmetric monoidal structures are distributive.
The first case is just commutativity of colimits with colimits, and the second is [NS22, Prop 3.2.5]. □
1.4.2. O-monoids. We will identify algebras in C I−× with the following.
Definition 1.50. Fix O ⊗ an I-operad and C a T -∞-category. Then, an O-monoid in C is a T -functor
M : TotT O ⊗ → C satisfying the condition that, for each orbit V ∈ T , each finite V -set S ∈ FV , and each
S-tuple X = (XU ) ∈ OS , the canonical maps M(X) → CoIndVU M(XU ) realize M(X) as the indexed product
S
Y
M(X) ≃ M(XU ). ◁
U
Indeed, we prove the following equivariant lift of [HA, Prop 2.4.2.5] as Proposition A.16.
Proposition 1.51. Given O ⊗ an I-operad and C a T -∞-category with I-indexed products, the forgetful functor
   
AlgO C I−× −→ FunT TotT O ⊗ , C
is fully faithful with image spanned by the O-monoids.
Corollary 1.52. Given O ⊗ an I-operad and D an ∞-category with finite products, the forgetful functor
   
AlgO CoeffG (D)I−× −→ Fun TotTotT O ⊗ , D
is fully faithful with image spanned by SegTotTotT O⊗ (D).
Proof. After Proposition 1.51, it suffices to characterize the image of O-monoids under the equivalence
Fun(TotTotT O ⊗ , D) ≃ FunT (TotT O ⊗ , CoeffG (D)).
`
By [Nar17, Ex 1.17], given a finite V -set S ∈ FV and writing TotS ≃ U ∈Orb(S) T/U for the total ∞-category of
the associated V -category, the above identification turns S-indexed products into right Kan extensions:
QS
FunT (S, CoeffT (D)) CoeffT (D)
≃
≃

RKE
Fun (TotS, D) Fun(T op , D)
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 23

Thus the image of MonO (CoeffT D) is those functors TotTotT O ⊗ → D whose image of an object ((XU ), S) ∈
TotTotT O ⊗ is right Kan extended along elementary maps, which is exactly the relevant Segal condition. □
Corollary 1.53. Given O ⊗ an I-operad and D an ∞-category with finite products, the forgetful functor
AlgO (CoeffG (D)I−× ) −→ Fun(TotO ⊗ , D)
is fully faithful with image spanned by SegTotO⊗ (D).
Proof. Apply Corollary 1.52 and the equivalence SegTotTotT O⊗ (C) ≃ SegTotO⊗ (C) constructed in [Ste25, § A]. □
We finally identify Perspectives (i) to (iii) from the introduction.
Corollary 1.54 (“CMon = CAlg”). There is a canonical equivalence CMonI (C) ≃ CAlg (C I−× ) over C.
I
Proof. Our proof is similar to that of [Nar16, Thm 6.5]; there is a pullback square over C

CMonI (C) CAlgI (C I−× ) ≃ FunI−× (Span(FT ), C)

⌟
 
FunT (C op , CMonI (S T )) FunT C op , CAlg (C I−× ) ≃ FunT (C op , FunI−× (Span(FT ), S T ))
I

so it suffices to prove this in the case C = S T . There, we simply compose equivalences as follows
1.8 1.53
CMonI (S T ) CMonI (S) CAlgI (S I−×
T ) □

Remark 1.55. As with much of the rest of this subsection, Corollary 1.54 possesses an alternative strategy
where both are shown to furnish the I-semiadditive closure, the latter using [CLL24, Thm B]. The above
argument was chosen for brevity, as its requisite parts are also needed elsewhere. ◁
Remark 1.56. In the case C ≃ S G , and I is an indexing category, the analogous result was recently proved in
[Mar24] for a Dwyer-Kan localization of algebras over the corresponding graph G-operads. To the knowledge
of the author, this is one of the first concrete higher-categorical indications that the genuine operadic nerve
of [Bon19] may induce equivalences between ∞-categories of algebras. ◁
1.4.3. F -unitality. We now study I-operadic unitality, beginning with the following definition.
Definition 1.57. We say that an I-operad O ⊗ is unital if O(∅V ) = ∗ for all V ∈ υ(I), and reduced if also
O(∗V ) = ∗ for all V ∈ c(I). More generally if F ⊂ υ(I) is a family, we say that O ⊗ is F -unital if O(∅V ) ≃ ∗ for
all V ∈ F and F -reduced if also O(∗V ) = ∗ for all V ∈ F ; equivalently, O ⊗ is F -unital (resp. F -reduced) if
and only if BorII∩FF O ⊗ is unital (reduced). ◁
An under-appreciated case of unitality is the (equivariantly) symmetric monoidal case.
Observation 1.58. If C ⊗ is an I-symmetric monoidal ∞-category with unit υ(I)-object 1• and X ∈ CV , then
the Segal conditon for multimorphisms constructs an equivalence
MapO⊗ (∅V , X) ≃ MapCV (1V , X);
hence C ⊗ is F -unital if and only if 1• ∈ Γ F C is initial. In particular, if BorII∩FF C ⊗ is cartesian, then C ⊗ is
F -unital if and only if U C is F -pointed. ◁
We can identify this via an algebraic mapping-in property as follows.
Lemma 1.59. If C ⊗ is an I-symmetric monoidal ∞-category, then C ⊗ is F -unital if and only if the forgetful
T -functor U : Alg (C) → C is an equivalence.
E0,F

Proof. The forward implication follows from the computation Lemma A.11 in the case IF0 , so assume U is
1 /
an equivalence. Then, for all V ∈ F , CVV ≃ AlgE0,F (C)V → C is an equivalence, so 1V ∈ CV is initial. Thus
Observation 1.58 implies the lemma. □
We can replace E⊗
0,F with an arbitrary F -unital I-operad, retaining the above property.
Lemma 1.60 (Incomplete [NS22, Thm 5.2.11]). If O ⊗ is an F -unital I-operad and C ⊗ is an I-symmetric
monoidal ∞-category, then Alg (C) is F -unital.
O
24 NATALIE STEWART

Proof. Using the same trick as Lemma 1.59, we prove this when F = υ(I) = T . Then, in light of Observa-
tion 1.58, this is simply [NS22, § 5.2.11]. □

1.4.4. O-comonoids, indexed semiadditivity. We prove the following fundamental fact as Lemma A.11.
Proposition 1.61. Let C be a T -∞-category and O ⊗ a unital I-operad. The forgetful functor is an equivalence
  ∼
Alg C I−⊔ −−−−−→ FunT (U O, C).
O
We use this to construct another recognition result for cocartesian I-symmetric monoidal ∞-categories.
Construction 1.62. Let C ⊗ be an I-symmetric monoidal ∞-category satisfying the property that 1• ∈ Γ υ(I) C is
initial and let (XU ) ∈ CS be an S-tuple for some S ∈ FI . The ⊗-Wirthmüller map for (XU ) is the map
S
a S
O
WS,(XU ) : XU −→ XU
U U
classified by the summand maps
WS,(XU ),U ′ NS
XW ResVU W XW .
≃

≃
NResVU S−W (id;!) NResVU S−W o(U ′ )
XW ⊗ U′ 1U ′ XW ⊗ U′ ResU ′ Xo(U ′ )

where o(U ′ ) ∈ Orb(S) is the orbit whose restriction contains U ′ . 1U exists and is initial by almost-unitality.
Dually, if C ⊗ is an I-symmetric monoidal ∞-category such that 1• ∈ Γ υ(I) C the ⊗-co-Wirthmüller map
co
NS QS ⊗,vop .
WS,(X ):
U U XU → U XU is the Wirtmüller map for (XU )S in the fiberwise opposite C ◁

Lemma 1.63. Let C ⊗ be an I-symmetric monodial ∞-category.

(1) C ⊗ is cocartesian if and only if 1• ∈ Γ υ(I) C is initial and WS,(XU ) is an equivalence for all S ∈ FI and
(XU ) ∈ CS ;
co
(2) C ⊗ is cartesian if and only if 1• ∈ Γ υ(I) C is terminal and WS,(X is an equivalence for all S ∈ FI and
U)
(XU ) ∈ CS .
Proof. (1) is Observation A.24, and part (2) follows directly from part (1) under Proposition 1.48. □

Observation 1.64. When C ⊗ is cartesian, the assumption that 1• is initial is precisely the assumption that C
is υ(I)-pointed; moreover, unwinding definitions, WS,(XU ) matches the Wirthmüller map of Definition 1.4. ◁
Finally, we find the indexed semiadditivity of [CLL24; Nar16] within equivariant higher algebra.
Corollary 1.65 (Equivariant [GGN15, Prop 2.3]). Suppose C is a T -∞-category with I-indexed products. Then,
the following conditions are equivalent.
(a) C is I-semiadditive.
(b) There exists an I-symmetric monoidal equivalence C I−× ≃ C I−⊔ lying over the identity.
(c) The forgetful T -functor CMonI (C) → C is an equivalence.
Proof. We have proved this redundantly. (a) ⇐⇒ (b) is Lemma 1.63 and Observation 1.64. (b) =⇒ (c) is
Corollary 1.54 and Proposition 1.61. (c) ⇐⇒ (a) is [CLL24, Cor 7.8]. □

1.4.5. Pointwise indexed tensor products of O-monoids. Last, we characterize Alg⊗ (C I−× ) categorically.
O

Lemma 1.66. Fix F : D⊗ → C⊗

an I-symmetric monoidal functor.
(1) If F is I-coproduct preserving then there is a homotopy FWS,(XU ) ∼ WS,F(XU ) .
co co
(2) If F is I-product preserving then there is a homotopy FWS,(X ∼ WS,F(X .
U) U)
(3) If F is I-coproduct-preserving and conservative and C is cocartesian then D⊗ is cocartesian.
⊗

(4) If F is I-product preserving and conservative and C ⊗ is cartesian then D⊗ is cartesian.

(5) If F is a fiberwise-monadic right T -adjoint and C ⊗ is cartesian then D⊗ is cartesian.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 25

Proof. For (1), since F is I-coproduct preserving, we’re tasked with constructing a homotopy making the
following diagram commute
 NResV S−U 
(id;!)
 NResV S−U 
F XU ⊗ W U 1W F XU ⊗ W U XW

≃
NResVU S−U (id;!) NResVU S−U
FXU ⊗ W 1U FXU ⊗ W FXW

In fact, there is a contractible space of such choices. (2) follows by applying fiberwise-opposites.
For (3), applying (1) and Lemma 1.63 shows that FWS,(XU ) is an equivalence for all S ∈ FI and (XU ) ∈ DS ,
so conservativity implies that WS,(XU ) is an equivalence and Lemma 1.63 concludes that D⊗ is cocartesian.
(4) follows similarly from (2). (5) is directly implied by (4), since fiberwise monadicity implies conservativity
and right T -adjoints are I-product preserving. □
 
Applying (5) of Lemma 1.66 to U : Alg⊗ C I−× → C I−× immediately yields the following.
O
 
Corollary 1.67. If C has I-indexed products, then Alg⊗ C I−× is a cartesian I-symmetric monoidal ∞-category.
O

2. I-commutative algebras
Philosophical remark 2.1. On one hand, it follows from Proposition 1.24 that I-operads are determined con-
servatively by their theories of algebras in I-symmetric monoidal categories; indeed, it suffices to characterize
their algebras in the universal case S I−×
T .
On the other hand, the right adjoint Cat⊗ I → OpI is full on cores, since automorphisms in the slice
category Cat/ SpanI (FT ) automatically preserve cocartesian morphisms. Hence the associated map of spaces

Cat⊗,≃
I Op≃
I Fun(OpI , Cat)≃
∈

C⊗ Alg(−) (C)

is a summand inclusion. That is, an I-symmetric monoidal ∞-category is determined (functorially on

equivalences) by its categories of O-algebras for each I-operad O. ◁
Following along these lines and using Proposition 1.51, we will generally characterize algebraic theories
in arbitrary settings by reducing to the universal case of S TT −× , which we study using category theoretic means.
Indeed, in Sections 2.1 and 2.2 we use this to bootstrap I-semiadditivity of CMonI (C) to I-cocartesianness of
CAlg⊗ (C) for C ⊗ an arbitrary I-operad. Using work from Appendix A, we use this to conclude generalizations
I
of Theorem C and Corollary D, answering Questions (I) and (II).
We take this to its logical extreme in Section 2.2, using this to completely characterize the smashing
localizations associated with ⊗-idempotent weak N∞ -operads. As promised in the introduction, we use this
classification to prove a generalization of Theorem F, answering Question (III). Following this, in Section 2.3
⊗ BV ⊗
we show that our results are sharp; if I is not almost essentially unital, then NI∞ ⊗ NI∞ fails to be connected,
⊗ BV
so NI∞ is (abstractly) idempotent under ⊗ if and only if I is almost essentially unital.
2.1. Indexed tensor products of I-commutative algebras. Fix I an almost-unital weak indexing system. In
Proposition 1.61, we showed that every object in a cocartesian I-symmetric monoidal structure bears a
canonical I-commutative algebra algebra structure, i.e. CAlg (C) → C is an equivalence. In this subsection,
I
we demonstrate the converse, i.e. we show the following.
Theorem 2.2 (Indexed tensor products of I-commutative algebras). The following are equivalent for C ⊗ ∈ Cat⊗
I.
⊗
(a) C is cocartesian.
(b) For all unital I-operads O ⊗ , the forgetful functor AlgO (C) → FunT (U O, C) is an equivalence.
(c) The forgetful T -functor CAlg (C) → C is an equivalence.
I
∼
(d) There exists an I-symmetric monoidal ∞-category D⊗ and an equivalence CAlg⊗ (D) −−−→ C ⊗ .
I
26 NATALIE STEWART

The implications (a) =⇒ (b), (c) are simply Proposition 1.61. For the implication (b) =⇒ (a), note
that Lemma 1.60 states that C ⊗ is unital; hence Yoneda’s lemma applied to Opuni
I constructs an I-operad
equivalence C ⊗ ≃ C I−⊔ , which is an I-symmetric monoidal equivalence by Philosophical remark 2.1. The
implication (c) =⇒ (d) follows by neglect of assumptions. To summarize, we’ve arrived at the implications
(d)
(10)
(b) (a) (c)
Our workhorse lemma for closing the gap is the following.
Lemma 2.3. The following are equivalent for P ⊗ ∈ OpT :
(e) The T -∞-category Alg (S T ) is I-semiadditive.
P
(f ) For all O ⊗ ∈ Opuni
I , the forgetful functor
AlgO⊗P (S T ) ≃ AlgO Alg⊗ (S T ) −→ Algtriv(U O) Alg⊗ (S T ) ≃ FunT (U O, Alg (S T ))
P P P
is an equivalence.
⊗ BV ⊗
(g) For all O ⊗ ∈ Opuni ⊗ BV ⊗
I , the map triv (O) ⊗ P → O ⊗ P is an equivalence.
(h) For all O ⊗ ∈ Opuni
I and Q⊗ ∈ OpI , the forgetful T -operad map
Alg⊗ (Q) ≃ AlgO Alg⊗ (Q) −→ Alg⊗ Alg⊗ (Q)
O⊗P P triv(U O) P
is an equivalence
Proof. Since Corollary 1.67 shows that BorTI Alg⊗ (S T ) is cartesian, Corollary 1.65 identifies the bi-implication
O
(e) ⇐⇒ (f) with (a) ⇐⇒ (b) applied to BorTI Alg⊗ (S T ). (f) =⇒ (g) follows from Proposition 1.24, and the
P
implications (g) =⇒ (h) =⇒ (f) are obvious. □
Proof of Theorem 2.2. After the implications illustrated in Eq. (10), it suffices to show for all D⊗ ∈ Cat⊗I that
CAlg (D) satisfies (b), i.e. (d) =⇒ (b); by Lemma 2.3, it suffices to prove that CAlg (S T ) is I-semiadditive.
I I
But in fact, Corollary 1.54 constructs an equivalence CAlg (S T ) ≃ CMonI (S T ) and the latter is I-semiadditive
I
by Cnossen-Lenz-Linsken’s result, Theorem 1.9. □
Rephrasing things somewhat, we’ve arrived at the following theorem.
Theorem C’. Let O ⊗ be an almost essentially reduced T -operad. Then, the following properties are equivalenent.
(a) The T -∞-category Alg S T is AO-semiadditive.
O
⊗
(b) The unique map O ⊗ → NAO∞ is an equivalence.
Furthermore, for any almost essentially unital weak indexing system I and I-symmetric monoidal ∞-category
C ⊗ , the I-symmetric monoidal ∞-category CAlg⊗ C is cocartesian.
I
Proof. By Lemma 1.63, Corollary 1.67, and Theorem 2.2, Condition (a) is equivalent to the forgetful T -functor
CAlg (S T ) ≃ Alg CAlg⊗ (S T ) ≃ CAlg Alg⊗ (S T ) Alg (S T )
AO O AO AO O O

being an equivalence, which is equivalent to Condition (b) by Proposition 1.24. The remaining statement
follows immediately from the implication (d) =⇒ (a) of Theorem 2.2. □
⊗
∈ Opred
BV
This implies that NI∞ I is ⊗ -absorptive.
⊗ ⊗ BV
Corollary 2.4. Let O ⊗ be an almost-reduced I-operad. The map F : NI∞ → NI∞ ⊗ O ⊗ is an equivalence.
Proof. By Theorem 2.2, the forgetful map
F ∗ : AlgO⊗NI∞ (S T ) ≃ AlgO Alg⊗ (S T ) → Alg (S T )
NI∞ NI∞
is an equivalence. The statement then follows from Proposition 1.24. □
Remark 2.5. At this point, we may answer Question (I) of the introduction; if is almost-essentially unital O⊗
⊗ ⊗ BV ⊗
and NI∞ → NI∞ ⊗ O is an equivalence, then Proposition 1.44 implies that O ⊗ is an almost-reduced I-operad;
in particular, the assumptions of Corollary D are necessary and sufficient. ◁
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 27

