The autonomous-driving story is finally moving from “future technology” to “commercialization.”
But that doesn’t mean every AV stock deserves to be bought on a red day.
The opportunity is in understanding which layer of the stack is actually monetizing today — and which companies are still selling the dream.
My AV watchlist breaks into four layers:
Compute → Integrators → Perception → OEMs
And the risk/reward is very different across each.
1. The Brain — Compute & Silicon
$NVDA — My highest-quality AV infrastructure play
NVIDIA remains one of the strongest ways to own the broader autonomous-driving compute cycle.
The advantage isn’t simply GPUs.
It’s the combination of:
- CUDA
- AI training infrastructure
- inference
- DRIVE platform
- software ecosystem
- enormous developer base
The beauty of NVDA is that you don’t need autonomous vehicles to become a massive standalone business for the stock to work.
AV is additional optionality on top of an already enormous AI infrastructure business.
$MBLY — One of the cleanest AV pure plays
Mobileye is much more directly tied to the autonomous-driving thesis.
It has exposure to:
ADAS → SuperVision → Chauffeur → autonomous driving
The risk is execution, competition and the amount of capital required to move from advanced driver assistance toward true autonomy.
But unlike many speculative AV names, Mobileye already has meaningful OEM relationships and an installed technology base.
For investors wanting pure AV exposure without betting on a tiny lidar company, MBLY remains one of my favorites.
$QCOM / $AMBA
These are more specialized plays.
QCOM: Snapdragon Ride gives Qualcomm exposure to automotive compute and ADAS.
AMBA: much higher beta, but interesting because of its edge-AI and computer-vision exposure.
AMBA is the type of name I would rather trade around support than make a huge core position.
2. Integrators — Where I See the Biggest 2026 Inflection
This is the layer I find most interesting.
Why?
Because we’re finally getting something investors can measure:
Miles. Vehicles. Customers. Revenue.
That’s much more valuable than another autonomous-driving demonstration.
$AUR — One of my favorite AV risk/reward setups
Aurora is probably the most interesting commercialization story in autonomous trucking.
The company had nearly 440,000 driverless miles by the end of June 2026 and has launched its second-generation driverless truck platform.
More importantly, Aurora says it is fully allocated to exit 2026 with approximately 200 driverless trucks operating commercially.
That’s the key transition:
Pilot → commercial network → scale
Aurora is also targeting a major reduction in hardware costs with its second-generation system.
And that’s where the economics get interesting.
A driverless Class 8 truck doesn’t need to sleep.
It doesn’t take weekends off.
It doesn’t have hours-of-service limitations.
The potential utilization improvement is enormous.
I’d rather own AUR after fear than after euphoria.
$PONY — One of the most interesting robotaxi growth names
Pony is showing something I want to see from an AV company:
revenue growth + fleet growth + commercial deployment.
Q2 2026 revenue increased 68.8% YoY to $36.2M, while robotaxi revenue jumped 691.2% YoY to $12.1M. Its robotaxi fleet reached 1,975 vehicles by June 30.
Even more interesting:
Pony is targeting more than 3,500 robotaxis by year-end.
That’s a huge jump from simply proving the technology works.
It’s attempting to build an actual business around it.
The risk?
China.
Regulation.
Cash burn.
Competition.
And valuation.
So I wouldn’t treat PONY like NVDA.
I’d treat it as a high-beta growth position.
$GOOG — The Quiet AV Giant
This is one of the most interesting setups in the entire sector.
Waymo isn’t a PowerPoint presentation anymore.
It’s operating a commercial driverless service.
In September, Waymo expanded public rides into Denver, San Diego and Tampa, bringing the number of cities where it provides fully autonomous trips to 14.
And that’s buried inside Alphabet.
That’s the part investors sometimes overlook.
You’re not buying a company that depends entirely on robotaxis succeeding.
