The Two Critical Driverless Technologies You Must Own

Today
Wednesday, February 5, 2020
A  A

 Print
One of our favorite topics in investing – and analysis in general – is understanding the
influence of the human mind and the biological impact it has on our thinking.

Through hundreds of thousands of years, the human brain has developed many tendencies
that dominate the way we view the world. Even with all of the progress of modern society,
these tendencies still have an overwhelming influence on our understanding of events.

One of the most powerful of these impulses is called “extrapolation bias.” This is a simple
tendency in which the human mind takes a small number of recent events and then projects
that they will continue indefinitely into the future.

This is an overwhelming impulse for the human mind… and it’s one of the many reasons
why most people are so bad at accurately predicting the future.

In particular, humans tend to be bad at predicting big changes. Most of these tend to happen

suddenly, but by extrapolating individual data points, our minds usually build much more
gradual predictions.

The disconnect between the reality of big, rapid changes versus the expectation of
gradual change can present tremendous investment opportunities.

We were reminded of this as we watched futurist Tony Seba’s fascinating presentation at the
Robin Hood Investors Conference in New York in late 2019. (If you haven’t seen it, you can
watch right here.)

One of Seba’s main points was that big technological changes don’t usually occur gradually,
but quickly… much quicker than the so-called “experts” typically expect.

For example, in 1985, leading U.S. telecommunications company AT&T (T) paid esteemed
consulting firm McKinsey millions of dollars to predict how new mobile-phone technology
would develop over the subsequent 15 years.
After much research, analysis, and deliberation (gotta justify those high fees!), McKinsey
predicted that there would be 900,000 mobile phone customers by the year 2000. In reality,
there were more than 100 million. McKinsey’s estimate was off by a factor of more than
100!

One of the main reasons these errors occur more frequently in the tech sector is because of
the exponential curves that characterize technology development. We can see this with the
concept of Moore’s Law.

The idea, which originated from Intel (INTC) founder Gordon Moore, is that the number of
transistors on a microchip will double every two years, thereby cutting the cost of computing
power in half.

This is an increase of 32 times in five years, 1,024 times in a decade, and more than 1 million
times in two decades. This clearly does not fit the gradual and linear model of how the human
mind works.

As you can see in the following images, we’ve seen similar rapid growth among various other
technologies in just the last few decades. 

As recently as 20 years ago, few Americans owned a cellphone or a personal computer.

Today, they’re virtually ubiquitous in U.S. households…
The same can also be said about the Internet. It’s hard to believe that as the dot-com bubble
was bursting in 2001, only half of U.S. households had Internet access. Today, nearly
everyone has high-speed Internet access at home…
And social media, which literally didn’t exist until the mid-2000s, is now a part of everyday
life. Today, 2.5 billion people – nearly one-third of the world’s population – use leading
social media platform Facebook (FB) on a monthly basis, and this doesn’t count Instagram,
Twitter, WeChat, and many others…
In the years ahead, we’re going to see a massive expansion along the lines of something
that’s straight out of science fiction… Like the Internet itself, it had its beginnings as part of a
project funded by the United States’ Department of Defense Advanced Research Projects
Agency (“DARPA”).

The Birth of Autonomous Driving

Back in the early 1980s, folks at DARPA approached researchers at Carnegie Mellon’s
Robotics Institute about a crazy project – a self-driving car. We now refer to these as
“autonomous vehicles” (or “AVs”). Today, auto manufacturers, leading technology
companies, and entrepreneurs have invested upwards of $100 billion into this idea.

When analysts discuss AVs, they talk about “levels.” The Society of Automotive Engineers
(“SAE”) has defined them as follows…

Level 0: Like most cars today, which require full human control.

Level 1: Level 1 cars feature various technologies that are commonplace in newer cars today
such as parking assistance, collision warning (typically combined with emergency braking),
and “adaptive cruise control,” where your car will speed up or slow down to maintain a safe
distance from the car in front of you.
Level 2: Level 2 vehicles have partially automated driving – also known as “hands off” – like
what’s found in Tesla’s Autopilot, Cadillac’s Super Cruise, and Volvo’s Pilot Assist. Level 2
cars can usually steer, accelerate, and brake on their own but can really only drive themselves
under the easiest conditions like clear weather and on highways or clearly marked roads.
Drivers must always pay attention and be ready to take the wheel in case the self-driving
system fails.

Level 3: Level 3 vehicles can steer, accelerate or decelerate, and pass other cars without
human input. They can also maneuver around incidents or traffic jams. Where level 2 cars
require drivers to have at least a finger on the wheel, level 3 systems enable drivers to take
their hands off the wheel and feet off the pedals — but only in specific situations. Humans
still need to be ready to take back control when the car requests it.