⊗
2.2. The smashing localization for NI∞ and Blumberg-Hill’s conjecture. In view of Corollary 2.4, when I is
an almost-unital weak indexing category the unique map triv⊗ ⊗
T → NI∞ induces an equivalence
⊗ ⊗ ∼
⊗ triv⊗ ⊗ ⊗
BV BV
NI∞ ≃ NI∞ T −
−−→ NI∞ ⊗ NI∞ ,
⊗
i.e. it uniquely witnesses NI∞ as an idempotent object in the sense of [HA, Def 4.8.2.1]. To conclude
⊗
Theorem F, we will characterize the smashing localization classified by NI∞ -modules.4
⊗ ⊗ BV
2.2.1. The smashing localization classified by NI∞ . (−) ⊗ NI∞ classifies algebraic Wirthmüller isomorphisms.
Theorem 2.6. Let I be an almost essentially unital weak indexing system. Then, a T -operad O ⊗ possesses an
BV
⊗
equivalence P ⊗ ⊗ NI∞ ≃ P ⊗ if and only if the following conditions are satisfied:
(a) c(P ) = c(I), and
(b) Alg (S T ) is I-semiadditive.
P
By the arity support computation of Proposition 1.44, Theorem 2.6 is equivalent to the following.
Proposition 2.7. Let I be an almost-unital weak indexing system. Then, an at-least one color T -operad P ⊗
BV
⊗
satisfies P ⊗ ⊗ NI∞ ≃ P ⊗ if and only if Alg (S T ) is I-semiadditive.
P
Proof. Just as in the proof of Lemma 2.3, note that Theorem 2.2 implies that the conditions of Lemma 2.3
are equivalent to the additional condition
(j) The forgetful T -functor Alg (S T ) ≃ CAlg Alg⊗ (S T ) → Alg (S T ) is an equivalence.
NI∞ ⊗P I P P
But this is equivalent to the desired equivalence by Proposition 1.24 □
2.2.2. The proof of Blumberg-Hill’s conjecture. We start by answering Question (III).
⊗ ⊗ ⊗ BV
Proposition 2.8. When I and J are almost-unital, there is an equivalence NI∞ ⊗ NJ∞ ≃ NI∨J∞ .
⊗ ⊗ BV
Proof. By [CSY20, Prop 5.1.8], NI∞ ⊗ NJ∞ ∈ OpT ≃ OpT ,triv⊗ / is an idempotent object classifying the
T
⊗ ⊗
conjunction of the properties which are classified by NI∞ and NJ∞ ; that is, T -operad O ⊗ is fixed by
BV
⊗ BV
⊗
(−) ⊗ NI∞ ⊗ NJ∞ if and only if Alg (S T ) is I-semiadditive and J-semiadditive. Proposition 1.6, identifies this
O
⊗ BV
with the property that Alg (S T ) is I ∨ J-semiadditive , i.e. O ⊗ is fixed by (−) ⊗ NI∨J . Thus, we have
O
BV
⊗ ⊗ ⊗ ⊗BV
⊗ BV BV
NI∨J∞ ≃ NI∨J∞ ⊗ NI∞ ⊗ NJ∞ ≃ NI∞ ⊗ NJ∞ . □
We may now conclude the full theorem, which we restate in the atomic orbital case.
⊗
Theorem F’. N(−)∞ : wIndexT → OpT restricts to a fully faithful symmetric monoidal T -right adjoint
A

wIndexaEuni,⊗ OpaEuni,⊗ .
⊣

T T

⊗
N(−)∞

Furthermore, the resulting tensor product of weak N∞ -operads is computed by the Borelified join
⊗ ⊗
≃N⊗
BV
(11) NI∞ ⊗ NJ∞ .
BorTcI∩cJ (I∨J)∞

Hence whenever I, J are almost-unital weak indexing categories and C ⊗ is an I ∨ J-symmetric monoidal
∞-category, there is a canonical equivalence of I ∨ J-symmetric monoidal ∞-categories
(12) CAlg⊗ CAlg⊗ (C) ≃ CAlg (C).
I J I∨J
4 The identification between idempotent objects and smashing localizations is stated in [CSY20; HA] under the unnecessary
specification that the constructions live in a given symmetric monoidal ∞-category, as this leads to canonical lifts of idempotent
objects to idempotent algebras. In fact, their arguments for the identification only make use of the underlying A2 -structure and the
existence of a braiding A ⊗ B ≃ B ⊗ A, separately for each pair. Even before Section 3.1, (OpT , ⊗ , triv⊗
BV

BV
T ) with the braiding determined
by symmetry of the universal property for ⊗ is certainly such a structure.
28 NATALIE STEWART

Proof of Theorem F’. The T -adjunction is precisely Proposition 1.36, and Eqs. (11) and (12) will follow from
⊗
symmetric monoidality of N(−)∞ and the support computation of Proposition 1.44.
BV
We’re left with proving that almost essentially unital weak N∞ -operads are closed under ⊗ , i.e. the
⊗ BV ⊗ ⊗
unique map ϕ : NI∞ ⊗ NJ∞ → NI∨J∞ is an equivalence. By Observation 1.43, it suffices to prove that
T
BorcI∩cJ (ϕ) is an equivalence, i.e. we may assume that I and J are almost-unital; this is Proposition 2.8. □
⊗
2.2.3. NI∞ classifies I-cocartesianness. We now study a variant of Theorem 2.6, motivated by the following.
Observation 2.9. The computation of [HA, § 2.3.1] and resulting theory may be stated simply: the operad
E⊗ ⊗ ⊗ ⊗
0 is an idempotent object in Op under the unique map triv → E0 , and the corresponding smashing
localization classifies unitality. In particular, a symmetric monoidal ∞-category C ⊗ is a E⊗ 0 -module with
BV
respect to ⊗ if and only if the unit object 1C ∈ C ⊗ is initial, i.e. for all X ∈ C, the unique map X ⊔∅ → X ⊗∅ is
an equivalence; that is, (−) ⊗ E⊗
BV

0 classifies I0 -cocartesianness in the symmetric monoidal case. ◁

We say that a T -operad O ⊗ is I-cocartesian if the identity on U O is adjunct to an equivalence
∼
BorTI O ⊗ −−−−−→ U O I−⊔ . We begin with an I-operadic variant of Theorem 2.2 and Proposition 2.7.
Proposition 2.10. Given I an almost-unital weak indexing category, the following are equivalent for P ⊗ ∈ OpT .
(a) P ⊗ is I-cocartesian.
(b) For all unital I-operads O ⊗ , the forgetful functor AlgO (P ) → FunT (IO, P ) is an equivalence.
(c) The forgetful T -operad map CAlg⊗ (P ) → P ⊗ is an equivalence.
I
∼
(d) There exists a T -operad Q⊗ and an equivalence of T -operads CAlg⊗ (Q) −−−→ P ⊗ .
I
⊗BV
(k) The canonical map P ⊗ → P ⊗ ⊗ NI∞ is an equivalence.
Proof. The proof of the equivalence between Conditions (a) to (d) is identical to Theorem 2.2, so we omit
it. Now, by a standard two-out-of-three argument, P ⊗ is local for the smashing localization associated
⊗ BV
⊗
with NI∞ if and only if pullback along the localization map O ⊗ → O ⊗ ⊗ NI∞ induces an equivalence
⊗ ≃ ≃ ∼ ≃ ⊗
AlgO CAlg (P ) ≃ AlgO⊗NI∞ (P ) −−−−→ AlgO (P ) for all O ∈ OpT , which is equivalent to the condition
I
that CAlg⊗ ⊗
I (P ) → P is an equivalence by Yoneda’s lemma. □

Applying Proposition 1.44 to Proposition 2.10 yields a variant of Theorem 2.6, answering Question (II).
Proposition 2.11. Let I be an almost essentially unital weak indexing category and O ⊗ a T -operad. Then, O ⊗
⊗
admits an (essentially unique) NI∞ -module structure if and only if the following conditions hold:
(a) c(O) = c(I), and
(b) O ⊗ is I-cocartesian.
Remark 2.12. If J ⊂ I and O ⊗ is I-cocartesian, then O ⊗ is J-cocartesian. In particular, applying this to
I ∩ E⊗ ⊗
0,υ(I) ⊂ I shows when I is unital that the conditions of Proposition 2.11 implies that O is unital. ◁
Given a related pair of weak indexing categories I ⊂ J, let OpI−cocart
J ⊂ OpJ be the full subcategory of
I−cocart BV
I-cocartesian J-operads. We find that OpJ is absorptive under ⊗ and the internal hom.

Corollary 2.13. Suppose O ⊗ , P ⊗ are at-least-one-color J-operads such that either O ⊗ or P ⊗ are I-cocartesian.
Then, O ⊗ ⊗ P ⊗ and Alg⊗ (P ) are I-cocartesian.
BV

O
BV
Proof. I-cocartesianness of O ⊗ ⊗ P ⊗ follows from Proposition 2.11. If P ⊗ is I-cocartesian, then Proposition 2.10
constructs equivalences
CAlgI Alg⊗ (P ) ≃ AlgO CAlg⊗ (P ) ≃ AlgO (P ),
O I
so the result follows from another application of Proposition 2.10. If O ⊗ is I-cocartesian, the result follows
from two more applications of Proposition 2.10. as we acquire equivalences
(13) CAlg Alg⊗ (P ) ≃ Alg (P ) ≃ Alg (P ). □
I O NI∞ ⊗O O

Finally, we vastly generalize the results of [HA, § 2.3.1].

 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 29

Corollary 2.14. Suppose I ⊂ J is a related pair of almost-unital weak indexing categories. Then, OpI−cocart
J ⊂
OpJ is a smashing localization and a cosmashing colocalization:
⊗ BV
NI∞ ⊗ (−)

⊣
OpI−cocart
J OpJ

⊣
CAlg⊗ (−)
I

Proof. Proposition 2.11 exhibits the top adjunction and Theorem 2.2 shows that CAlg⊗ : OpJ → OpJ factors
I
through OpJI−cocart ⊂ OpJ . Moreover, applying Γ T (−)≃ to Eq. (13) yields a natural equivalence
   
MapOp O ⊗ , P ⊗ ≃ MapOpI−cocart O ⊗ , CAlg⊗ (P )
J J I

for all J-operads O⊗, P ⊗ such that O⊗ is I-cocartesian, yielding the bottom adjunction. □

2.2.4. (Co)localization to unital I-operads. The specialization of Theorem 2.2 and Proposition 2.11 to I ∩ I0
is quite useful, so we state it explicitly here.
Corollary 2.15. Given I an almost-unital weak indexing system and O ⊗ ∈ OpI , the following are equivalent:
(a) For all unital I0 -operads P ⊗ , the forgetful I-operad map Alg⊗ (O) → Alg⊗ (O ⊗ ) is an equivalence.
P triv(U P )
(b) The forgetful I-operad map Alg⊗ (O) → O ⊗ is an equivalence.
E0,υ(I)
(c) O ⊗ is unital.
(d) There exists an equivalence O ⊗ ≃ E⊗
BV
⊗
0,υ(I) ⊗ O .
(e) For all I-operads C ⊗ , the forgetful I-operad map Alg⊗ Alg⊗ (C) → Alg⊗ (C) is an equivalence.
O E0,υ(I) O
(f ) For all I-operads C ⊗ , the I-operad Alg⊗ (C) is unital.
O
(g) The T -∞-category MonO (S) is υ(I)-pointed.
In particular, Opuni
I ⊂ OpI is a smashing localization and cosmashing colocalization:

E⊗
BV
0,υ(I)
⊗ (−)
⊣

Opuni
I OpI
⊣

Alg⊗ (−)
E0,υ(I)

The double adjunction is Corollary 2.14 and the following diagram shows how to recover the corollary.
2.11 Yoneda 2.10
(c) (d) (e) (f)
2.10 1.59 2.3 1.58

(a) (b) (g)

obvious

2.2.5. The underlying T -∞-category. We get an immediate corollary from Theorem A’ and Theorem 2.2.
Corollary 2.16. Suppose I is almost-unital. Then, Uuni : Opuni → CatT is left T -adjoint to (−)I−⊔ .
I

Warning 2.17. Corollary 2.16 shows that no nontrivial T -colimit of one-color T -operads has one color; in
particular, no one-color T -operads are the result of a nontrivial induction. ◁
We use this to compute the T -∞-category underlying Boardman-Vogt tensor products.
Proposition 2.18. The underlying category functor U |uni : Opuni
I → CatT sends
 BV 
U O ⊗ ⊗ P ⊗ ≃ U (O ⊗ ) × U (P ⊗ ).
30 NATALIE STEWART

Proof. Corollaries 2.13 and 2.16 yield a string of natural equivalences

  BV    
FunT U O ⊗ ⊗ P ⊗ , C ≃ AlgO⊗P C I−⊔
 
≃ AlgO Alg⊗ C I−⊔
P
 I−⊔
≃ AlgO FunT U (P ⊗ ), C
  
≃ FunT U (O ⊗ ), FunT U (P ⊗ ), C
 
≃ FunT (U (O ⊗ ) × U (P ⊗ ), C ,
so the result follows by Yoneda’s lemma. □

Applying Observation 1.43 and Propositions 1.44 and 2.18, we acquire the following.
Corollary 2.19. The full subcategories Opred a red
⊂ OpaE red BV

T ⊂ OpT T ⊂ OpT are closed under ⊗ .

2.3. Failure of the nonunital equivariant Eckmann-Hilton argument. We say that a T -operad O ⊗ with at
most one color is n-connected if the nonempty structure spaces O(S) are each n-connected. We write the full
subcategory of n-connected T -operads as
Op≤oc ≤oc
T ,≥n+1 ⊂ OpT .

By Proposition 1.24, this is equivalent to the condition that the forgetful functor CAlgAO (C) → AlgO (C) is an
equivalence for all T -symmetric monoidal (n + 1)-categories , which itself is equivalent to the same condition
BV
in the case C ≃ S≤n . We first observe compatibility with ⊗ in the almost-essentially reduced setting.
red
Corollary 2.20. OpaE
BV

T ,≥n+1 is closed under ⊗ .

red
Proof. By Proposition 1.44 and Theorem F’, O ⊗ , P ⊗ ∈ OpaE
T ,≥n+1 participate in a string of natural equivalences

AlgO⊗P (S T ,≤n ) ≃ AlgO Alg (S T ,≤n )

P
≃ CAlgAO CAlg (S T ,≤n )
AP
≃ CAlgAO∨AP (S T ,≤n )
≃ CAlgA(O⊗P ) (S T ,≤n ),
BV
⊗ BV
induced by the unique map O ⊗ ⊗ P ⊗ → NA(O⊗P ⊗ ⊗
) , so Proposition 1.24, implies that O ⊗ P is n-connected.
BV
Corollary 2.19 implies that O ⊗ ⊗ P ⊗ is almost essentially reduced. □

Remark 2.21. The unit object triv⊗ T ∈ OpT is n-connected for all n, so n-connected T -operads are closed
under k-fold tensor products for all k ∈ N. ◁
The example triv⊗
BV
⊗ ⊗
T ⊗ O ≃ O demonstrates that this is the best we can say without further assumptions
on the T -operads in question; the author hopes to return to this question in forthcoming work, constructing
analogues to [SY19]. For the time being, we demonstrate that Corollary 2.20 dramatically fails without the
almost essentially unital assumption, exhibiting a failure of the nonunital Eckmann-Hilton argument.
Observation 2.22. Fix I a weak indexing system. By Proposition 1.44, there is a contractible space of diagrams
of the following form:
⊗ ⊗ id ⊗!
⊗ triv⊗ ⊗ ⊗ ⊗
BV BV
NI∞ ≃ NI∞ c(I) −
−−−→ NI∞ ⊗ NI∞ → NI∞ ;
⊗ ⊗ ⊗
furthermore, the composite NI∞ → NI∞ is homotopic to the identity since NI∞ has contractible endomorphism
space. In particular, this implies that there is a unique natural split diagonal diagram

CAlgI CAlg⊗ (−)

I
δ U

CAlgI (−) CAlgI (−)

δ takes a structure to two interchanging copies of itself, and U simply forgets one of the structures. ◁
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 31

A weak ∞-categorical form of the Eckmann-Hilton argument for I-commutative algebras would state
⊗ BV ⊗
that the functor U is an equivalence, or equivalently, δ is an equivalence, i.e. NI∞ ⊗ NI∞ is ∞-connected; the
⊗ BV
⊗
specialization to (n + 1)-categories is that NI∞ ⊗ NI∞ is n-connected. Unfortunately, this does not hold for all
I ∈ wIndexT . The following simple counterexample was pointed out to the author by Piotr Pstrągowski.
Example 2.23. Let R be a nonzero commutative ring and let Comm⊗ nu be the weak N∞ -∗-operad associ-
ated with the ∗-weak indexing system Fnu = F − {∅}. Then, the Abelian group underlying R supports a
Comm⊗
BV
⊗
nu ⊗ Commnu structure given by the two multiplications

µ(r, s) = rs, µ0 (r, s) = 0,

which are easily seen to satisfy interchange but be distinct. This lies outside of the essential image of
δ : AlgCommnu (Ab) −→ AlgCommnu Alg (Ab),
Commnu

so δ is not an equivalence; by Proposition 1.24, this implies that Comm⊗

BV
⊗
nu ⊗ Commnu is not connected. ◁
BV
In the positive direction, [SY19] yields a classification of ⊗ -idempotent algebras in reduced ∞-operads.
In fact, Example 2.23 shows that the associated unitality assumption only misses one example among
nonequivariant one-color weak N∞ -operads.
Corollary 2.24. A weak N∞ -∗-operad O ⊗ possesses a map triv⊗ → O ⊗ inducing an equivalence
∼
O ⊗ −−−−−→ O ⊗ ⊗ O ⊗
BV

if and only if O ⊗ is equivalent to triv⊗ , E⊗ ⊗

0 , or E∞ .

