The Bet Inside the Bet:What Tesla's Valuation Requires You to Believe

This memo focuses on Tesla's robotaxi business. Not because Optimus or energy storage are unimportant to the valuation. They are. The Stacked Bets section addresses both. But the robotaxi thesis is where the valuation is most testable right now.

Waymo provides a live operational comparison that does not exist for humanoid robotics. There is no Waymo of Optimus. The robotaxi market is nascent but real, which means the gap between what Tesla's valuation requires and what Tesla's operational record shows can actually be measured.

That is where the analysis is sharpest, and that is where this memo focuses.

ARK Invest, led by Cathie Wood, carries a $2,600 price target on Tesla implying a roughly $9 trillion company by 2029.

ARK's published model attributes approximately 88% of projected enterprise value and earnings to the robotaxi business alone. Electric vehicle sales represent just 9% of enterprise value.

Wedbush analyst Dan Ives, one of Wall Street's most vocal Tesla bulls, carries a $600 price target with a 12-to-18-month horizon and a bull-case scenario of $3 trillion in market cap by end of 2026. He called the Austin robotaxi launch 'the golden era of autonomous' and estimates the AI and autonomous opportunity alone is worth at least $1 trillion for Tesla.

Two of the most prominent Tesla bulls. One thinks the company is worth $9 trillion. The other thinks it could reach $3 trillion this year.

$9 trillion by 2029 or $3 trillion this year. Both are just disagreements about which speculative bet resolves, and when. That specific dynamic frames everything else that follows.

These valuations require you to believe simultaneously:

There also is a Musk fandom variable that does something specific as well. Musk believers tend to dismiss Waymo's operational lead as irrelevant because they believe Musk will leapfrog it.

That dismissal is not a technology assessment. It is a faith-based, history-driven assumption dressed as competitive analysis. 

The Musk skeptics make the mirror error. They dismiss the technology because of the person. 

Neither position is actual analysis, however.

Before examining the robotaxi thesis specifically, it is worth understanding exactly how much of Tesla's actual current valuation rests on businesses that do not yet exist at commercial scale.

Andrew Percoco, the Morgan Stanley analyst covering Tesla, breaks the company into five pillars in his sum-of-the-parts model: core automotive at $55 per share; energy storage at $40 per share; Network Services including FSD software at $145 per share; Mobility including Robotaxi at $125 per share; Optimus at $39 per share; and being a 3rd party supplier at $11 per share. Total: the $415 price target.

The two businesses generating meaningful current revenue total $95 per share. This is against a stock trading above $400, more than three quarters of Tesla's current market cap rests on businesses that have not yet demonstrated commercial scale.

ARK carries a $2,600 target by 2029. Wedbush is at $600 this year. And Morgan Stanley sits at $415. 

And then there’s Ron Baron, whose Baron Partners Fund holds Tesla as its second largest position representing 19% of the fund (only second to SpaceX). Baron recently told CNBC that Tesla could reach $2,000 to $2,500 per share over the next ten years. 

Baron's conviction deserves a specific note. It is not irrational. Baron Capital first bought Tesla in 2014 at an average price of about $45 per share. The position has generated an estimated $7 billion in gains to date - around a 1,500% return. Baron has famously noted that he does not expect to sell a single personal share of Tesla in his lifetime.

Baron entered SpaceX in 2017 at a reported $20 billion valuation. Both bets looked improbable. Both paid off substantially. The track record of betting on Musk's ability to execute on improbable things is real.

Is conviction earned on bets that have already paid off the right analytical tool for evaluating bets that require a specific market structure outcome that Musk has not yet won?

Building electric cars at scale against a skeptical industry is a different problem than capturing the majority of robotaxi economics against an already-operational, well-capitalized competitor actively scaling its own fleet. The conviction is the same. The problem is not.

How much of Tesla's valuation reflects a genuine probability-weighted belief that Tesla will solve autonomy and capture the robotaxi market? And how much reflects something else - a founder premium, a fandom premium?

Those are different questions. Yet, many have collapsed them into one price.

Musk fandom is not a proxy for whether vision-only autonomy works. Musk skepticism is not a proxy for whether it fails. The technology and competitive questions, along with the sentiment question need to be read separately before the investment thesis can be assessed clearly.

Tesla's stock contains multiple signals simultaneously. That makes it very difficult to read what the market actually believes about robotaxi execution versus what it believes about Elon Musk.

This matters for investors on three sides, not just two. If you are long Tesla primarily because you believe in Musk, you are taking a position on a person rather than a technology and competitive thesis. If you are short Tesla primarily because you distrust Musk, you may be underweighting a technology program that has real capability and a cost structure that is genuinely difficult to replicate.

The third position is the one financial coverage rarely acknowledges. It’s the investor who has decided that when sentiment and technology are this entangled in a single price, the stock is simply too difficult to underwrite with conviction in either direction. That investor is not indifferent to Tesla. They are making a considered judgment that the Musk variable introduces an uncertainty that no amount of robotaxi operational analysis fully resolves. Sitting out is its own analytical conclusion.

All three positions can be held by serious, informed investors. What none of them should be is a substitute for reading the operational record carefully.

What makes the Musk variable analytically difficult to isolate is that it does not sit alongside the financial models. It operates inside them.

Belief in Musk's execution ability shapes the input assumptions. How large the robotaxi market will be. How much of it Tesla will capture. On what timeline. At what margin. Those assumptions run through a cash flow model and produce a price target that looks like financial analysis.