The first of these vehicles, the Audi A8, was originally scheduled for availability in late 2019,
but Audi decided that it needed better answers regarding liability, responsibility, and
regulation before rolling out the technology.

Level 4: According to SAE guidelines, a level 4 car should be able to drive itself safely,
“even if a human driver does not respond appropriately to a request to intervene.” A level 4
car will slow down, pull over, or park itself at a safe spot if the driver doesn’t take control
when requested, which might happen in tougher driving conditions like off-road driving or
unmapped roads. This is also known as “mind off,” because the car can drive itself so well
that you can read a book or even go to sleep.

Level 5: Level 5 is the holy grail… Fully autonomous vehicles that never require human
involvement, even under the most difficult conditions. In fact, they wouldn’t have a steering
wheel or other controls. Instead, think of a futuristic “pod.”
Search for “autonomous vehicles” on the Internet, and you’ll see predictions about when the
more advanced levels will be available. The overwhelming consensus is that it will take
decades.

This forecast is consistent with how the human mind works, as we discussed earlier… But as
we know from history, the reality is almost always different.

What we find most interesting as we study AVs is that this is no longer a technology issue.
The technology exists across the board – cameras, computers, and drive systems – to achieve
full autonomy.
Rather, the reason we don’t have fully autonomous cars today is due to data, equipment
costs, and humans.

It’s astounding how much data we take in as we drive our cars. We make decisions large
(following the hand signals of a police officer), medium (avoiding obstacles), and small
(stopping at a stop sign) without even really thinking too much, using one of the greatest
computers ever made – our brain – and years of experience.

The first AVs lacked both the computing power and the experience. But they’re catching up
in a big way… The first generations of artificial intelligence (“AI”) took years to master
skills, but the latest generation is doing it in hours.

Now combine this with the fact that many of the major technology companies exploring
autonomous driving – Tesla (TESLA), Uber (UBER) and Alphabet’s (GOOGL) Google (via
Waymo) – have all of their cars plugged into their data networks.

More than 650,000 Teslas are on the road today, and all of them are reporting back every
single piece of data they gather. As of mid-2019, that covered more than 1.6 billion miles of
data. By 2022, it will be more than 10 billion miles. And Google’s Waymo already has more
than 10 million miles of data from actual AVs driving on regular roads.

Given the continued developments in computing power and the massive amount of data being
gathered, this technology will be driving cars better than human drivers can… and
much sooner than anyone imagines.

Technological advancements are also solving the issues of equipment cost and size. The
components needed for the automation systems in AVs are all related to electronics, optics,
and data. With the exponential growth in capacity development in every one of these areas,
we believe that the costs and size will come down much sooner than current expectations.

So that leaves the real obstacle: humans. The problem here? Fear of the unknown… and
plenty of uncertainty from the public about self-driving cars. In light of this, will regulators
permit AVs? And, if so, with what restrictions?

The other common dismissal of the near-term prospects for autonomous driving is that the
regulatory environment isn’t ready for it and that safety concerns will prohibit it.

Again, though, history can guide us… and historical reality is that regulations around new
technologies tend to develop fairly quickly.
Right now, the software and vehicle manufacturers are expected to be liable for any at-fault
collisions based on existing automobile product-liability laws. Given the potential positive
ramifications for society with AVs, though, it’s likely that we’ll see an adjustment to existing
laws to reflect the new environment.

Today, large companies that operate fleets (truck drivers, buses, taxis, etc.) are able to
purchase liability insurance that protects them against their drivers’ liability. If the liability in
an accident would indeed still reside with the manufacturer of the vehicle or software, then
why couldn’t we view the vehicles they’ve sold as a “fleet” and insure it in the same way?
Given the guarantee that the number of accidents with AVs will be substantially less than
with human-driven fleets, this would be an economical solution.

Once the data equation has been properly calculated and technology is ready for true
commercial rollout, regulation will likely follow soon after.

So… if the current perception about the timing of rolling out AVs is wrong, we want to find
the chief beneficiaries that we can invest in today ahead of these big changes. And in this
special report, we’ve found the companies that we think are the best way to profit from these
exciting advancements…

AV Stock No. 1: Aptiv (NYSE: APTV)

Our first beneficiary dates back to a decision by carmaker General Motors (GM) in 1994…

That year, GM decided to organize its auto parts into a division called the Automotive
Components Group. That company would eventually change its name to Delphi Automotive
Systems. It went bankrupt in 2005 due to accounting irregularities and then re-emerged
during the depths of the financial crisis.

A group of private investors eventually bought and reorganized the company and took it
public in late 2011. In 2017, it changed its name again, becoming Aptiv (NYSE: APTV).

Aptiv is the dominant company in advanced electrical and safety systems for automobiles. Its
largest division is the Signal and Power Solutions business, which accounts for 70% of
revenue and 85% of profits.