Proof. [SY19, Cor 5.3.4] covers the reduced case, so it suffices to assume that O(∅) = ∅ and show that
O ⊗ ≃ triv⊗ . Note that Comm⊗ ⊗ ⊗
nu is the terminal nonunital N∞ -∗-operad, i.e. there exists a map O → Commnu ,
yielding a diagram
ϕ
O⊗ ⊗ O⊗ Comm⊗ ⊗
nu ⊗ Commnu

O⊗ Comm⊗
nu

Pulling back Example 2.23, we find that if O(n) = ∗ for any n , 1, then ϕ ∗ R ∈ AlgO⊗O (Ab) is not in the image
of the diagonal; contrapositively, O(n) = ∅ when n , 1, i.e. it’s equivalent to triv⊗ . □
n o
We saw in [Ste24] that triv⊗ → E⊗ ⊗
0 → E∞ = Op∗
auni, weak-N∞
. In this section, we introduce an equivariant
analogue to this argument in order to prove the following proposition.
⊗ ⊗
BV
Proposition 2.25. Suppose NI∞ ⊗ NI∞ is connected. Then, I almost essentially unital.
By combining Proposition 2.25 and Corollary 2.4, we conclude the remaining part of Corollary D.
⊗ ⊗
Corollary 2.26. NI∞ ⊗ NI∞ is a weak N∞ -operad if and only if I is almost essentially unital; in particular,
⊗ ⊗ ∼ ⊗ BV ⊗
there exists a (necessarily unique) map triv⊗ → NI∞ inducing an equivalence NI∞ −−−−−→ NI∞ ⊗ NI∞ if and
only if I is almost-unital.
To show Proposition 2.25, we pass to a universal case. First, the weak indexing system.
Recollection 2.27. In [Ste24], we computed the terminal weak indexing system with unit family F to be


FV
 V ∈F;
FF ⊥ −nu,V = 
FV − {S | ∀U ∈ Orb(S), U ∈ F } V < F ;


in particular, FI fails to be almost essentially unital if and only if there is some non-contractible W -set in
Fυ(I)⊥ −nu,W ∩ FI,W for some W ∈ υ(I)⊥ . We refer to the associated weak indexing category as IF ⊥ −nu ; note
that IF ⊥ −nu ⊂ FT is the wide subcategory of maps T → S such that either S, T ∈ FF or S, T ∈ FF ⊥ . ◁
Now, we construct a family of problematic NI⊗ ⊥ −nu ⊗ NI⊗ ⊥ −nu -algebras.
BV

F F
32 NATALIE STEWART

Construction 2.28. Let M be a T -commutative monoid in pointed sets. We define a new functor
M 0 : h1 SpanIF ⊥ −nu (FT ) → Set∗

which agrees with M on objects, backwards maps, forwards maps lying in FF , but whose forward maps
lying in FF ⊥ are zero. This is evidently functorial on backwards and forward maps, and the restriction to
backwards maps is product-preserving. We’re left with verifying the double coset formula that, given a
cartesian square as on the left such that S, T , R ∈ T and f , f ′ ∈ IF ⊥ −nu , the right square commutes, where
(−)∗ denotes covariant functoriality and (−)∗ contravariant.

R ×S T M 0 (R ×S T )
g′ f′ g ′∗ f∗′
⌟

R T M 0 (R) e 0 (T )
M
f g f∗ g∗
S M 0 (R)

The assertions that f , f ′ ∈ IF ⊥ −nu and that F is a family together imply that T ∈ F if and only if the entire
diagram lives in FF , and T ∈ F ⊥ if and only if the entire diagram lives in FF ⊥ . In the former case, the right
diagram commutes by the double coset formula for M, and in the latter case it commutes as each composite
map is zero. ◁
Lemma 2.29. For all T -commutative algebras M, the IF ⊥ −nu -commutative algebras M and M0 interchange.
Proof. Note a diagram of T -coefficient systems in a 1-category commutes if and only if the V -fixed point
diagram commutes for all V ∈ T ; the V -fixed points of the diagram in Recollection 1.32 in our case correspond
with the diagram
 S  T (tr0S )T
XT ≃ X S×T ≃ XS XT
(trT )S trT
tr0S
XS X
where tr∗ is the indexed multiplication in M and tr0∗ is the indexed multiplication in M 0 ; when V ∈ F ⊥ , this
commutes as each of the composites factor through a zero map.
Moreover, note that BorTF IF ⊥ −nu is unital, so Corollary 2.4 implies that BorTF IF ⊥ −nu -algebras interchange
with themselves; in particular, the interchange relation of Recollection 1.32 for M with itself implies the same
relation for M and M0 whenever V ∈ F . □

We are left with constructing a highly nontrivial T -commutative algebra; we choose a universal one.
Construction 2.30. Since the “isomorphism classes of objects” functor π0 : Cat → Set preserves limits, pushfor-
ward along it lifts to a functor
π0 : Cat⊗
T ≃ CMonT (Cat) → CMonT (Set);

the effective Burnside T -commutative monoid is AT B π0 FTT −⊔ . We denote its image under the maps
CMonT (Set) ≃ CMonT (Set∗ ) → CMonIF ⊥ −nu (Set∗ )

implied by Corollaries 1.54 and 2.4 by A

e .
T ◁

Lemma 2.31. The S-indexed multiplication in A e0 are distinct for all S ∈ FF ⊥ −nu,V − {∗V } and V ∈ F ⊥ .
e and A
T T

Proof. It suffices to prove that, for all S , ∅V ∈ FV , the S-ary multiplication of AT takes some element to
another element other than the unit; unraveling definitions, this is equivalent to the property that some
nonempty V -set can be expressed as an S-indexed coproduct. S provides such an example. □

We now exhibit failure of the non-almost-essentially-unital 1-categorical Eckmann-Hilton argument.

 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 33

Proof of Proposition 2.25. Note that

⊗ BV
⊗ ⊗ ⊗ ⊗ BV
⊗
NI∞ ⊗ NI∞ is connected ⇐⇒ h1 NI∞ ⊗ NI∞ ≃ NA(N ≃ NI∞
I∞ ⊗NI∞ )
=⇒ CAlgI (Set∗ ) −→ AlgNI∞ ⊗NI∞ (Set∗ ) is essentially surjective.
⊗ ⊗ BV
Furthermore, Lemmas 2.29 and 2.31 construct an Nυ(I) ⊥ −nu∞ ⊗ Nυ(I)⊥ −nu∞ -algebra A satisfying the condition

that its two individual structure maps A(S) → A(∗V ) differ whenever V ∈ υ(I)⊥ and S , ∗V . Since I is
not almost essentially unital, it must admit some noncontractible S ∈ FI,V for V ∈ υ(I)⊥ , so the pullback
⊗ BV ⊗
NI∞ ⊗ NI∞ structure on A has two distinct underlying I-algebra structures, implying it is outside of this
⊗ BV ⊗
essential image. The contrapositive shows that NI∞ ⊗ NI∞ is not connected. □

Remark 2.32. Using the above argument, one can show that if O ⊗ is a idempotent object in T -operads, then
its nullary spaces O(∅V ) are nonempty. If additionally O(∅V ) are assumed to be contractible (i.e. O ⊗ is
almost-unital), then Proposition 2.18 shows that the underlying fixed point catgeories OV are all idempotent
algebras, i.e. they are contractible. Hence O ⊗ will be shown to be almost-reduced. In forthcoming work, we
will develop an equivariant lift of [SY19], which would imply that every idempotent almost-unital T -operad
is a weak N∞ -operad. ◁

3. Corollaries in higher algebra

We now indulge in a number of corollaries. We begin in Section 3.1 by making the equivariant
BV
equifibered perspective symmetric monoidal, establishing that ⊗ is T -bifunctorial and distributive, fulfilling
Assumption (b) in full generality. We make use of the distributivity in Section 3.2, where we apply the
disintegration and assembly procedure of Appendix B to compute tensor products of T -space colored T -
operads. Then, in Section 3.3 we show how Corollary D constructs an I-symmetric monoidal structure on
right-modules over an I-commutative algebra.
Moving on from this, in Section 3.4 we spell out the cases of equivariant Dunn additivity that follows from
Corollary D and Theorem F; in Section 3.5 we show that equivariant factorization homology is G-symmetric
monoidal and spell out how to infinitely iterate equivariant factorization homology and TCR.

3.1. Coherences and restrictions of equivariant Boardman-Vogt tensor products. We would like to construct
BV
coherences for ⊗ using the argument of [BS24a], but it is currently not known whether EnvI is monic in Cat,
so we must modify their argument: we use that the sliced envelope fully-faithful [BHS22, Prop 4.2.1].
Now, we proved in Corollary 2.26 that Comm⊗ T bears a unique structure as an idempotent object;
in particular, Proposition 1.34 shows that Env is compatible with finitary tensor products, so it induces a
⊛-idempotent object structure on FTT −⊔ ∈ Cat⊗T . By [HA, Prop 4.8.2.9] this underlies a unique E∞ -algebra
under the mode structure. Then, Corollary C.8 constructs a symmetric monoidal T -∞-category structure on
the T -overcategory Cat⊗ T −⊔ whose underlying tensor functor has value
T ,/FT

πC ⊛πD ∼
C ⊛ D −−−−−−→ FTT −⊔ ⊛ FTT −⊔ −−−−−→ FTT −⊔ .

and whose unit is

  η
Env triv⊗
T − → FTT −⊔ .
We acquire existence and uniqueness of the Boardman-Vogt symmetric monoidal structure.

Corollary E’. Op ⊂ Cat⊗ is a symmetric monoidal subcategory under ⊛, with unit corresponding
T T ,/FTT −⊔
with triv⊗
BV

T and tensor bifunctor corresponding with ⊗ . Hence there exists a unique symmetric monoidal
T -∞-category Op⊗ and symmetric monoidal T -functor
T

Op⊗ → Cat⊗,⊛
T T ,/FTT −⊔

lifting the sliced T -symmetric monoidal envelope.

34 NATALIE STEWART

T −⊔
Proof. We’re tasked with proving that the image of Env/FT (−) contains the unit and is closed under binary
tensor products. The unit is Proposition 1.34, and by construction we have a commutative diagram
BV
 BV  Env(πO ⊗ πP )   Env(id ⊗η)  
Env Comm⊗ ⊗
Env Comm⊗
BV
Env O ⊗ ⊗ P ⊗ T ⊗ CommT ∼ T

≃
≃ ∼
Env (O ⊗ ) ⊛ Env (P ⊗ ) πEnv(O ⊗ ) ⊛πEnv(P ⊗ )
FTT −⊔ ⊛ FTT −⊔ FTT −⊔
id ⊛η

Inverting Env(id ⊗η) and id ⊛η yields the desired equivalence

T −⊔   T −⊔   T −⊔  
Env/FT O ⊗ ⊗ P ⊗ ≃ Env/FT O ⊗ ⊛ Env/FT P⊗ .
BV
□
BV
Corollary 3.1. There exists a G-bifunctor ⊗ : Op × Op −→ Op whose V -value OpV × OpV → OpV is
T T T
 
∧
O ⊗ ⊗ P ⊗ ≃ LOpV O ⊗ × P ⊗ −→ Span(FV ) × Span(FV ) −−−−→ Span(FV ) .
BV

This recovers the correct construction in the nonequivariant case.

Corollary 3.2. When T = ∗, there is an equivalence of symmetric monoidal ∞-categories
Op⊗ ⊗
∗ ≃ Op ,

where the latter is the Boardman-Vogt symmetric monoidal ∞-category of [BS24a].

Proof. In [Ste25] we supplied an equivalence Op∗ ≃ Op, so it suffices to upgrade this to a symmetric monoidal
equivalence. In fact, the unslicing functor Cat⊗ ⊗
∞,/F⊔ → Cat∞ bears a symmetric monoidal structure (see
Corollary C.6), so Corollary E’ constructs a symmetric monoidal structure on the composite Op⊗ ⊗
∗ → Cat∞ .
Thus [BS24a, Thm E] constructs a symmetric monoidal equivalence extending the equivalence Op∗ ≃ Op. □

Corollary 3.3. Let I be a one color weak indexing system and n ∈ N ∪ {∞}. Then, OpI ⊂ OpT is a symmetric
monoidal subcategory, Opuni
I ⊂ OpI is a smashing localization, and the following are symmetric monoidal full
subcategory inclusions:
red red
OpaE
I,≥n ⊂ OpaE
I ⊂ OpaEuni
I ⊂ OpI

Opred red
I,≥n ⊂ OpI ⊂ Opuni
I

Proof. The first statement follows from Proposition 1.44 and the second from Corollary 2.15. triv⊗ T and
E⊗
0,υ(I) are ∞-connected, so in particular, the symmetric monoidal units are compatible with each of the
above subcategory inclusions. We’re left with verifying that each subcategory inclusion is closed under tensor
products; the lefthand inclusions both follow from Corollary 2.20, the middle inclusions both follow from
Proposition 2.18, and the righthand inclusion is Corollary 2.19. □

We finish the subsection by confirming a convenient structural result.

Corollary 3.4. Op⊗ , and OpJ−cocart,⊗ are presentably symmetric monoidal T -∞-categories.
I I

To see this, Op⊗ is presentable by the localizing inclusion Op ⊂ CatT ,/FI−⊔ , and it is distributive by the
I I I
tensor-hom T -adjunction (−) ⊗ O ⊗ ⊣ Alg⊗ (−). The remaining case follows from the following easy lemma.
BV

Lemma 3.5. If L : D⊗ → C ⊗ is a smashing T -localization and D⊗ is a presentably symmetric monoidal

T -∞-category, then C ⊗ is a presentably symmetric monoidal T -∞-category.
Proof. It’s clear that C is a presentable T -∞-category, so we’re left with verifying that − ⊗ C : C → C possesses
a right T -adjoint; by the usual argument, it suffices to show this on fixed points, so we may assume T = ∗.
We claim that − ⊗ C ⊣ homD (C, −), the latter denoting the D-internal hom. It suffices to verify that
homD (C, D) is L-local for all D ∈ D. We apply the standard argument: if f : X → Y is an L-equivalence, then
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 35

C ⊗ f ∼ C ⊗ Lf : C ⊗ X → C ⊗ Y is an equivalence, so the horizontal arrows in the following are equivalences.

f∗
Map(Y , hom(C, D)) Map(X, hom(C, D))

≃
(C⊗f )∗
Map(Y ⊗ C, D) Map(X ⊗ C, D)

The fact that the top arrow is an equivalence is the desired locality. □

3.2. Disintegration and equivariant Boardman-Vogt tensor products. We show the following generalization
of the main results of [HA, § 2.3.3-2.3.4] in Appendix B.
Theorem 3.6 (Disintegration and assembly). Let X be a T -space. Taking fibers yields an equivalence
Op ≃ FunT (X, Op ).
I,/X I−⊔ I
The counit of this specifies a natural equivalence
 
⊗ ∼
colimx∈X ResTstab(x) O ⊗ ×ResT X I−⊔ IV ∞ −
N −−−→ O ⊗ .
stab(x)

Here, colimx∈X refers to a T -colimit of an X-indexed diagram. Given x ∈ X V we’ve written stab(x) B V .
Given O ⊗ a T -operad, I a one-color T -weak indexing category, and x ∈ OV a V object, we define the reduced
endomorphism IV -operad of x to be the pullback
ιx
EndI,red
x (O) ResTV O ⊗
⌟
! η

NI⊗V ∞ ResTV U O I−⊔

{x}

In the case I = T , we simply write Endred T ,red

x (O) B Endx (O).
Remark 3.7. If O ⊗ is unital (resp. almost-unital) then Endred
x (O) is reduced (almost-reduced). ◁
We acquire the following from Theorem 3.6.
Corollary 3.8. Suppose O ⊗ is a T -operad whose underlying T -∞-category U O is a T -space and I is a one-color
weak indexing system. Then, the inclusion maps ιx assemble to a T -colimit diagram in I-operads:
∼
colimx∈U O EndI,red
x (O) −−−−−→ BorTI O.
In essence, this says that an at-least one color T -operad O ⊗ whose underlying T -∞-category is a T -space
disintegrates into a U O-local system of one color T -operads, and the U O-indexed colimit T -operad (i.e. the
Grothendieck construction) assembles O ⊗ from this local system. In particular, O-algebras are U O-indexed
systems of Endred
x (O)-algebras:

Alg (C) ≃ Alg (C)

O colimx∈U O Endred
x (O)
≃ limx∈U O Alg (C).
Endred
x (O)

The corresponding picture for O ⊗ ⊗ P ⊗ -algebras is U O × U P -local systems of Endred red

BV BV

x (O) ⊗ Endy (P )-algebras:

that is, we can compute tensor products of at-least one color T -operads in terms of one color T -operads, as
long as they are T -space colored.
Corollary 3.9 (Disintegration of tensor products). Suppose O ⊗ , P ⊗ , Q⊗ are at-least one colored T -operads
BV
whose underlying T -∞-categories are T -spaces and ϕ : O ⊗ ⊗ P ⊗ → Q⊗ is a map such that
(a) the underlying map of T -spaces U ϕ : U O × U P → U Q is an equivalence, and
(b) for all pairs (x, y) ∈ U O × U P , ϕ pulls back to an equivalence
ϕ(x,y) : Endred red red
BV

x (O) ⊗ Endy (P ) → End(x,y) (Q).