You’re buying:
Google Search + YouTube + Cloud + AI + Gemini + Waymo
Waymo is essentially a massive embedded option.
GOOG is one of my favorite lower-risk ways to own AV
If Waymo eventually becomes a massive transportation platform, the upside is meaningful.
If it takes longer than expected, Alphabet still has its core businesses.
$TSLA — Highest upside, highest debate
Tesla is completely different.
It’s trying to own multiple layers:
AI compute → software → vehicle → fleet → robotaxi network
Tesla currently lists autonomous Robotaxi service in Austin, Dallas, Houston, Miami, Orlando and Tampa, while Cybercab rides are available in limited areas of Austin.
That’s real progress.
But Tesla also carries enormous expectations.
The bull case:
Millions of vehicles → autonomous software → robotaxi network → enormous operating leverage
The bear case:
Regulatory limitations, safety questions, slower deployment and valuation.
So TSLA isn’t a simple AV trade.
It’s an entire ecosystem bet.
3. Perception — The Most Dangerous Layer
This is where I become much more selective.
Lidar has produced incredible technology.
It has also produced an incredible graveyard of stocks.
Names like:
$HSAI
$AEVA
$ARBE
$OUST
$MVIS
$INVZ
can produce huge moves.
But that’s exactly the problem.
Technology leadership does not automatically equal shareholder returns.
You can have:
Great technology
Major OEM interest
Increasing deployments
…and still have:
Dilution
Cash burn
Margin pressure
Pricing wars
$HSAI
One of the more interesting lidar names because of scale and Chinese automotive exposure.
But China + lidar + valuation makes this a higher-risk combination.
$AEVA
The FMCW lidar story remains technically interesting.
If Aeva converts technology into large-scale production and OEM wins, the upside could be significant.
But this is still speculative.
$OUST
Interesting technology and potentially a survivor in a consolidated lidar industry.
But I would still classify it as speculative rather than core.
4. OEMs — The Indirect AV Plays
Then we have:
$TSLA
$TM
$NIO
$LI
$XPEV
$GM
$RIVN
$LCID
$MBGYY
The important distinction:
For most of these companies, AV isn’t the reason you buy the stock.
It’s an additional catalyst.
My AV Ranking
If I were building an AV basket today, I wouldn’t weight every layer equally.
Core / Higher Conviction
$NVDA — AV compute + AI infrastructure
$GOOG — Waymo + massive core business
$MBLY — pure AV/ADAS exposure
$AUR — autonomous trucking commercialization
$TSLA — vertically integrated autonomy ecosystem
Higher Growth / Higher Risk
$PONY — robotaxi commercialization
$QCOM — automotive compute
$AMBA — edge AI / computer vision
$HSAI — lidar scale
$AEVA — differentiated lidar
$OUST — lidar consolidation survivor
Speculative / Catalyst Trades
$MVIS
$INVZ
$ARBE
These can move violently, but I wouldn’t confuse volatility with a durable investment thesis.
The Biggest Shift I’m Watching
The AV thesis is changing from:
“Who will eventually achieve full autonomy?”
to:
“Who can commercially deploy autonomy at scale while improving unit economics?”
That’s a massive difference.
Aurora is showing the trucking side of the equation with real driverless miles and commercial fleet expansion.
Waymo is showing the robotaxi side with expansion into additional cities.
Pony is showing that robotaxi revenue can start scaling rapidly alongside fleet deployment.
Tesla is pursuing the vertically integrated version of the same thesis.
That’s why I think 2026–27 could be an important period for AV investors.
Not because every AV company will win.
Quite the opposite.
The next phase could be about separating the companies that can actually commercialize autonomy from the companies that simply have impressive technology.
My strategy remains the same:
Don’t chase the headline.
Wait for the pullback.
Build in tranches.
Own the strongest layer of the stack.
Let the winners prove themselves.
AV is no longer one trade. It’s a stack — and I want to own the layers where commercialization is actually accelerating.
NFA.