Thirty years ago, the power requirements and complexity of these systems were relatively
simple. But today, the average car operates with massive computing power and significantly
more sophisticated electrical and power requirements.
Historically, vehicle manufacturers operated with an incremental approach – adding
additional systems to pre-existing architecture. Given the increase in complexity, however,
this is becoming untenable… and companies need to completely redesign these systems.

As you might imagine, the move to autonomous and electric vehicles (“EVs”) will require
exponentially more electrical and power complexity.

Aptiv has by far the most technologically advanced solutions for what is essentially the
“nervous system” for AVs.

The benefits of having a cutting-edge technological nervous system to go along with the
“brain” (computing power) include more safety for passengers, lower emissions, and new
opportunities in connectivity.

Aptiv’s second-largest division is Advanced Safety and User Experience, which makes up the
other 30% of revenues and 15% of profits. This segment develops safety technology and
other technologies that are vital to AVs.

Remember, only the Audi A8 is available right now with Level 3 autonomy. Much of its
platform uses Aptiv’s sensors and software, which highlights the company’s dominant
position in providing these systems.

Aptiv estimates that its incremental content per vehicle for Level 4 or Level 5 autonomy
could be worth as much as $5,000. So, as AVs grow exponentially, Aptiv (and its
shareholders) stand to benefit tremendously…

Aptiv is also uniquely positioned to tap other growth opportunities in the sector, including
mobility solutions and connected services.

In recent years, Aptiv has done a good job of growing its revenues and earnings. But
weakness in the global automobile industry is presenting us with a unique opportunity right
now, as Aptiv’s stock trades 2015 levels…

Current analyst estimates for Aptiv are based on the consensus view that there will only be a
gradual rollout of both AVs and EVs. If it happens faster, as we expect, the company will be
a major beneficiary, and quickly.

For instance, if Aptiv were to provide its technologies to 2.8 million vehicles (3.5% of the
market), at $5,000 each, the company’s revenues would double.
But even if Aptiv-powered AVs were to capture only 1% of the market, it would still result in
25% incremental revenue growth and up to $1.50 in earnings per share (“EPS”), a 20%
increase to current 2022 EPS estimates.

If the rise of AVs and EVs plays out as we expect, Aptiv could make upwards of $8 per share
in earnings in 2022. Using a 20-times multiple gives us a price target of $160…
nearly double today’s levels.

ACTION TO TAKE
Buy Aptiv (NYSE: APTV) up to $105 per share.
AV Stock No. 2: Nvidia (Nasdaq: NVDA)
Today’s video-game environments in popular multiplayer “ecosystems” – like Minecraft and
Fortnite – involve millions of simultaneous players operating in virtual worlds that involve
trillions of pieces of data.

The amount of computing power that’s necessary to create these sophisticated simulations is
immense… and one company dominates that market: Nvidia (Nasdaq: NVDA).

Nvidia is also well positioned to dominate the advanced processors that will be necessary for
AVs to “play” the real-life video game that is driving.

Nvidia was founded in 1993 with an initial seed investment of just $40,000. Today, it sports a
$145 billion market cap.

Nvidia designs graphics processing units (“GPUs”) for gaming and professional markets
along with system on a chip (“SoC”) units for mobile computing and automobiles.

When you think about the complexity of a video-game environment and the amount (and
speed) of processing necessary to create it and manage all of the actions that take place within
it, you can see how it relates to the real-time data required for autonomous driving. Look at
how much this screenshot from an AV car is like a video game…
Last year, Nvidia launched its Nvidia Drive platform and Optimus Ride partner to compete
with Waymo.

Like Aptiv, the opportunity for Nvidia is massive here…

The company has a dominant 80% share of the high-end gaming GPUs market… and its
technology and willingness to invest could put it in a similar position with AVs.

Nvidia’s chipset might cost around $1,000 per vehicle. At that price, every 1 million AVs
produced would be an incremental $800 million of revenue for Nvidia (assuming 80% market
share). That would equal a 7% increase in the company’s revenue, but Nvidia’s incremental
margins exceed 60%, so this could mean an incremental 10% boost to earnings.

Keep in mind that is a scenario where AVs only make up less than 1.5% of overall
automobile production. Over time, there could be as many as 40 million AVs produced every
year, which, using the same math, could result in $32 billion in additional revenue…
nearly triple Nvidia’s current sales.

In the meantime, Nvidia is an incredible business with remarkable financial and operational
performance: it has tripled revenues over the past decade while increasing operating income
more than 50-fold.
Analysts expect this growth to continue, driven by its core GPU business, dominant market
share, and the strong secular growth in its end markets – like PC gaming, data centers, and
autonomous machines.