Then ϕ is an equivalence.
36 NATALIE STEWART

Proof. Corollaries 3.4 and 3.8 construct equivalences of arrows

colimx∈U O Endred red

colim(x,y)∈U O×U P Endred red
BV BV BV
O⊗ ⊗ P ⊗ ≃ x (O) ⊗ colimy∈U P Endy (P ) ≃ x (O) ⊗ Endy (P )
ϕ ∼

Q⊗ ≃ colim(x,y)∈U O×U P Endred

(x,y) (Q) colim(x,y)∈U O×U P Endred
(x,y) (Q)

The right vertical arrow is an equivalence by assumption, so ϕ is an equivalence by two out of three. □
We will make crucial use of this in forthcoming work concerning variants of E⊗
V with tangential structure.
3.3. Norms of right-modules over I-commutative algebras. Let I be an indexing category, C ⊗ an I-symmetric
monoidal ∞-category, t : W → V an I-admissible transfer, A an IV -commutative algebra, and M a right
module over the associative algebra underlying ResVW A. Then, we may define the A-module norm of M by
the base-changed A-module
V V
A NW M B A ⊗N V ResV A NW M; W W
V
that is, the normed multiplication recognizes A as an NW ResVW A-module, and the A-module norm of M
V V
is the free A-module on the normed NW ResW A-algebra of M; see [Yan23] for a detailed account in the
Cp -equivariant case.
⊗
≃ E⊗ ⊗
BV
In this subsection, we use the equivalence NI∞ 1 ⊗ NI∞ to lift this to an I-symmetric monoidal
structure, yielding coherent functoriality and a coherent double coset formula for A-module norms. To
do this, we begin by bootstrapping G-symmetric monoidality of the right module construction from the
non-equivariant case.
Observation 3.10. Fix O ⊗ a T -operad. By [HA, Rmk 4.8.3.8], functors F : TotTotT O ⊗ → CatAlg are data
C⊗
AF
πF

E⊗
1 × TotTotT O
⊗ E⊗
1 × TotTotT O
⊗

such that π is a cocartesian fibration whose fibers CV⊗ → E⊗ 1 are the unstraightenings of small monoidal
∞-categories and such that the composite arrows E⊗ 1 × {O} ,→ C ⊗ are associative algebras. Moreover,
unwinding definitions and applying Lemma A.18, the condition that F corresponds with an O-monoid
TotT O ⊗ → CoeffT CatAlg corresponds with the condition that each of the fibers Cn⊗ → TotTotT O ⊗ is an
O-monoidal ∞-category. ◁
Given C ⊗ ∈ Cat⊗ and A ∈ AlgE1 ⊗O (C), we acquire a functorial diagram
E 1 ⊗O

RModA (C)⊗
TotT O ⊗
(A,C ⊗ )

CoeffT CatAlg ×CoeffT Alg(Cat) TotT O ⊗ CoeffT CatAlg CoeffT CatMod CoeffT Cat
Θ Y
⌟ U
T
Coeff Alg(Cat)
≃

C⊗
TotT O ⊗ Alg(CoeffT Cat)

θ and Y are product preserving functors, so RModA (C)⊗ is an O-monoid in Cat, i.e. an O-monoidal ∞-category.
Unwinding definitions, this proves the following proposition.
Proposition 3.11. Let O ⊗ be a T -operad and let C ⊗ an E1 ⊗ O-monoidal ∞-category. There is a lift
RMod⊗
(−)
(C) Cat⊗
O

ΓT
RMod(−) (C)
AlgO⊗E1 (C) AlgE1 (C) Cat,
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 37

natural separately in O ⊗ and C ⊗ ; that is, left modules over E1 ⊗ O-algebras bear a natural O-algebra structure.
We immediately acquire the following corollary, confirming a hypothesis of [Hil17, Rmk 3.15].
Corollary 3.12. Let O ⊗ be a T -operad whose underlying I ∞ -operad is E∞ and C ⊗ an O-monoidal ∞-category.
There is a lift
RMod⊗
(−)
(C) Cat⊗
O

RMod(−) (C)
AlgO (C) AlgE1 (C) Cat

natural separately in O ⊗ and C ⊗ . In particular, if I is an indexing category, the ∞-category of right-modules

over an I-commutative algebra admits a natural I-symmetric monoidal structure.
3.4. Equivariant infinitary Dunn additivity. In [Bon19], a genuine operadic nerve 1-categorical functor was
constructed between a model of graph-G operads and a model of G-operads. In [Ste25], we lifted this to a
conservative functor of ∞-categories N ⊗ : gOpG → OpG . Given V an orthogonal G-representation, we define
E⊗ ⊗
V B N DV ,
where DV is the little V -disks graph G-operad of [GM17], whose n-ary G × Σn space is the configuration space
DV (n) ≃ Confn (V )
by [GM17, Lem 1.2]. The resulting unital G-operad EV was studied in [Hor19], who showed for instance that
EV (S) ≃ ConfH H
S (V ) B colim ConfS (W ),
W ⊂V
fin.dim

in view of the fact that the assignment O 7→ O(S) preserves sifted colimits [Ste25, § 2.3]; here, ConfH
S (W ) is
the space of H-equivariant configurations of S into W under the compact open topology.
A weak form of the following easy claim appears to be folklore.
Proposition 3.13. Let G be a topological group, H ⊂ G a closed subgroup, S ∈ FH a finite H-set admitting an
configuration ι : S ,→ W , and V , W orthogonal G-representations whose associated map
ConfH H
S (V ) ,→ ConfS (V ⊕ W )

is an equivalence. Then, ConfH

S (V ) is contractible.

Proof. Linear interpolation to ι yields a deformation of MapH (S, V ⊕ W ) onto {ι}. The path of a point
beginning in the subspace ConfH H
S (V ) ⊂ ConfS (V ⊕ W ) consisting of configurations with zero projection
to W lands within ConfH S (V ⊕ W ) at all times; composing this deformation after the deformation retract
∼
ConfHS (V ⊕ W ) −
→ Conf S H 5
H (V ) yields a deformation retract of ConfS (V ⊕ W ) onto {ι}, so it is contractible. By
the equivalence ConfH H H
S (V ) ≃ ConfS (V ⊕ W ), the space ConfS (V ) is contractible as well. □

Remark 3.14. This argument only produces contractibility, whereas the nonequivariant argument using Fadell
and Neuwirth’s fibration [FN62] sharply characterizes n-connectivity of Confk (Rn ), and hence of E⊗k ; the
author will equivariantize this in forthcoming work. ◁
We say that V is a weak universe if it is a direct sum of infinitely many copies of a collection of irreducible
orthogonal G-representations; equivalently, there is an equivalence V ≃ V ⊕ V . Given V an orthogonal G-
representation, we let AV B AEV , i.e. AV corresponds with the weak indexing system FV = FAV of finite
H-sets admitting an embedding into V . The following corollary follows immediately from Proposition 3.13.

5 Said explicitly, let h : [0, 1] → ConfH (V ⊕ W ) be the deformation retract onto those configurations with zero projection to W .
S
Then, our deformation retract h′ onto ι(w) is computed by

1,
h(2t) t≤

h′ (t) =  2

(2 − 2t) · h(1) + (2t − 1) ι
 t≥ 1.
2

The second is an isotopy since h(1) and ι are pointwise-linearly independent embeddings.
38 NATALIE STEWART

Corollary 3.15. If there exists an equivalence E⊗ ⊗ G ⊗ ⊗

V ≃ EV ⊕W , then the canonical map BorAW EV → NAW is an
equivalence; in particular, if V is a weak universe, then there is a unique equivalence
∼
E⊗
V −
⊗
−−−→ NAV .
Observation 3.16. If V is a universe (i.e. it is a weak universe admitting a positive-dimensional fixed point
locus), then it admits embeddings of all finite sets with trivial G-action; in this case, E⊗
V is not just a weak
N∞ -operad, but an N∞ -operad. ◁
Much study has been dedicated to the less general setting of universes; for instance, Rubin has given a
complete and simple characterization of those indexing systems (equivalently, transfer systems) occurring as
the arity-support of an EV -operad in [Rub19] for G abelian.
An inclusion V ⊂ W yields a map of graph G-operads DV → DW , hence a map E⊗ ⊗
V → EW . This yields a
map of weak indexing systems FV → FW ; in [Ste24] we showed that this is additive, i.e.
(14) FV ∨ FW = FV ⊕W .
Corollary 3.17 (Equivariant infinitary Dunn additivity). Let G be a finite group and V , W real orthogonal
G-representations satisfying at least one of the following conditions:
(a) V , W are weak G-universes, or
(b) the functoriality map E⊗ ⊗
V → EV ⊕W is an equivalence.
Then there is a canonical equivalence
E⊗ ⊗ ⊗
BV

V ⊗ EW → EV ⊕W ;
equivalently, for any G-symmetric monoidal category C, there are canonical equivalences6
AlgEV Alg⊗ (C) ← AlgEV ⊕W (C) → AlgEW Alg⊗ (C).
EW EV

Proof. Given Corollary 3.15, case (a) follows from Theorem F and Eq. (14) and case (b) follows from
Corollary D. □
Remark 3.18. In [Szc24], an ostensibly similar result to Corollary 3.17 is proved: given DV the little Disks
graph G-operad, Szczesny constructs a non-homotopical Boardman-Vogt tensor product ⊗ and a canonical
map DV ⊗ DW → DV ⊕W , which he shows to be a weak equivalence of graph G-operads in [Szc24, Thm 7.1].
Neither this result nor Corollary 3.17 imply each other.
On one hand, Szczesny’s result concerns a tensor product with no known homotopical properties, so it
is incomparable with results concerning ∞-categories of algebras defined by homotopy-coherent universal
properties. On the other hand, while Corollary 3.17 is homotopical, it only concerns cases where at least one
of the representations induces I-symmetric monoidal ∞-categories of algebras whose indexed tensor products
are indexed coproducts; this property will not be satisfied for any nontrivial indexed tensor products in the
finite-dimensional case, so the range of representations in Szczesny’s result is significantly larger. The author
will address the general case in forthcoming work. ◁
3.5. Norms on Real topological Hochschild and cyclic homology.
3.5.1. Factorization homology in general. In classical algebra, there is a well-known tensor products of functors
F, G : C → D using monoidal structure of D: the pointwise tensor product sets F ⊗ G(C) B F(C) ⊗ G(C). We
will use a lift of this due to Nardin-Shah.
Theorem 3.19 ([NS22, Thm 3.3.1, 3.3.3]). Let K be a T -∞-category and C ⊗ a T -operad. Then, there exists a
unique (functorial) I-operad structure FunT (K, C)⊗−ptws on FunT (K, C) satisfying the universal property
AlgO (FunT (K, C)⊗−ptws ) ≃ FunT (K, Alg (C))
O
for O ∈ OpI . Furthermore, when C ⊗ is I-symmetric monoidal, FunT (K, C)⊗−ptws is I-symmetric monoidal
and satisfies the universal property
   
FunI−⊗
T D, FunT (K, C) ⊗−ptws
≃ FunT K, FunI−⊗
T (D, C) .

6 What we mean by “canonical” depends on the case; for case (a), there is a contractible space of equivalences, and for case (b), this
⊗ id ⊗! ⊗ ι⊗id ⊗ BV ⊗
equivalence comes from inverting arrows of the zigzag E⊗ ⊗ ⊗
BV BV
V ⊕W ≃ EV ⊕W ⊗ trivG −−−−→ EV ⊕W ⊗ EW ←−−−− EV ⊗ EW .
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 39

⊗−ptws
If additionally, S is I-admissible, then the S-indexed tensor product of (FU ) ∈ FunT (K, C)S has value
NS
∆S (FU )
DV DS CS CV
S
N
FU
U

Observation 3.20. Suppose F : K′ → K is a functor. Then, the restriction and left Kan extension natural
transformations    
F! : FunT K′ , FunI−⊗
T (D, C) ←
−−−
−−
−−
−−
−−
−−
→
−
I−⊗
−− FunT K, FunT (D, C) : F
∗

yield I-symmetric monoidal functors FunT (K′ , C)⊗−ptws ⇄ FunT (K, C)⊗−ptws extending the left Kan extension
and restriction functors between functor categories via Yoneda’s lemma. In particular, given X ∈ Γ T K
this yields an I-symmetric monoidal lift evX : FunT (K, C)⊗−ptws → C ⊗ of the ordinary evaluation T -functor
FunT (K, C) → FunT ({X} , C) ≃ C. ◁
The following proposition is easy, so we omit its proof.
Proposition 3.21. There exists a natural equivalence FunT (K, C)⊗−ptws ≃ Alg⊗ (C).
triv(K)

The structure functor Env(O) → O is adjunct to a T -operad map triv(Env(O))⊗ → O ⊗ , which yields a
natural pullback T -symmetric monoidal functor
(15) U : Alg⊗ (C) → FunT (Env(O), C)⊗−ptws .
O
In particular, this constructs a G-symmetric monoidal lift for genuine equivariant factorization homology.
Corollary 3.22. Given M a V -framed smooth G-manifold, M-factorization homology lifts to a G-symmetric
monoidal functor
Z
: Alg⊗ (C) → C ⊗ ;
EV
M
in particular, it further lifts to a G-symmetric monoidal endofunctor
Z
: CAlg⊗ (C) → CAlg⊗ (C).
AV AV
M

Proof. In the notation of [Hor19], let ι⊗ : DiskG,V −f r,⊔ → MfldG,V −f r,⊔ beR the symmetric monoidal inclusion
of V -framed G-disks into V -framed G-manifolds. By [Hor19, Prop 4.1.4], M may be presented as the G-value
of a composition
Z   U   ι!   evM
: Alg (C) ≃ Fun⊗ G Disk
G,V −f r
, C −−→ FunG DiskG,V −f r , C −→ FunG MfldG,V −f r , C −−−−→ C.
EV
M
R
To construct the lift of M , we may compose G-symmetric monoidal lifts of U , ι! , and evM ; these are given by
Eq. (15) and Observation 3.20, respectively. □

3.5.2. Multiplication and norms on THR. We specialize Corollary 3.22 to (V , M) = (σ , S σ ).

Corollary 3.23. Real topological Hochschild homology lifts to a C2 -symmetric monoidal functor
THR : Alg⊗ (C) → C;
Eσ

in particular, THR lifts to a C2 -symmetric monoidal endofunctor

THR : Alg⊗ (C) → Alg⊗ (C).
EV +∞σ EV +∞σ

Given A ∈ AlgEV +∞σ (C), there is an equivalence

THR(A) ≃ colimS σ A,
with colimit taken in AlgEV +∞σ (C), naturally in A.
40 NATALIE STEWART

Proof. The last sentence is the only part which does not follow immediately from combining Horev’s
facorization homology formula [Hor19, Rmk 7.1.2] with Corollaries 3.17 and 3.22. It suffices to show
⊗
the colimit property for O ≃ O ⊗ Eσ -algebras whenever BorTAσ O ⊗ ≃ NAσ ∞ , which holds for EV +∞σ by
Proposition 3.13. In any case, naturality of the dihedral bar construction together with the the Wirthmüller
maps of Construction 1.62 yields a diagram
..
. A⊔µ3 A⊔µ2 A colimS σ A
ϕ
..
. A⊗µ3 A⊗µ2 A THR(A);

where µn is the n-element “dihedral” C2 -set, i.e. the unique σ -admissible C2 -set of size n, and each row is a
geometric realization diagram. When the domain category is Aσ -semiadditive, the vertical maps between the
bar constructions are equivalences, so ϕ is an equivalence. The result then follows by Aσ -semiadditivity of
O-algebras, as in Theorem F’. □
Remark 3.24. The computation THR(A) = colimS σ A when A is pulled back from a C2 -commutative algebra is
not new; indeed, it appears as [QS19, Rmk 5.4]. In fact, the ambiguity induced by the potential discrepancy
between our construction Alg⊗ (C) and that of [NS22, Thm 5.3.4] vanishes for the I-symmetric monoidal
O
structure on CAlgI (C) by applying Theorem A’ in view of the fact that each are cocartesian [NS22, Thm 5.3.9].
The new element of this identification is that the operation on C2 -commutative algebras is induced canonically
from the operation on Eσ -algebras and that the colimit formulas need only an E∞σ -algebra structure. ◁
Remark 3.25. In the above, we only needed C to be an Eσ -monoidal ∞-category; however, to easily understand
O-algebras, one ought to assume that C is an O-monoidal ∞.category. ◁
Now, we can construct a circle action on THR.
Construction 3.26. Applying Observation 3.20 along an inclusion K0 ⊂ K yields a T -functor
 
K0 → FunI−⊗T FunT (K, C) ⊗−ptws ⊗
, C .

In Corollary 3.22, we are such a situation; define DiffV −f r (M) ⊂ EmbV −f r (M, M) to be the topological
subspace of diffeomorphisms and embeddings of M with conjugation G-action, considered as a (grouplike)
E1 -G-space. Applying the above construction for BDiffV −f r (M) ≃ BAutMfldV −f r (M) ⊂ MfldG,V −f r yields a
DiffV −f r (M)-action on M (−) through G-symmetric monoidal natural transformations, where BDiffV −f r (M)
R

is the unique connected G-space with ΩBDiffV −f r (M) ≃ DiffV −f r (M) as E1 -G-spaces (see [HA, § 5.2.6]). In
particular, this yields a natural lift
V −f r
Alg (C)BDiff (M)
O

Alg (C) R C.
O⊗EV
M

Applying the left-action S σ → Diffσ −f r (S σ ) yields a lift

σ
Alg (C)BS
O

Alg (C) C
O⊗Eσ THR

We refer to this as the S σ action on THR. ◁

In fact, this type of action has been seen in previous work.
Remark 3.27. For ψ : G → G/N a surjective topological group homomorphism, the N -free G-family is
ψ n o op
BG/N N = [G/H] ResG N ([G/H]) is a free N -set ⊂ OG ;
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 41

[QS19, Lem 4.14] recognizes the “quotient by N ” forgetful functor

ψ op
BG/N N → OG/N ,
ψ
as the unstraightening of a G/N -space, which we also refer to as BG/N N . In particular, for an arbitrary
Abelian group A, they define BtC2 A with respect to the semidirect product extension A → A ⋊ C2 → C2 for the
C2 -action on A by inversion. In the case that A = T is the circle group, this is given by the usual extension
T = SO(2) → O(2) → C2 .
We may explicitly identify BtC2 T by hand; it’s well-known that there are exactly two conjugacy classes of
subgroups (H) ⊂ O(2) whose homogeneous spaces [O(2)/H] are T-free: the reflections and the trivial subgroup.
The normalizer of a subgroup generated by a reflection is a dihedral group of order 4, so
AutO(2) ([O(2)/C2 ]) ≃ WO(2) (C2 ) ≃ D4 /C2 ≃ C2 .
In particular, we may picture the unstraightening of BtC2 T via the diagram

Unstraightening
BtC2 T C2 [O(2)/C2 ] [O(2)/e] O(2)

∗C2 [C2 /C2 ] [C2 /e] C2

 e  C2
In particular, taking fibers, we find that ΩBtC2 T = T, ΩBtC2 S 1 = C2 , and the restriction map C2 → T
is the unique nontrivial such homomorphism. This evidently agrees with the circle group structure on S σ ,
inducing an equivalence BS σ ≃ BtC2 T, so we really have acquired a natural lift

Alg (C)hC2 T
O

Alg (C) C
O⊗Eσ THR
 
where we write C hC2 T ≃ FunC2 BtC2 T, C . ◁
The following reformulation of the equivalence THR(A) ≃ colimS σ A may be familiar.
Observation 3.28. Let Γ be a grouplike E1 -T -space. Evaluation and left Kan extension yield an adjunction
− ⊗ Γ : Γ T C ⇄ FunT (BΓ , C) : U ;
unwinding the left Kan extension formula along ∗T → BΓ shows that the T -object underlying − ⊗ Γ is the
constant Γ = ΩBΓ -indexed T -colimit functor; in particular, Corollary 3.23 understands THR(A) to be the free
EV +∞σ -object on A with C2 -equivariant S σ -action. This free action agrees with that of Construction 3.26,
and in particular this identifies our action with that of [QS22a, § 5]. ◁
3.5.3. Multiplication and norms on TCR. Having produced a C2 -symmetric monoidal construction
   ⊗−ptws
THR : AlgEσ SpC2 → Sp⊗ B FunC2 BtC2 T, Sp⊗
BC2 T C2

which lifts to Quigley-Shah’s construction, we’re poised to become the “future work” indicated in [QS19,
Warning 0.12] by constructing a lax symmetric monoidal p-typical (Borel) Real topological cyclic homology
functor which lifts to Quigley-Shah’s construction. Now, consider the C2 -space map i : BtC2 µp∞ → BtC2 T
defined by the following colimit
BtC2 e ··· BtC2 µpn BtC2 µpn+1 ··· BtC2 µp∞
i
BtC2 T
42 NATALIE STEWART

Postcomposing with i ∗ yields a C2 -symmetric monoidal functor Alg⊗ (SpC2 ) → Sp⊗ .

Eσ BC2 µp∞
Following Nikoulaus and Scholze [NS18, Cons IV.2.1] Quigley-Shah defined the C2 -operad of (Borel)
Real p-cyclotomic spectra by the pullback C2 -operad
 ⊗−ptws
C
RCycSp⊗ FunC2 Infle 2 ∆1 , Sp⊗
p C2
⌟
(ev0 ,ev1 )

Sp⊗t t µp 
Sp⊗ × Sp⊗
BC µp∞ C2 C2

2 U ,(−) C2

where (−)tC2 µp has the lax C2 -symmetric monoidal structure of [QS19, Setup 2.1]; this is a lax equalizer of
C2 -operads. Pullback-stability of cocartesian fibrations guarantees that RCycSp⊗ → Sp⊗ is a cocartesian
C2
fibration of C2 -operads, and in particular, RCycSp⊗ is a C2 -symmetric monoidal ∞-category.
In [QS22a, Const 5.5], Quigley-Shah define a lax C2 -symmetric monoidally natural dihedral Tate diagonal
 ⊗−ptws tC µp
C

functor ∆p : Sp⊗ → FunC2 Infle 2 ∆1 , Sp⊗ whose composites are ev1 ◦∆p A ∼ A⊗µp 2 and ev0 ∼ id.
C2 C2

Remark 3.29. Note that the full subcategory FσC2 = µn | n ∈ N ⊂ FC2 of σ -admissible C2 -sets participates in a


weak indexing system with Fσe = F; in particular these together are closed under indexed coproducts. One can
easily identify the underlying set of output by the double coset formula for indexed coproducts, which in this
C
case simply shows that SU XU = U ∈Orb(S) |U | · |XU |; since Rese 2 : FσC2 → Fσe = F is injective, the dihedral
` P
⊗µ tC µp t µ
 
Tate diagonal of an indexed tensor power has signature X ⊗µn −→ (X ⊗µn ) p 2 ≃ prnX ⊗µnp C2 p . ◁
Construction 3.30 (Dihedral Frobenius). Naturality of the dihedral Tate diagonal, commutativity of C2 -colimits
with C2 -colimits, and the coassembly map for C2 -limits yields a natural diagram

··· X ⊗µ3 X ⊗µ2 X THR(X)

∆p (X ⊗3 ) ∆p (X ⊗2 ) ∆p (X)
 tC mup  tC µp  tC µp  tC µp
··· X ⊗µ3p 2 X ⊗µ2p 2 X ⊗µp 2 colim X ⊗µp 2

 tC2 µp
  tC µp
⊗µ3p ⊗µ2p ⊗µp 
colim X ⊗µp
 2
 · · · X X X

 

 tC2 µp
⊗µ3p ⊗µ2p ⊗µp
THR(X)tC2 µp
 
···

···

···

···

···

···

 · · ·
 X ··· X ··· X ··· X 

We acquire a (lax C2 -symmetric monoidally-) natural composite map ϕp : THR(X) → THR(X)tC2 µp , yielding
a lax C2 -symmetric monoidal functor
 
Alg⊗ Sp
Eσ C2 ϕp

THR
]
 ⊗−ptws
RCycSp⊗
C
FunC2 Infle 2 ∆1 , Sp⊗ ◁
p C2
THR ⌟
(ev0 ,ev1 )

Sp⊗t t µp 
Sp⊗ × Sp⊗
BC µp∞ C2 C2

2 U ,(−) C2
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 43

Now, in [QS19, § 2.3] they also define a lax C2 -symmetric monoidal functor TCR(−, p) : RCycSp⊗ → Sp⊗ .
p C2
From this, we may conclude the following.
Corollary 3.31. Given O ⊗ a C2 -operad, there is a commutative diagram of lax C2 -symmetric monoidal functors
     
THR TCR(−,p)
CAlg⊗ Sp CAlg⊗ RCycSp CAlg Sp
C2 C2 C2 p C2 C2

     
Alg⊗ Sp Alg⊗ RCycSp Alg⊗ Sp
Eσ ⊗O C2 O p O C2

 
TCR(−,p)
Alg⊗ Sp RCycSp⊗ Sp⊗
Eσ C2 p C2

THR
Sp⊗
C2

whose top row recovers the constructions of [QS19]. In particular, if I is a is a C2 -weak indexing system
containing the C2 -set n·[C2 /e]⊔ϵ ·∗C2 for al n ∈ N and ϵ ∈ {0, 1}, then Quigley-Shah’s p-typical Real topological
cyclic homology is lifted from a lax C2 -symmetric monoidal endofunctor of I-commutative ring spectra.
Remark 3.32. The conditions of Corollaries 3.23 and 3.31 were varied for the sake of diversity of examples;
they can be interchanged, and each can be weakened to construct (lax) C2 -symmetric monoidal endofunctors
C ⊗
of O-algebras whenever BorAσ2 O ⊗ ≃ E⊗
σ ∞ ≃ NAσ ∞ . ◁
Remark 3.33. For simplicity, we made the p-typical construction; however, Real cyclotomic spectra and their
Real topological cyclic homology were constructed integrally in [QS22a], and we may lift our constructions to
a lax C2 -symmetric monoidal Real cyclotomic structure as follows.
 
Alg⊗ Sp
Eσ C2 (ϕp )

THR
]
!⊗−ptws
⊗ C
Infle 2 ∆1 , Sp⊗t
Q
RCycSp FunC2
p prime BC T
THR ⌟ 2

(ev0 ,ev1 )

Sp⊗t
Q
Sp × Sp
BC T t µp  BtC T BtC T

2 U ,(−) C2 p prime 2 2
p

The lax C2 -symmetric monoidal Real topological Cyclic homology functor may be defined to be the composite
  ]
THR TCR
Alg⊗ Sp −−−−−−−→ RCycSp⊗ −−−−−−−→ Sp⊗ . ◁
Eσ C2 C2

3.5.4. Speculations on genuine equivariant THH for other groups. Merling posed the following question.
Question 3.34 ([AimPL, Prob 1.6]). Is it possible to build a version of THH for G-ring spectra which is
G-symmetric monoidal? ◁
On its face, Question 3.34 receives a positive answer by setting C = Sp⊗ in the following.
G
Corollary 3.35. Given C⊗ a G-symmetric monoidal ∞-category, there is a G-symmetric monoidal functor
Z
: Alg⊗ (C) → C ⊗
E1
S1
whose H-value functor AlgE1 (CH ) → CH is THH.

Proof. Specialize Corollary 3.22 to M = S 1 . □

This is possibly unsatisfying; it certainly recovers the notion of [MQS24], but it recovers neither twisted
nor Real THH. We can give a more general result recovering the latter as follows.
44 NATALIE STEWART

Corollary 3.36. Let C ⊗ be a G-symmetric monoidal ∞-category, V be an orthogonal G-representation, and

S ⊂ V an embedded G-set. Then, there exists a G-symmetric monoidal functor
THHS,V : Alg⊗ (C) → C ⊗ ,
EV
natural in G-symmetric monoidal strongly G-colimit preserving G-functors in C, satisfying
THHS,V (A) ≃ A ⊗A⊗S A.
Proof. Define the invariant closed topological subspace
S (S,V ) B RS ∪ {∞} ⊂ S V .
Giving S V the standard metric, pick some small ε > 0 and let τS (S,V ) ⊂ S V be the open ball around S (S,V )
of radius ε; this is canonically a V -framed open G-submanifold of S V . We define
Z
THHS,V (A) B A : Alg⊗ (C) → C ⊗ .
EV
τS (S,V )
We’re left with verifying the tensor product formula, which itself will be an application of Horev’s G-⊗-excision
result [Hor19, Prop 5.2.3]. We need a collar decomposition; define the function f : S V → [−1, 1] by



−1 |v| < 2ε;
 |V |−3ε

f (v) =  2ε ≤ |v| ≤ 4ε;
 ε

1

|v| > 4ε.
The restriction of this to τS (S,V ) is a collar decomposition with positive and negative part framed-diffeomorphic
to D(V ) and with interior V -framed diffeomorphic to the indexed disjoint union S · D(V ). ⊗-excision yields
Z Z
THHS,V (A) ≃ (S,V )
A ⊗ R
(S,V ) A (S,V )
A;
τS+ τS+ ∩τS (S,V ) τS−
Z Z
≃ A ⊗R A A
D(V ) S·D(V ) D(V )op
≃ A ⊗A⊗S Aop . □

Appendix A. Technicalities on (co)cartesian I-symmetric monoidal ∞-categories

For the duration of this appendix, we assume the notation of [Ste25, § A]. Fix P ⊂ T an atomic orbital
subcategory and I ⊂ FPT an almost-unital weak indexing category. We define the ∞-category of Γ -I-preoperads
PreOpΓI B Catint−cocart
T ,/F ,
I,∗

so that the results of [BHS22] recognize as a localizing subcategory.7

OpI ⊂ PreOpΓI
This appendix can be understood as a lift of [HA, § 2.4.1-2.4.3] to the setting of (co)cartesian I-symmetric
monoidal ∞-categories, working in the specific model of Γ -I-preoperads; we proceed by an essentially similar
strategy, complicated only by less convenient combinatorics. We suggest only readers in need of the minutiae
inquire within, and the shunt remaining readers to the summaries contained in Section 1.4.
First, define the T -1-category Γ ∗I to have V -values
 
∗ s.i.
ΓI,V B U+ −−→ S+ U ∈ T/V ⊂ Ar(FI,∗ )V ;
∗
that is, the objects of ΓI,V are pointed I-admissible V -sets with a distinguished orbit, and the morphisms of
ΓI,V preserve distinguished orbits. This possesses a target forgetful T -functor t : Γ ∗I → FI,∗ . We use this to
construct an ∞-category TotTotT C over TotFI,∗ in Appendix A.1 satisfying the following universal property.

Proposition A.1. Given C a T -∞-category, there exists an ∞-category TotTotT C I−⊔ over TotFI,∗ satisfying
the universal property that there is a natural equivalence
Fun/TotFI,∗ (D, TotTotT C I−⊔ ) ≃ Fun/T op (D ×TotFI,∗ TotΓ ∗I , TotC);
7 Here, Γ is a reference to Segal’s category Γ , whereas the undecorated version centers the effective Burnside category.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 45

Second, define the (non-full) T -subcategory ΓI× ⊂ Ar(FI,∗ ) to have V -objects given by summand inclusions
  V -sets S + ,→ S+ and morphisms of V -objects given by maps α : S+ → T+ with the property that
of pointed
−1
α T + ⊂ S + . In Appendix A.1 we prove the following.

eI−× over F satisfying the

Proposition A.2. Given C a T -∞-category, there exists an ∞-category TotTotT C I,∗
universal property that there is a natural equivalence
eI−× ) ≃ Fun/T op (K ×TotF TotΓ × , TotC).
Fun/TotFI,∗ (K, TotTotT C I,∗ I

Note that there is an equivalence

{S+ } ×FI,∗ Γ ×I ≃ PV (S),

where PV (S) is the V -poset with U -value given by subsets of ResVU S ordered under inclusion. In particular,
eI−× as V -functors PV (S)op → CV . Let TotTotT C I−× ⊂ TotTotT C
for S+ ∈ FI,∗ , we view objects in C eI−× be
S+
the full subcategory whose objects over V are spanned by those functors F : PV (S) → CV satisfying the
op

property that, for all U → V and T ⊂ ResVU S, the maps ResVW F(T ) → F(W ) exhibit F(T ) as the T -indexed
product F(T ) ≃ TW F(U ) in C.
Q
Following Appendix A.1, we construct cocartesian lifts and characterize algebras and I-symmetric
monoidal functors into C I−⊔ and C I−× in Appendices A.2 and A.3. We spell out a corollary in Appendix A.6
relating LOpI -equivalences to the Morita theory of algebraic patterns.

A.1. Quasicategories modeling C I−⊔ and C I−× .

A.1.1. The cocartesian case. Let T op be a quasicategory and TotC ∈ sSetcocart

/T a cocartesian fibration to T .
There exists a simplicial set TotTotT C I−⊔ satisfying the universal property
(16) Hom/TotFI,∗ (K, TotTotT C I−⊔ ) ≃ Hom/T op (K ×TotFI,∗ TotΓ ∗I , TotC), ∀K ∈ sSetcocart
/TotF .
I,∗

Lemma A.3. The map TotTotT C I−⊔ → TotFI,∗ is an inner fibration; hence TotC I−⊔ is a quasicategory.

Proof. The proof is exactly analogous to [HA, Prop 2.4.3.3]: apply the universal property
f0 a
Λni TotTotT C I−⊔ Λni ×FI,∗ Γ ∗I ≃ Λni TotC
U ∈Orb(S)
◦
f (U )∈Sn,+

a
∆n TotFI,∗ ∆n ×FI,∗ Γ ∗I ≃ ∆n T op
(S0,+ →···→Sn,+ )
U ∈Orb(S)
◦
f (U )∈Sn,+

to note that inner horn lifts of TotC I−⊔ → TotFI,∗ correspond with tuples of inner horn lifts along TotC → T op ,
which exist by assumption that it is a cocartesian fibration (hence an inner fibration). The remaining claim
follows by noting that TotFI,∗ is a quasicategory, so the composite map TotTotT C I−⊔ → TotFI,∗ → ∗ is an inner
fibration. □

Proof of Proposition A.1. We’ve verified that TotTotT C I−⊔ is a quasicategory over TotFI,∗ . Fixing some
quasicategory D over FI,∗ and applying Eq. (16) for K B D × ∆n , we find that Fun(D, TotTotT C I−⊔ ) ≃
Fun/T op (D ×TotFI,∗ TotΓ ∗I , TotC). The result then follows by replacing “quasicategory” with “∞-category.” □

eI−× ∈ sSet/TotF by the universal property

A.1.2. The cartesian case. Now, we define TotTotT C I,∗

(17) eI−× ) ≃ Hom/T op (K ×TotF TotΓ × , TotC),

Hom/TotFI,∗ (K, TotTotT C I,∗ I ∀K ∈ sSetcocart
/TotF .
I,∗
46 NATALIE STEWART

Recollection A.4 ([NS22, Def 2.1.2]). A morphism f in TotFI,∗ from S+ ∈ FI,∗,U to T+ ∈ FI,∗,V may be modelled
as a morphism of spans
f◦
S f −1 (T ) T
ιf
ResVU S

U V V
such that f ◦ ∈ I (c.f. Construction 1.17). Such a morphism is πFI,∗ -cocartesian if f ◦ and ιf are both
∼
equivalences, i.e. it witnesses an equivalence ResVU S+ −
→ T+ . ◁
Let f : T+ → S+ be a map in TotFI,∗ lying over an orbit map U → V and let S ⊂ S be an element
of Γ ×I
lying over S+ . We would like to construct a Cartesian edge landing  on S ⊂ S; we do so by setting
T B f −1 (ResVU S) ⊂ f −1 (ResVU S) ⊂ T , and letting the associated map t : f −1 (ResVU S) ⊂ T → S ⊂ S be the
canonical one. The following lemma then follows by unwinding definitions, where U : Γ ×I → FI,∗ denotes the
forgetful functor.
Lemma A.5. t is a U -cartesian arrow; in particular, U is a cartesian fibration.
The following lemma then follows from [HTT, Cor 3.2.2.12].