Right now, almost nothing in any of these numbers reflect the opportunity for AVs…
Remember, the current consensus is that this rollout will happen gradually.

NVDA shares are up more than 1,100% over the past five years… But if we begin to see
sales to the AV sector materially impact results in 2022 or 2023, the stock will likely
skyrocket from here.

Nvidia’s shares can be volatile. In late 2018, they got cut in half due to lowered guidance and
a broad market pullback but have mostly recovered since.

The company has also seen its price-to-earnings (“P/E”) multiple expand with the recent
gains, but they’re still well below their previous highs.

Using current expectations for earnings growth and assuming a constant P/E ratio, we arrive
at a target price of $315 by 2023… more than 25% above Nvidia’s current share price.

However, if we’re right about the growth of AVs, this would both boost growth (adding 10%
to EPS) and likely the P/E ratio as well to previous highs of 45 times (if not higher), which
could lead to a doubling of the stock from here.

In light of its valuation and volatility, Nvidia is the most speculative of our three picks… But
given the size of the opportunity in AV and the company’s market-leading position, it’s an
attractive speculation over the next few years.

ACTION TO TAKE
Buy Nvidia (Nasdaq: NVDA) up to $350 per share.
Bonus AV Stock No. 3: Alphabet (Nasdaq: GOOGL)
As we mentioned above, DARPA sponsored the first real work on AVs. In 2004, after two
decades of work, the agency established the “Grand Challenge,” a long-distance competition
for self-driving cars with a $1 million prize to the winner.

The initial competition ended with no winner, as none of the vehicles were able to complete
the course… So, for the 2005 race, DARPA increased the prize to $2 million. Five cars
finished. The winner was a vehicle named Stanley, created by Carnegie Mellon alumni
Sebastian Thrun, who ran the Stanford Artificial Intelligence Laboratory at the time.

Thrun is also the founder of the primary effort in AVs for tech titan and Google parent
company Alphabet (Nasdaq: GOOGL).

Alphabet’s products are ubiquitous. Chances are good you use several of them every day.
They include the Google search engine, Android operating system, Google Maps, Google
Chrome browser, YouTube streaming service, Google Play app store, Gmail e-mail service,
and many others. Each of these products has more than 1 billion monthly users, and most
have dominant market shares.

Alphabet has an incredible business model… Because it provides information electronically,

its products can scale to nearly every human being on Earth with no additional cost. This
generates huge free cash flow.

Even with its massive size (more than $150 billion in revenue), the company is still
continuing to grow its sales at nearly 20% annually. Alphabet has been able to build these
dominant positions and maintain this growth because of its willingness to invest in new areas
and technologies.

Its latest group of investments – which includes Google’s self-driving car project, “Waymo”
– falls under their category of “Other Bets.” They currently generate less than $600 million in
annual revenue and lose $3.4 billion a year. But Alphabet is so insanely profitable that it
almost doesn’t matter. And it has a strong track record of creating valuable businesses from
these early stage investments, so investors aren’t complaining.

The name Waymo comes from its mission: “a new way forward in mobility.” It began in
2009 as part of cofounder Sergey Brin’s secretive “X” project.

Brin hired Thrun to helm Waymo, which revealed its first prototype car in 2014… But the
real progress began in 2018 when it rolled out a ride-hailing service in Phoenix. Waymo
chose Arizona because of the good weather, simple and straight roads, and favorable
regulation.

This service is similar to ride-sharing company Uber’s (UBER) platform: You hail a car via
the Waymo app, and a Chrysler Pacifica minivan picks you up.
Until recently, the car arrived with a safety driver in the vehicle who could intervene if
necessary. But now, many of the cars are completely self-driving. (You can watch a video of
the whole experience right here.)

Waymo hasn’t detailed its plans for future rollouts, but we expect it to dramatically expand
the service in the next few years.
But the service itself isn’t even the most interesting aspect. What fascinates us the most is the
data and experience that Alphabet gains from operating it…

Similar to what it did with cellphones and the Android operating system, Alphabet wants to
dominate the operating system for AVs in the same way it does with mobile phones, web
browsing, and Internet searches.

If it is even partially successful, Alphabet could reap billions in incremental annual profits
because this industry is so large. The global automobile industry generates $4 trillion in
annual revenue, and Americans, on average, spend more than two full weeks in their cars
every year.

By dominating the operating system, Alphabet – following its own playbook from Android
and Google Maps – may have opportunities in advertising, e-commerce, and other licensing
and transactions.

In summary, Alphabet is investing more than any other company in AVs… has nearly
unlimited resources… and is attacking a huge market. As a result, we’re confident that it will
be a major beneficiary of the coming revolution in AVs.

ACTION TO TAKE
Buy Alphabet (Nasdaq: GOOGL) up to $1,600 per share.