Lemma A.6. Let e eI−× → TotF be the projection and let α̃ : F → G be a C

p: C eI−× -morphism lying over a
I,∗
TotFI,∗ -morphism α : T+ → S+ lying over an orbit map U → V . Then, α̃ is e p-cocartesian if and only if, for
all T ′ ⊂ T , the induced map
F(α −1 (ResVU T ′ )) → ResVU G(T ′ )
p is a cocartesian fibration, so C
is an equivalence; in particular, e eI−× is a quasicategory.

Proof of Proposition A.2. We concluded in Lemma A.6 that TotTotT C eI−× is a quasicategory satisfying Eq. (17),
so it models an ∞-category satisfying our universal property by the same argument as Proposition A.1. □

A.2. O-comonoids and (co)cartesian rigidity. An object of TotTotT C I−⊔ may be viewed as S+ a pointed V -set
and C = (CW ) ∈ CS an S-tuple of elements of C; a morphism f : C → D may be viewed as a TotFI,∗ -map
f
(S+ → VS,+ ) −
→ (T+ → VT ,+ ) together with a collection of maps
n o
fW : IndU
W CW → DU | W ∈ f
−1
(U )

for all U ∈ Orb(T ). Unwinding the universal property for cocartesian arrows, we find the following.8
Proposition A.7. A map f : C → D is π-cocartesian if and only if {fW } witness DU as the indexed coproduct
−1 (U )
fa
∼
CW −−−−→ DU
W

for all U ∈ Orb(T ). In particular, f is inert if and only if the following conditions are satisfied:
(a) The projected morphism π(f ) : S → T is inert.
(b) The associated map Cf −1 (U ) → DU is an equivalence for all U ∈ Orb(T ).
Hence π : TotTotT C I−⊔ → TotFI,∗ presents a Γ -I-preoperad C I−⊔ .

Corollary A.8. C I−⊔ is an I-operad which is an I-symmetric monoidal ∞-category if and only if C admits
I-indexed coproducts.
Proof. It follows from Proposition A.7 that TotTotT C I−⊔ → TotFI,∗ is a cocartesian fibration if and only if C
admits I-indexed coproducts, so it suffices to verify the following conditions:

8 It is here that we use almost-unitality for the cocartesian setting; if I was not almost essentially unital, then there would exist
some S whose I -admissible orbits do not together cover S , so C I−⊔ → TotFI,∗ would not be an inert-cocartesian fibration.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 47

(b) cocartesian transport yields an equivalence

Y
CS ≃ CU ;
U ∈Orb(S)

(c) cocartesian transport yields an equivalence

Y
TU →U
 
MapTTotO
→S
⊗ (C, D) ≃ MapTotO ⊗ CU , D .
U ∈Orb(S)

In fact, each condition follows from Proposition A.7. □

Observation A.9. It follows from Proposition A.7 that the indexed tensor product functor ⊗S : CSI−⊔ → CVI−⊔ is
left adjoint to ∆S , i.e. indexed tensor products in C I−⊔ are indexed coproducts. ◁
Given O ⊗ a unital I-operad, define a diagram of Cartesian squares in CatT .
ι
O OΓ⊗ O⊗
⌟ ⌟

∗T Γ ∗I FI,∗

Note that the objects of OΓ⊗,V consist of triples (S+ , U , X) where U ∈ Orb(S) and X ∈ OS , and the image of ι is
equivalent to the triples where S ∈ T/V , hence U = S.
Further note that cocartesian transport along the inert morphism U+ ,→ S+ induces an equivalence
MapO⊗ (ιY , (S+ , U , X))) ≃ MapO⊗ (ιY , (U+ , U , XU )))
Γ ,V Γ ,V

for all Y ∈ O.9 In particular, ι witnesses O as a colocalizing T -subcategory, with colocalization T -functor
R(S+ , U , X) ≃ (U+ , U , XU ).
This interacts with Kan extensions via the following lemmas.
Lemma A.10. Suppose F : C → D is a T -functor and L : C → E is T -left adjoint to R : E → C. Then, FR is
the T -left Kan extension of F along L.
Proof. Using [Sha23, Thm 10.5] we simply repeat the nonequivariant proof: Yoneda’s lemma yields
NatT (L! F, G) ≃ NatT (F, GL)
≃ NatT (FR, G),
so another application of Yoneda’s lemma constructs a natural equivalence L! F ∼ FR. □

Lemma A.11. Fix a T -functor A : OΓ⊗ → C. Then, the following are equivalent
(a) The corresponding map O ⊗ → C I−⊔ is a functor of I-operads.
(b) For all morphisms α in OΓ⊗ whose image in O ⊗ is inert, A(α) is an equivalence in C.
(c) If f : (S+ , U , X) → (U+ , U , XU ) is a cocartesian lift of the corresponding inert morphism, then A(f ) is
an equivalence.
(d) A is T -left Kan extended from O.
Furthermore, every functor F : O → C admits a left Kan extension along O ,→ OΓ⊗ ; in particular, the forgetful
functor Alg (C) → FunG (O, C) is an equivalence.
O

Proof. (a) ⇐⇒ (b) follows immediately from Proposition A.7. (b) ⇐⇒ (c) is immediate by definition. (c)
⇐⇒ (d) and the remaining statement both follow by Lemma A.10. □

9 This utilizes unitality of O ⊗ , as we implicitly use that, for each orbit U ′ ∈ Orb(S) other than U , the space O(∅ ′ ; X ′ ) is
U U
contractible.
48 NATALIE STEWART

A.3. O-monoids. We start with the following.

Proposition A.12. TotTotT C I−× → TotFI,∗ is a cocartesian fibration, so in particular, it presents a Γ -I-preoperad
C I−× ; moreover, its straightening is an I-symmetric monoidal ∞-category if and only if C admits I-indexed
products, in which case the indexed tensor functors in C are indexed products.
Proof. For the first statement, it suffices to observe that the cocartesian arrows described in Lemma A.6 lie
in TotTotT C I−× . For the second, note by unwinding definitions that cocartesian transport induces a fully
faithful functor Y
CS → CU
U ∈Orb(S)

Moreover, this is essentially surjective if and only if C admits I-indexed products, as desired. □

We organize ourselves around the following observations.

Observation A.13. The projection TotTotT O ⊗ ×TotFI,∗ TotΓ ×I → TotTotT O ⊗ admits a left adjoint L sending
X ∈ OS⊗+ to (X, S ⊂ S); the unit map of this adjunction is evidently an equivalence, so L : TotTotT O ⊗ →
TotTotT O ⊗ ×FI,∗ Γ ×I is fully faithful. ◁
Observation A.14. The section L : TotFI,∗ → TotΓ ×I induces a natural transformation K ≃ K ×TotFI,∗ TotFI,∗ →
K ×TotF TotΓ × , which induces a natural transformation TotTotT C
I,∗ I
eI−× → TotC under Yoneda’s lemma. Un-
eI−× → C.
winding definitions using Lemma A.6, this presents a T -functor TotT C ◁
ϕ
Given a T -functor TotT O⊗ eI−× , we acquire a corresponding functor
−−−−−→ TotT C
L ϕ′
TotTotT O ⊗ −−−−→ TotT O ⊗ ×TotFI,∗ TotΓ ×I −−−−−→ TotC
over T op . Now, the following observation is important.
Observation A.15. The description of cocartesian arrows of Lemma A.6 and Proposition A.12 implies that ϕ ′
and L are unstraightened from T -functors. ◁
Now, given a Γ -I-preoperad O ⊗ , we say that an O-monoid in C is a T -functor TotT O ⊗ → C satisfying
the condition that, for all X ∈ CS , the maps ResVU F(X) → F(XU ) induced by cocartesian transport witness
F(X) as the indexed product
YS
F(X) ≃ F(XU ).
U
Proposition A.16. Fix C a T -∞-category and O ⊗ a Γ -I-preoperad. Then, the postcomposition functor
AlgO (C I−× ) → FunT (TotT O ⊗ , C) is fully faithful with image spanned by the O-monoids.
Lemma A.17. The following conditions are equivalent:
(a) ϕ is a map of Γ -I-preoperads.
(b) For all morphisms α in TotTotT O ⊗ ×TotFI,∗ TotΓ ×I whose image in O ⊗ is inert ϕ ′ (α) is an equivalence..
(c) If f : (S + → V+ , S, F, X) → (S+ → V+ , S, F, X) is a cocartesian lift of the corresponding inert morphism,
then ϕ(f ) is an equivalence.
ϕ
(d) The composite TotT O ⊗ −−−−→ TotT C eI−× → C is homotopic to ϕ ′ .
ϕ
eI−× → C is T -right Kan extended from ϕ ′ along L.
(e) The composite TotT O ⊗ −−−−→ TotT C
Proof. Lemma A.6 immediately implies that (a) ⇐⇒ (b) ⇐⇒ (c) ⇐⇒ (d). (d) ⇐⇒ (e) follows from
Lemma A.10. □

The following lemma follows by unwinding definitions.

Lemma A.18. Suppose ϕ is a functor of Γ -I-preoperads. Then, the following conditions are equivalent:
(a) ϕ factors through a Γ -I-preoperad map ϕ : O ⊗ → C I−× .
(b) ϕ ′ is an O-monoid.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 49

Proof of Proposition A.16. Since TotT C I−× ,→ TotT C̃ I−× is fully faithful, the first of the following functors is
fully faithful      
AlgO C I−× ,→ AlgO C eI−× ,→ FunT TotT O ⊗ , C .
By Lemma A.17, the second is fully faithful, and by Lemma A.17 the image of the composite functor consists
of the O-mononids. □
We may additionally characterize I-symmetric monoidal functors via a lift of [HA, Prop 2.4.1.7], which
also follows immediately from Lemma A.6.
Lemma A.19. Suppose C has I-indexed products, TotT O ⊗ → FI,∗ is a cocartesian fibration, and Lemma A.18
is satisfied. Then, the following conditions are equivalent.
(a) ϕ is a T -functor of cocartesian fibrations over FI,∗ .
(b) If f : D → D′ is an active arrow in TotTotT O ⊗ , then map ϕ ′ (f ) is an equivalence.
Now, we will also lift [HA, Prop 2.4.1.6]. We have a fully faithful T -functor ι : O ,→ TotT O ⊗ .
Lemma A.20. Suppose ϕ satisfies the conditions of Lemma A.19 and the action maps ⊗S : OS → OV are right
adjoint to the restriction maps ∆S : OV → OS . Then, the functor ϕ e : TotT O ⊗ → C is right Kan-extended from
⊗
the I-product-preserving functor O → TotT O → C along ι.
NS
Proof. The assumptions imply that there is a right T -adjoint R : TotT O ⊗ → O to ι, sending (XU ) 7→ U XU ≃
QS
U XU . The O-monoid assumption shows that ϕ e ∼ ϕ ◦ ι ◦ R, which shows that ϕe is right Kan extended from
ϕ ◦ ι along ι; moreover, the O-monoid assumption shows that ϕ ◦ ι is I-product-preserving. □
A.4. (Co)cartesian rigidity.
Proposition A.21. Suppose O ⊗ is an I-symmetric monoidal ∞-category such that ⊗S : OS → OV is right
adjoint to ∆S : OV → OS for all S ∈ FI,V . Then, the forgetful functor
 
U : Fun⊗ ⊗ I−×
I O ,C → FunT (O, C)
is fully faithful with essential image spanned by the I-product preserving T -functors.
g (TotT O ⊗ , C) ⊂ Fun (TotT O ⊗ , C) be the equivalent image of Fun⊗ (O ⊗ , C ⊗ ). Lemma A.20
Proof. Let Fun T T I
constructs the solid portion of a diagram
ι∗

FunI−×
T (O, C)
g T (TotT O ⊗ , C)
Fun FunT (TotT O ⊗ , C)
ι∗ U ι∗

FunI−×
T (O, C) FunT (O, C)
It suffices to verify that the dashed arrow exists, i.e. right Kan extensions of I-product-preserving functors
along ι satisfy the conditions of Lemma A.19; but this follows by unwinding definitions. □

 A.22. IfvopC has I-indexed products, then there exists a unique I-symmetric monoidal equivalence
Corollary
C I−× ≃ (C vop )I−⊔ lying over the equivalence C ≃ (C vop )vop ; if C has I-indexed coproducts, then there exists
 vop
a unique I-symmetric monoidal equivalence C I−⊔ ≃ (C vop )I−× lying over the equivalence C ≃ (C vop )vop .
Proof. By conservativity of the underlying category (see Lemma 1.11), it suffices to construct a unique
I-symmetric monoidal functor lying over the identity in each case. For the first case, by Proposition A.21 it
vop
suffices to note that (C vop )I−⊔ has ∆S ⊣ ⊗S . The second case follows from the first under the following
equivalence of arrows, where D B C vop .
 vop 
vop )I−⊔ , D I−× vop ≃ Fun⊗ (D vop )I−⊔ vop , D I−× vop
        vop 
Fun⊗ (D ≃ Fun⊗ (D vop )I−⊔ , DI−×
I I I

   
FunT (Dvop , Dvop ) ≃ FunT (Dvop )vop , (Dvop )vop ≃ FunT D, (Dvop )vop
50 NATALIE STEWART

Corollary A.23. Suppose O ⊗ is an I-symmetric monoidal ∞-category such that ⊗S : OS → OV is left to

∆S : OV → OS for all S ∈ FI,V . Then, the forgetful functor
 
U : Fun⊗I O ⊗ I−⊔
, C → FunT (O, C)
is fully faithful with essential image spanned by the I-coproduct preserving T -functors.
Proof. This follows by taking vertical opposites of Proposition A.21 in light of Corollary A.22. □

We are now ready to prove our main generalization of Theorem A’ (see p. 21).

Proof of Theorem A’. We begin with the cartesian cases. To see that (−)I−× is fully faithful, it suffices to
combine Propositions A.12 and A.21. The compatibility with U is obvious, and the description of the
image follows immediately from Proposition A.21. The cocartesian case follows by the same argument using
Corollary A.23. □

A.5. Wirtmüller maps. Suppose O ⊗ is an I-symmetric monoidal ∞-category and C has I-indexed coproducts.
The equivalence AlgO (C I−⊔ ) ≃ FunT (O, C) embeds I-symmetric monoidal functors O ⊗ → C I−⊔ as a full
subcategory of T -functors O → C. We now record the property identifying this full subcategory.
Observation A.24. Let F : D⊗ → E ⊗ be a lax I-symmetric monoidal functor. Then, the universal property for
cocartesian arrows constructs, for each active arrow IndTV S → V , an arrow
S
 S 
O O 
µS : F(−) =⇒ F  −
U U
such that F is I-symmetric monoidal if and only if µS is an equivalence for all S ∈ FI . In particular, in the
case of the lax I-symmetric monoidal functor F : O ⊗ → C I−⊔ classified by a functor G : O → C, the arrow µ
has the type
S
 S 
a O 
G(−) =⇒ G  − ;
U U
moreover, unwinding definitions, in the case that O⊗ =C⊗ is an I-symmetric monoidal structure on C
and, G is the identity, and 1• ∈ Γ υ(I) O is initial, this map is precisely the ⊗-Wirthmüller map constructed
in Construction 1.62. In` particular, we’ve observed that the identity classifies an I-symmetric monoidal
∼
equivalence C ⊗ −−−−−→ C I− if and only if C ⊗ has I-admissible ⊗-Wirthmüller isomorphisms. ◁

A.6. A technical corollary on n-Morita equivalences. A Segal morphism of algebraic patterns ϕ : O → P is

called an n-Morita equivalence if, for all complete (n + 1)-categories C, the induced functor
f ∗ : SegP (C) → SegO (C)
is an equivalence; in fact, it suffices to check this in the case C = S≤n [Bar23, Prop 2.1.9]. We have the
following corollary.
Corollary A.25. Suppose ϕ : P ⊗ → O ⊗ is a morphism of Γ -I-preoperads such that the induced T -functor
U P → U O is essentially surjective. Then, ϕ is an n-Morita equivalence if and only if the associated map of
T -operads hn+1 LOpT P ⊗ → hn+1 LOpT O ⊗ is an equivalence.

Proof. There is an equivalence SegO (C) ≃ AlgO (C I−× ) ≃ AlgLOp O (C

I−× ) natural in Segal morphisms over
T
Span(FT ), so the result follows from the recognition result for n-equivalences of I-operads [Ste25]. □

Now, we define the I-preoperads PreOpI B Catint−cocart

T ,/ Span (F ) . In [Ste25, § A.1] we proved that, under the
I T
assumption that O ⊗ is an I-operad, the canonical strong Segal morphism s∗ O ⊗ → O ⊗ is a Morita equivalence
and s∗ O ⊗ is soundly extendable; in fact, we only used that O ⊗ is an I-operad to conclude that s∗ O ⊗ is soundly
extendable, and the same proof shows that s∗ O ⊗ → O ⊗ underlies a natural Morita equivalence when O ⊗ is an
I-preoperad, and in particular, a natural n-Morita equivalence.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 51

Moreover, this underlies the counit map of I-preoperads ε : s! s∗ O ⊗ → O ⊗ , which is an LOpG -equivalence.
In particular, ε becomes an equivalence after Alg(−) (S G,≤n+1 ). This yields a chain of natural equivalences.
AlgO (S G,≤n+1 ) ≃ Algs! s∗ O (S G,≤n+1 ) ≃ Algs∗ O (S G,≤n+1 ) ≃ Segs∗ O (S≤n+1 ) ≃ SegO (S≤n+1 )
In particular, the same proof as Corollary A.25 yields the following.
Corollary A.26. Suppose ϕ : P ⊗ → O ⊗ is a morphism of I-preoperads such that the induced T -functor
U P → U O is essentially surjective. Then, ϕ is an n-Morita equivalence if and only if the associated map of
T -operads hn+1 LOpT P ⊗ → hn+1 LOpT O ⊗ is an equivalence.

Appendix B. I-operadic disintegration and assembly

In this appendix, we assume familiarity with the minutiae of [BHS22; Ste25] and prove Corollary 3.8.

B.1. The algebraic pattern for families of I-operads. Fix C a T -∞-category.

Construction B.1. The totally inert pattern structure on C has
 el  int  act
Totint C B Totint C B TotC; Totint C B (TotC)≃ .
We define the C-family pattern
 
Totint C × FI,∗ B TotC int ×T op,int TotFI,∗ . ◁
The following observation is as crucial as it is immediate.
Observation B.2. If C is a T -space, then Totint (C × FI,∗ ) → TotFI,∗ is an inert-cocartesian fibration, and the
domain has the induced pattern structure. In particular, in this case, Totint (C × FI,∗ ) is the pattern underlying
the Γ -I-preoperad C × FI,∗ . ◁
We begin by identifying C-indexed diagrams of I-operads.
Proposition B.3. There exists a natural equivalence
    
Fbrs Totint C × FI,∗ ≃ FunT C, Op ,
I
so that when C ≃ ∗, this is the usual equivalence Fbrs(FI,∗ ) ≃ Op .
I

To prove this, we use the equifibered theory, focusing on the following lemmas.
 
Lemma B.4. There exists a natural equivalence SegTot(C×F ) (D) ≃ FunT C, CMonI (D) .
I,∗

Proof. First off, we get an embedding

SegTot(C×F ) (D) ⊂ Fun(Tot(C × FI,∗ ), D);
I,∗

≃ FunT (C × FI,∗ , CoeffT (D));

  
≃ FunT C, FunT FI,∗ , CoeffT D
 
characterized by the Segal condition that the restricted functor Tot C × Fint
I,∗ → D is right Kan extended from
 
el
Tot C × FI,∗ . Now, unwinding conditions, this corresponds with the condition that the value V -functors
V
Fint
IV ,∗ → Coeff D are each right-Kan extended from Fint
IV ,∗ , i.e. the corresponding functor factors through
 
T
CMonI (D) ⊂ FunT FI,∗ , Coeff D (c.f. [Nar17, Ex 1.17]). □

Lemma B.5. EnvTot(C × FI,∗ ) corresponds naturally with the constant T -functor over FI−⊔
I,∗ ; a natural trans-
formation F → Env(C × FI,∗ ) is equifibered if and only if it is pointwise-equifibered.
Proof. This follows by explicitly identifying the active arrows in Tot(C × FI,∗ ) as products of equivalences and
active arrows of TotFI,∗ . □
52 NATALIE STEWART

Proof of Proposition B.3. The above work constructs a string of natural equivalences
 pointwise−equifibered
Fbrs(Tot(C × FI,∗ )) ≃ FunT C, CMonI (D)
/∆FI,∗
 
equifibered
≃ FunT C, CMonI (D)/F
I,∗
 
≃ FunT C, Op . □
I

Now, this is closely related to (−)I−⊔ , as described by the following construction.

Construction B.6. Fix C a T -∞-category. Then, pullback along the projection (−) ×TotFI,∗ TotΓ ∗I → (−)
determines a natural transformation
     
Fun/TotFI,∗ −, TotTotT C I−⊔ ≃ Fun/T op − ×TotFI,∗ TotΓ ∗I , C ← Fun/T op (−, C) ≃ Fun/TotFI,∗ −, C × TotFI,∗ ,

which corresponds with a functor γ : TotC × TotFI,∗ → TotTotT C I−⊔ under Yoneda’s lemma. Note that
γ(C, S) ≃ ∆S C. Moreover, this is compatible with Γ -I-preoperadic structure in the case C = TotT P ⊗ :
id ×πP
FI,∗ × TotT P ⊗ FI,∗ × FI,∗
γ ∧
I−⊔
 I−⊔
(TotT P ⊗) πPI−⊔ FI,∗ ∨ FI,∗
pr1
π
FI,∗

Pulling back to SpanI (FV ), we acquire a simpler diagram

id ×πP
SpanI (FV ) × TotP ⊗ SpanI (FT ) × SpanI (FT )
γ ∧
I−⊔
(TotV P ⊗) ρ SpanI (FT )
pr1 π
SpanI (FV )

so in particular, γ becomes a bifunctor under the alternative structure map ρ. ◁

In this paper, we mostly care about the case that C is a generic T -space. We will use the following
specialization of Barkan’s morita equivalence recognition result (c.f. [HA, Thm 2.3.3.23, Thm 2.3.3.26]).10
Proposition B.7 ([BHS22, Prop 3.1.16, Thm 5.1.1]). Suppose f : O ⊗ → P ⊗ is a strong Segal morphism of
Γ -I-preoperads such that P ⊗ presents an I-operad and the following conditions hold:
(a) the T -functor U O → U P is an equivalence, and 
act ≃
 ≃
(b) for every O ∈ U O, the map of spaces O/O → P/fact(O) is an equivalence.
Then, the functors f ∗ : MonP (C) → MonO (C) and f ∗ : OpI,/P ⊗ → Fbrs(O) are equivalences.
 
Proposition B.8. When X is a T -space, the functor γ : Totint X × TotFI,∗ → TotTotT X I−⊔ satisfies the
conditions of Proposition B.7; in particular, γ is an LOpI localization map.
Proof. By unwinding definitions we find that Totγ is an iso-segal morphism, and in particular it is a strong
Segal morphism. Moreover, condition (a) follows simply by unwinding definitions.
 act
For condition (b), note that the elements of TotI X I−⊔ correspond with maps f : T → S in
/γ(x,S)
FT together with elements (yU )T ∈ X T with distinguished paths yU ∼ x within X U ; in particular, by

10 To see this as a specialization of Barkan’s result, note that by [Ste25], TotF

I,∗ is soundly extendable, so TotT of an I -operad is
soundly extendable by [BHS22, Lem 4.1.15]. The remaining modifications necessary are the observation that O el ≃ TotU O (so condition
(1) implies condition (1) of [BHS22, Thm 5.1.1]) and the identifications Fbrs(P ) ≃ OpI,/P ⊗ of [Ste25] and [BHS22, Cor 4.1.17] as well as
MonP (C) ≃ SegP (C) of Corollary 1.54.
 ON TENSOR PRODUCTS WITH EQUIVARIANT COMMUTATIVE OPERADS 53

 act ≃
contracting paths, we may construct a deformation retract of TotI X I−⊔ onto the subspace F≃
I,/S ⊂
/γ(x,S)
 act ≃
TotI X I−⊔ of identity paths.
/γ(x,S)
 act ≃
Similarly, we may perform a deformation retract of X × Fi,∗ onto the summand F≃ I,/S ⊂
/(x,S)
 act  ≃
X × Fi,∗ of identity paths. It follows by unwinding definitions that these are taken isomorphically onto
/(x,S)
each other; alternatively, one may note that the induced endomorphism of F≃
Q `
I,/S ≃ U ∈Orb(S) T ∈FI,U B AutU (T )
is a product of coproducts of maps classified by torsor maps AutU (T ) → AutU (T ), which are automatically
isomorphisms, implying that our map of 1-truncated spaces is an isomorphism on π0 and on π1 at all
basepoints. □

Warning B.9. A closely related analog of Proposition B.8 is claimed in [HA, Rmk 2.4.3.6] in the case T = ∗
without the assumption that C is a space; as pointed out in [KK24, Rmk 2.3] Lurie’s claim (and hence proof)
is incorrect in general, but the claim was verified in loc. cit. when C is a space. ◁
We finish with the following proposition.
 
Proposition B.10. If TotT P ⊗ → Totint X × FI,∗ is a fibrous pattern, then LOpI P ⊗ is the T -colimit of the
T -functor X → Op associated with P ⊗ .
I
∗
Proof. Note that LOpI γ! πX I−⊔ ! ⊣ γ ∗ πX I−⊔ , and the latter is equivalent to ∆ : OpI → FunT (X, Op ); the above
I
presentation for the left adjoint is LOpI P ⊗ , and indexed colimits are also left adjoint to ∆, so the claim follows
from uniqueness of left adjoints. □

We now apply this in the language of disintegration and assembly.

B.2. Disintegration and assembly. Given X ∈ ST and O ⊗ ∈ OpI,/X I−⊔ , define the pullback Γ -I-preoperad
α
disI (O ⊗ ) TotT O ⊗
⌟

γ
X × FI,∗ TotI X I−⊔

We refer to disI (−) as the disintegration functor and α as the assembly map.
Proposition B.11. α is an LOpI -localization map.

Proof. We verified in [Ste25, § A] that the conditions of Proposition B.7 are pullback-stable, so α is a Morita
equivalence; by Corollary A.25 it is then an LOpI -equivalence. By assumption, TotT O ⊗ is LOpI -local, proving
the proposition. □

Proposition B.12. O ⊗ is the T -colimit of the T -functor X → Op associated with disI (O ⊗ ).

I
Proof. This is a straightforward application of Propositions B.10 and B.11. □

We spell out the following corollary, which summarizes the full power of what we’ve proved.
I−⊔
Corollary B.13. Let X be a T -space. The assignment x 7→ Ox B ResTV O ⊗ ×ResT X I−⊔ NI∞ , yields an equivalence
V

Op ≃ FunT (X, Op ).
I,/X I−⊔ I
The unit of this equivalence specifies a natural equivalence.
O ⊗ ≃ colimx∈X Ox⊗ .
Proof. The first claim follows by combining Observation B.2 and Propositions B.3, B.7 and B.8. The
remaining claim is proved identically to Proposition B.12. □
54 NATALIE STEWART

Appendix C. Algebraic patterns and the I-symmetric monoidal structure on overcategories

In this appendix, we repeat the arguments of [HA, § 2.2.2] in the setting of algebraic patterns, assuming
familiarity with the minutiae of [HA, § 2.2.2] and of [BHS22].

C.1. The fibrous pattern case. We fix O an algebraic patter and make the following temporary definitions.
Definition C.1. An O-monoidal ∞-category is a fibrous pattern C → O which is also a cocartesian fibration.
If C → O is a fibrous pattern, then the ∞-category of O-algebras in C is
Alg/O (C) B Funint−cocart
/O (O, C). ◁
Remark C.2. When O is sound, [BHS22, Prop 4.1.7] shows that O-monoidal ∞-categories are synonymous
with Segal fibrations to O in the sense of [CH21]. ◁
⊗
Warning C.3. If O underlies a T -operad, AlgO (C) and Alg/O (C) should not be confused; the latter consists
of algebras over O ⊗ . Nevertheless, in the case that O ⊗ → Comm⊗ ⊗
T is a monomorphism (i.e. O is a weak
N∞ -operad), these agree. ◁
For the duration of this appendix, we fix q : C → O a fibrous pattern and p : K → Alg/O (C) a K-indexed
diagram of O-algebras in C. Let q′′ : C/pO −→ O be the construction made in [HA, Def 2.2.2.1], interpreted
as a functor of ∞-categories via [HA, Lem 2.2.2.6]. The proof of the following theorem will involve essentially
no new ideas over that of [HA, Thm 2.2.2.4].
Theorem C.4. q′′ exhibits C/pO as a fibrous O pattern, which is an O-monoidal ∞-category if C is.
Proof. We apply [HA, Lem 2.2.2.7-9] on opposite categories. In particular, given an arrow g : X → q′′ (Y ) in
O over which C has a cocartesian lift, [HA, Lem 2.2.2.8] supplies a lift
{Y }
∗ C/pO
g q′′

∗▷ g O

such that g is a q-colimit diagram; moreover, [HA, Lem 2.2.2.7] guarantees that g is a q′′ -cocartesian lift of g.
Since C has inert-cocartesian lifts, so does C/pO , and when q is a cocartesian fibration, so is q′′ .
We’re left with verifying the Segal condition(s) for fibrous patterns; we use that of [BHS22, Prop 4.1.6].
As in the proof of [HA, Thm 2.2.2.4], it follows from a simple application of [HA, Lem 2.2.2.9] that each of
the relevant diagrams are limit diagrams, as they project to limit diagrams in C. □

Remark C.5. As in [HA, Thm 2.2.2.4.(2)], it follows from the above diagram that, given an arrow f in C/pO
such that C admits a q′′ (f )-cocartesian arrow, f is q′′ -cocartesian if and only if its image in C is q-cocartesian.
In particular the inert arrows in C/pO are the preimages of the inert morphisms of C, and if C is O-monoidal,
then the cocartesian arrows in C/pO are the preimages of cocartesian arrows in C. ◁
It is worthwhile to explicitly record following immediate corollary of Remark C.5, in part because it
establishes the “pointwise” nature of the coherences for slice O-monoidal structure.
Corollary C.6. If C is an O-monoidal ∞-category, then the unslicing functor C/pO → C is an O-monoidal
functor, i.e. it is a functor of cocartesian fibrations over O.
As claimed in [HA], we may pass through opposite categories to establish the following result about un-
dercategories without additional argument, noting that the additional assumption comes from the asymmetric
assumptions of [HA, Lem 2.2.2.7, 8]. Let q′ : CpO/ → O be the construction of [HA, Def 2.2.2.1].
Theorem C.7. q′ exhibits CpO/ as a fibrous O-pattern; moreover, if C is an O-monoidal ∞-category and the
value functors O → C are all O-monoidal functors, then q′ exhibits CpO/ as an O-monoidal ∞-category.
C.2. The I-symmetric monoidal case. We explicitly specialize Theorem C.4 to O B SpanI (FT ).
Corollary C.8. Let C ⊗ be an I-symmetric monoidal ∞-category and A ∈ CAlgI (C) an I-commutative algebra
⊗
in C. Then, Theorem C.4 supplies an I-symmetric monoidal ∞-category C/A such that
 REFERENCES 55

⊗
(1) The underlying T -∞-category of C/A agrees with Shah’s slice T -∞-category C/(A,T op ) [Sha22]; moreover,
if N∗V ⊂ FI,V , then the induced symmetric monoidal structure on C ⊗ T agrees with Lurie’s with
/A,ResV A
respect to the restricted E∞ -algebra ResTV
∈ CAlg(CV ).  
⊗
(2) The S-indexed tensor functor in C/A takes a tuple of maps fU : XU → ResTU A to the map
S
S NS S
U fU µ
O O
XU −−−−−−−−−→ ResTU A −−−−→ ResTV A.
U U

Proof. (1) is functoriality of the relative slice construction with respect to pullback of the base ∞-category;
this follows straightforwardly from the defining universal property.
For (2), we may apply the universal property of [HA, Def 2.2.2.1] along the functor ∆1 → SpanI (FT ) clas-
sifying an I-admissible active arrow ψ : IndTV S = IndTV S → V : active arrows lying over ψ are in correspondence
with dashed arrows (and homotopies) making the following diagram commute.

∆1 ∆1 × ∗◁ ∆1
ψ ψ

SpanI (FT ) C⊗ π SpanI (FT )

{A}

That is, active arrows over ψ correspond with commuting diagrams

(XU ) Y

(ResTU A) ResTV A

(with homotopies witnessing commutativity, between the bottom arrow and the active arrow (ResTU A) →
ResTV A, and between the underlying arrows in SpanI (FT ) and ψ). To compute the S-indexed tensor functor,
we are tasked with exhibiting cocartesian active arrows, and by Remark C.5 it suffices to construct an active
arrow whose top arrow is cocartesian. Indeed, the outer diagram of the following suffices.
S
N S
N
(XU ) XU XU
U U

S
(ResTU A) ResTU A ResTV A
N
U

References
[AimPL] AimPL: Equivariant techniques in stable homotopy theory. url: http : / / aimpl . org /
equivstable (cit. on p. 43).
[BH21] Tom Bachmann and Marc Hoyois. “Norms in motivic homotopy theory”. In: Astérisque 425
(2021), pp. ix+207. issn: 0303-1179,2492-5926. doi: 10.24033/ast. url: https://arxiv.
org/abs/1711.03061 (cit. on p. 4).
[Bar23] Shaul Barkan. Arity Approximation of ∞-Operads. 2023. arXiv: 2207.07200 [math.AT] (cit.
on p. 50).
[BHS22] Shaul Barkan, Rune Haugseng, and Jan Steinebrunner. Envelopes for Algebraic Patterns. 2022.
arXiv: 2208.07183 [math.CT] (cit. on pp. 2, 3, 8, 9, 14, 15, 19, 33, 44, 51, 52, 54).
[BS24a] Shaul Barkan and Jan Steinebrunner. Segalification and the Boardmann-Vogt tensor product.
2024. arXiv: 2301.08650 [math.AT] (cit. on pp. 9, 33, 34).
[BS24b] Shaul Barkan and Jan Steinebrunner. The equifibered approach to ∞-properads. 2024. arXiv:
2211.02576 [math.AT]. url: https://arxiv.org/abs/2211.02576 (cit. on p. 2).
56 REFERENCES

[Bar14] C. Barwick. Spectral Mackey functors and equivariant algebraic K-theory (I). 2014. arXiv:
1404.0108 [math.AT] (cit. on pp. 12, 13).
[Bar15] Clark Barwick. “Multiplicative structures on algebraic K-theory”. In: Doc. Math. 20 (2015),
pp. 859–878. issn: 1431-0635,1431-0643. arXiv: 1304.4867 [math.AT] (cit. on p. 4).
[BDGNS16] Clark Barwick, Emanuele Dotto, Saul Glasman, Denis Nardin, and Jay Shah. Parametrized
higher category theory and higher algebra: A general introduction. 2016. arXiv: 1608.03654
[math.AT] (cit. on pp. 2, 4).
[BGS20] Clark Barwick, Saul Glasman, and Jay Shah. “Spectral Mackey functors and equivariant
algebraic K-theory, II”. In: Tunisian Journal of Mathematics 2.1 (Jan. 2020), pp. 97–146. issn:
2576-7658. doi: 10.2140/tunis.2020.2.97. url: https://arxiv.org/abs/1505.03098
(cit. on p. 13).
[BH18] Andrew Blumberg and Michael Hill. “Incomplete Tambara functors”. In: Algebraic & Geo-
metric Topology 18 (Mar. 2018), pp. 723–766. issn: 1472-2747. doi: 10.2140/agt.2018.18.
Segalnumber={2}. url: https://arxiv.org/abs/1603.03292 (cit. on pp. 10, 12).
[BH15] Andrew J. Blumberg and Michael A. Hill. “Operadic multiplications in equivariant spectra,
norms, and transfers”. In: Adv. Math. 285 (2015), pp. 658–708. issn: 0001-8708,1090-2082. doi:
10.1016/j.aim.2015.07.013. url: https://arxiv.org/abs/1309.1750 (cit. on pp. 2, 7,
10).
[Bon19] Peter Bonventre. “The genuine operadic nerve”. In: Theory Appl. Categ. 34 (2019), pp. 736–780.
issn: 1201-561X. url: https://arxiv.org/abs/1904.01465 (cit. on pp. 10, 23, 37).
[BP21] Peter Bonventre and Luís A. Pereira. “Genuine equivariant operads”. In: Adv. Math. 381 (2021),
Paper No. 107502, 133. issn: 0001-8708,1090-2082. doi: 10.1016/j.aim.2020.107502. url:
https://arxiv.org/abs/1707.02226 (cit. on p. 10).
[CSY20] Shachar Carmeli, Tomer M. Schlank, and Lior Yanovski. Ambidexterity and Height. 2020.
arXiv: 2007.13089 [math.AT]. url: https://arxiv.org/abs/2007.13089 (cit. on pp. 9,
27).
[CH21] Hongyi Chu and Rune Haugseng. “Homotopy-coherent algebra via Segal conditions”. In: Adv.
Math. 385 (2021), Paper No. 107733, 95. issn: 0001-8708,1090-2082. doi: 10.1016/j.aim.
2021.107733. url: https://arxiv.org/abs/1907.03977 (cit. on pp. 8, 13, 14, 54).
[CHLL24a] Bastiaan Cnossen, Rune Haugseng, Tobias Lenz, and Sil Linskens. Homotopical commutative
rings and bispans. 2024. arXiv: 2403.06911 [math.CT] (cit. on p. 3).
[CHLL24b] Bastiaan Cnossen, Rune Haugseng, Tobias Lenz, and Sil Linskens. Normed equivariant ring
spectra and higher Tambara functors. 2024. arXiv: 2407 . 08399 [math.AT]. url: https :
//arxiv.org/abs/2407.08399 (cit. on pp. 4, 6, 10, 22).
[CLL24] Bastiaan Cnossen, Tobias Lenz, and Sil Linskens. Parametrized higher semiadditivity and the
universality of spans. 2024. arXiv: 2403.07676 [math.AT] (cit. on pp. 5, 8, 9, 11, 13, 23, 24).
[CLR25] Bastiaan Cnossen, Tobias Lenz, and Maxime Ramzi. Universality of Barwick’s unfurling
construction. 2025. arXiv: 2502.18278 [math.AT]. url: https://arxiv.org/abs/2502.
18278 (cit. on pp. 10, 22).
[Cra11] James Cranch. Algebraic theories, span diagrams and commutative monoids in homotopy theory.
2011. arXiv: 1109.1598 [math.AT]. url: https://arxiv.org/abs/1109.1598 (cit. on p. 3).
[Die09] Tammo tom Dieck. Representation theory. 2009. url: https://ncatlab.org/nlab/files/
tomDieckRepresentationTheory.pdf (cit. on p. 4).
[Dre71] Andreas W. M. Dress. Notes on the theory of representations of finite groups. Part I: The
Burnside ring of a finite group and some AGN-applications. With the aid of lecture notes,
taken by Manfred Küchler. Universität Bielefeld, Fakultät für Mathematik, Bielefeld, 1971,
iv+158+A28+B31 pp. (loose errata) (cit. on p. 4).
[Elm83] A. D. Elmendorf. “Systems of Fixed Point Sets”. In: Transactions of the American Mathematical
Society 277.1 (1983), pp. 275–284. issn: 00029947. url: https://people.math.rochester.
edu/faculty/doug/otherpapers/elmendorf-fixed.pdf (visited on 04/22/2023) (cit. on
p. 4).
[FN62] Edward Fadell and Lee Neuwirth. “Configuration spaces”. In: Math. Scand. 10 (1962), pp. 111–
118. issn: 0025-5521,1903-1807. doi: 10.7146/math.scand.a- 10517. url: https://www.
mscand.dk/article/download/10517/8538 (cit. on p. 37).
 REFERENCES 57

[Gep19] David Gepner. An Introduction to Higher Categorical Algebra. 2019. arXiv: 1907 . 02904
[math.AT] (cit. on p. 11).
[GGN15] David Gepner, Moritz Groth, and Thomas Nikolaus. “Universality of multiplicative infinite loop
space machines”. In: Algebr. Geom. Topol. 15.6 (2015), pp. 3107–3153. issn: 1472-2747,1472-
2739. doi: 10.2140/agt.2015.15.3107. url: https://arxiv.org/pdf/1305.4550 (cit. on
pp. 3, 24).
[GM11] Bertrand Guillou and J. P. May. Models of G-spectra as presheaves of spectra. 2011. arXiv:
1110.3571 [math.AT] (cit. on p. 6).
[GM17] Bertrand J. Guillou and J. Peter May. “Equivariant iterated loop space theory and permutative
G-categories”. In: Algebr. Geom. Topol. 17.6 (2017), pp. 3259–3339. issn: 1472-2747. doi:
10.2140/agt.2017.17.3259. url: https://arxiv.org/abs/1207.3459 (cit. on p. 37).
[GW18] Javier J. Gutiérrez and David White. “Encoding equivariant commutativity via operads”. In:
Algebr. Geom. Topol. 18.5 (2018), pp. 2919–2962. issn: 1472-2747,1472-2739. doi: 10.2140/
agt.2018.18.2919. url: https://arxiv.org/pdf/1707.02130.pdf (cit. on p. 10).
[HHKWZ24] Jeremy Hahn, Asaf Horev, Inbar Klang, Dylan Wilson, and Foling Zou. Equivariant nonabelian
Poincaré duality and equivariant factorization homology of Thom spectra. 2024. arXiv: 2006.
13348 [math.AT]. url: https://arxiv.org/abs/2006.13348 (cit. on p. 6).
[Hau23] Rune Haugseng. An allegendly somewhat friendly introduction to ∞-operads. 2023. url:
https://folk.ntnu.no/runegha/iopd.pdf (cit. on p. 11).
[HHLN23] Rune Haugseng, Fabian Hebestreit, Sil Linskens, and Joost Nuiten. “Two-variable fibrations,
factorisation systems and ∞-categories of spans”. In: Forum Math. Sigma 11 (2023), Paper
No. e111, 70. issn: 2050-5094. doi: 10.1017/fms.2023.107. url: https://arxiv.org/abs/
2011.11042 (cit. on p. 13).
[HK24] Rune Haugseng and Joachim Kock. “∞-operads as symmetric monoidal ∞-categories”. In: Publ.
Mat. 68.1 (2024), pp. 111–137. issn: 0214-1493,2014-4350. doi: 10.5565/publmat6812406.
arXiv: 2106.12975 [math.CT] (cit. on p. 2).
[HHR16] M. A. Hill, M. J. Hopkins, and D. C. Ravenel. “On the nonexistence of elements of Kervaire
invariant one”. In: Ann. of Math. (2) 184.1 (2016), pp. 1–262. issn: 0003-486X. doi: 10 .
4007/annals.2016.184.1.1. url: https://people.math.rochester.edu/faculty/doug/
mypapers/Hill_Hopkins_Ravenel.pdf (cit. on p. 4).
[Hil17] Michael A. Hill. Equivariant chromatic localizations and commutativity. 2017. arXiv: 1708.
03017 [math.AT]. url: https://arxiv.org/abs/1708.03017 (cit. on pp. 2, 37).
[Hil22] Michael A. Hill. “On the algebras over equivariant little disks”. In: J. Pure Appl. Algebra 226.10
(2022), Paper No. 107052, 21. issn: 0022-4049,1873-1376. doi: 10.1016/j.jpaa.2022.107052.
arXiv: 1709.02005 [math.AT] (cit. on p. 6).
[HH16] Michael A. Hill and Michael J. Hopkins. Equivariant symmetric monoidal structures. 2016.
arXiv: 1610.03114 [math.AT] (cit. on pp. 4, 6).
[Hil24] Kaif Hilman. Parametrised Presentability over Orbital Categories. 2024. arXiv: 2202.02594
[math.AT] (cit. on p. 22).
[HM23] Vladimir Hinich and Ieke Moerdijk. On the equivalence of the Lurie’s ∞-operads and dendroidal
∞-operads. 2023. arXiv: 2206.14033 [math.CT] (cit. on p. 3).
[Hor19] Asaf Horev. Genuine equivariant factorization homology. 2019. arXiv: 1910.07226 [math.AT]
(cit. on pp. 6, 37, 39, 40, 44).
[KK24] Manuel Krannich and Alexander Kupers. ∞-operadic foundations for embedding calculus. 2024.
arXiv: 2409.10991 [math.AT]. url: https://arxiv.org/abs/2409.10991 (cit. on p. 53).
[LLP25] Tobias Lenz, Sil Linskens, and Phil Pützstück. Norms in equivariant homotopy theory. 2025.
arXiv: 2503.02839 [math.AT]. url: https://arxiv.org/abs/2503.02839 (cit. on pp. 6, 10,
22).
[Lin76] Harald Lindner. “A remark on Mackey-functors”. In: Manuscripta Math. 18.3 (1976), pp. 273–
278. issn: 0025-2611,1432-1785. doi: 10.1007/BF01245921. url: https://people.math.
rochester.edu/faculty/doug/otherpapers/lindner.pdf (cit. on p. 4).
[HTT] Jacob Lurie. Higher topos theory. Vol. 170. Annals of Mathematics Studies. Princeton University
Press, Princeton, NJ, 2009, pp. xviii+925. doi: 10 . 1515 / 9781400830558. url: https :
//www.math.ias.edu/~lurie/papers/HTT.pdf (cit. on pp. 11, 21, 46).
58 REFERENCES

[HA] Jacob Lurie. Higher Algebra. 2017. url: https://www.math.ias.edu/~lurie/papers/HA.pdf

(cit. on pp. 2–4, 6, 9, 11, 19, 22, 27, 28, 33, 35, 36, 40, 44, 45, 49, 52–55).
[Mar24] Gregoire Marc. A higher Mackey functor description of algebras over an N∞ -operad. 2024.
arXiv: 2402.12447 [math.AT] (cit. on pp. 9, 10, 23).
[MQS24] David Mehrle, J. D. Quigley, and Michael Stahlhauer. Koszul Resolutions over Free Incomplete
Tambara Functors for Cyclic p-Groups. 2024. arXiv: 2407.18382 [math.AT]. url: https:
//arxiv.org/abs/2407.18382 (cit. on p. 43).
[Nar16] Denis Nardin. Parametrized higher category theory and higher algebra: Exposé IV – Stability
with respect to an orbital ∞-category. 2016. arXiv: 1608.07704 [math.AT] (cit. on pp. 5, 8,
11, 12, 23, 24).
[Nar17] Denis Nardin. Stability and distributivity over orbital ∞-categories (thesis). 2017. url: https:
//dspace.mit.edu/handle/1721.1/112895?show=full (cit. on pp. 22, 51).
[NS22] Denis Nardin and Jay Shah. Parametrized and equivariant higher algebra. 2022. arXiv: 2203.
00072 [math.AT] (cit. on pp. 2, 4, 9–11, 13–15, 17, 22–24, 38, 40, 46).
[NS18] Thomas Nikolaus and Peter Scholze. On topological cyclic homology. 2018. arXiv: 1707.01799
[math.AT]. url: https://arxiv.org/abs/1707.01799 (cit. on pp. 4, 42).
[PS18] Dmitri Pavlov and Jakob Scholbach. “Admissibility and rectification of colored symmetric
operads”. In: J. Topol. 11.3 (2018), pp. 559–601. issn: 1753-8416,1753-8424. doi: 10.1112/
topo.12008. arXiv: 1410.5675 [math.AT] (cit. on p. 9).
[Per18] Luís Alexandre Pereira. “Equivariant dendroidal sets”. In: Algebr. Geom. Topol. 18.4 (2018),
pp. 2179–2244. issn: 1472-2747,1472-2739. doi: 10.2140/agt.2018.18.2179. url: https:
//arxiv.org/abs/1702.08119 (cit. on p. 10).
[QS19] J. D. Quigley and Jay Shah. On the parametrized Tate construction and two theories of real
p-cyclotomic spectra. 2019. arXiv: 1909.03920 [math.AT]. url: https://arxiv.org/abs/
1909.03920 (cit. on pp. 40–43).
[QS22a] J. D. Quigley and Jay Shah. On the equivalence of two theories of real cyclotomic spectra.
2022. arXiv: 2112.07462 [math.AT]. url: https://arxiv.org/abs/2112.07462 (cit. on
pp. 41–43).
[QS22b] J. D. Quigley and Jay Shah. On the parametrized Tate construction. 2022. arXiv: 2110.07707
[math.AT]. url: https://arxiv.org/abs/2110.07707 (cit. on p. 13).
[Rub19] Jonathan Rubin. Characterizations of equivariant Steiner and linear isometries operads. 2019.
arXiv: 1903.08723 [math.AT] (cit. on p. 38).
[Rub21] Jonathan Rubin. “Operadic lifts of the algebra of indexing systems”. In: J. Pure Appl. Algebra
225.12 (2021), Paper No. 106756, 35. issn: 0022-4049,1873-1376. doi: 10.1016/j.jpaa.2021.
106756. url: https://par.nsf.gov/servlets/purl/10393698 (cit. on pp. 7, 10).
[SY19] Tomer M. Schlank and Lior Yanovski. “The ∞-categorical Eckmann-Hilton argument”. In:
Algebr. Geom. Topol. 19.6 (2019), pp. 3119–3170. issn: 1472-2747,1472-2739. doi: 10.2140/
agt.2019.19.3119. url: https://arxiv.org/abs/1808.06006 (cit. on pp. 3, 30, 31, 33).
[Seg74] Graeme Segal. “Categories and cohomology theories”. In: Topology 13 (1974), pp. 293–312. issn:
0040-9383. doi: 10.1016/0040-9383(74)90022-6. url: https://doi.org/10.1016/0040-
9383(74)90022-6 (cit. on p. 3).
[Sha22] Jay Shah. Parametrized higher category theory II: Universal constructions. 2022. arXiv: 2109.
11954 [math.CT] (cit. on pp. 11, 55).
[Sha23] Jay Shah. “Parametrized higher category theory”. In: Algebr. Geom. Topol. 23.2 (2023),
pp. 509–644. issn: 1472-2747,1472-2739. doi: 10 . 2140 / agt . 2023 . 23 . 509. url: https :
//arxiv.org/pdf/1809.05892.pdf (cit. on pp. 11, 47).
[Ste24] Natalie Stewart. Orbital categories and weak indexing systems. 2024. arXiv: 2409 . 01377
[math.CT]. url: https://arxiv.org/abs/2409.01377 (cit. on pp. 2, 5, 10–12, 21, 31, 38).
[Ste25] Natalie Stewart. Equivariant operads, symmetric sequences, and Boardman-Vogt tensor products.
2025. arXiv: 2501.02129 [math.CT]. url: https://arxiv.org/abs/2501.02129 (cit. on
pp. 2, 5–12, 14–18, 23, 34, 37, 44, 50–53).
[Szc24] Ben Szczesny. Equivariant Framed Little Disk Operads are Additive. 2024. arXiv: 2410.20235
[math.AT]. url: https://arxiv.org/abs/2410.20235 (cit. on p. 38).
 REFERENCES 59

[Yan23] Lucy Yang. On normed E∞ -rings in genuine equivariant Cp -spectra. 2023. arXiv: 2308.16107
[math.AT]. url: https://arxiv.org/abs/2308.16107 (cit. on p. 36).