The valuation spread between ARK at $2,600 and Percoco at $415 is not primarily a disagreement about arithmetic. It is a disagreement about the Musk assumption embedded in the inputs.

It means the debate about Tesla's valuation cannot be resolved by better modeling alone. It requires an honest accounting of what each analyst actually believes about one person's ability to win markets that aren't yet fully formed.

I think the Musk variable is a more significant and more durable component of Tesla's price than most financial coverage acknowledges. Ignoring it, in either direction, is its own kind of analytical error.

Waymo has completed more than 20 million fully autonomous trips. It operates approximately 4,000 vehicles across 10 U.S. metro areas, delivering roughly 500,000 paid rides per week.

Recently, Waymo began offering public rides in its 6th-generation Ojai vehicle platform across San Francisco, Los Angeles, and Phoenix. The company has announced plans to launch in more than 20 additional cities in 2026, including Tokyo, London, Nashville, Denver, Las Vegas, and New York City. Waymo is targeting one million fully autonomous rides per week by end of 2026.

The February 2026 funding round raised $16 billion at a $126 billion post-money valuation. Alphabet contributed approximately $13 billion -- more than 75% of the total -- as anchor investor, joined by Sequoia Capital, Dragoneer, DST Global, Andreessen Horowitz, Mubadala, Silver Lake, Tiger Global, T. Rowe Price, and Fidelity.

Waymo currently generates approximately $355 million in annualized revenue and operates at a loss as it scales.

Alphabet's roughly $2 trillion market cap does not depend on Waymo winning, and Alphabet can fund Waymo's losses indefinitely as a result.

The absence of that pressure is itself a competitive advantage. Waymo can expand deliberately, and enter each city on its own terms. Tesla's expansion cadence carries the weight of a valuation that has already priced in a future that has not arrived.

Tesla launched its Austin robotaxi service in June 2025 using Model Y vehicles with safety monitors. Unsupervised operation began in January 2026 on a limited basis, expanding to Dallas and Houston in April.

The fleet peaked at 33 unsupervised vehicles on May 4, with 19 in Austin, 6 in Houston, 5 in Dallas. But it has since contracted. As of late May 2026, approximately 20 active unsupervised vehicles remain across the three Texas cities.

Five additional cities have been announced: Phoenix, Miami, Orlando, Tampa, and Las Vegas. None have launched.

Tesla has disclosed approximately 700,000 cumulative paid robotaxi miles. Waymo has accumulated 127 million fully autonomous miles.

What that gap supports is a directional observation. Waymo is operating a commercial service and growing its scale. Tesla is running a controlled pilot that is a long way from scaling.

Cybercab production began April 24 at Gigafactory Texas, a milestone that signals real confidence in the program's commercial direction. Purpose-built vehicles without steering wheels or pedals represent a significant cost and design commitment.

It is worth noting that General Motors made a comparable commitment. The Cruise Origin - a purpose-built robotaxi - began production in September 2023 at a GM's factory near Detroit. Production was paused two months later following a safety incident that cost Cruise its California operating permits. GM killed the program entirely by July 2024, after investing approximately $10 billion over nearly a decade.

Manufacturing a purpose-built robotaxi vehicle and operating a commercially viable robotaxi business are not the same milestone. The Cruise arc is the most recent evidence of that distinction.

Tesla may in fact solve vision-only autonomy. The architecture of a single large end-to-end model trained on billions of real-world miles is a coherent engineering bet. Nvidia CEO Jensen Huang called it the most advanced autonomous vehicle stack in the world at CES 2026. That is not nothing.

But two things can be true simultaneously. Tesla may cross the technology threshold, and others will too. Waymo is already closer.

The question is not binary - Tesla solves it or Tesla doesn't. The question is whether solving it translates into market dominance, and on what timeline. A well-capitalized competitor is already at commercial scale and actively driving down costs.

Technology journalist Timothy B. Lee, who has covered autonomous vehicles for more than a decade at the Washington Post, Vox, and Ars Technica before launching his newsletter Understanding AI, argues that Tesla's billions of FSD miles are not equivalent training data for unsupervised robotaxi operation.

Consumer FSD data captures supervised driving with a human ready to intervene. Robotaxi operation requires edge case handling in a fully unsupervised environment. The domains overlap but are not identical.

The conventional framing is that Tesla's vision-only architecture is cheap. Whereas Waymo's sensor stack is expensive. Therefore Tesla has a permanent cost advantage. But this seems materially outdated today.

Waymo's 6th-generation Ojai robotaxi represents a dramatic reduction from the prior Jaguar I-Pace fleet estimated at $150,000 to $200,000 per vehicle fully. Cost estimates for the Ojai vary by source. Fifth Level Consulting puts the all-in figure at $50,000 to $55,000; investor Gary Black of The Future Fund estimates $75,000 delivered; Morgan Stanley's estimate is approximately $125,000.

The range reflects uncertainty.  What is not in dispute, however, is the cost direction. Waymo can now deploy significantly more vehicles per dollar of capital than it could 18 months ago.

The 6th-generation hardware suite uses 42% fewer sensors than the prior system while improving capability.

Tesla targets a retail sale price of under $30,000 for the Cybercab. That cost gap is real and Tesla still holds it. But the assumption that this gap is permanent and compounds indefinitely is not supported by Waymo's current trajectory.

A more useful question is at what fleet utilization rate does Waymo's cost structure become economically competitive with Tesla's on a per-mile total cost basis. 

The Ojai introduces a geopolitical exposure the cost analysis cannot ignore. The base vehicle is manufactured by Zeekr, a premium electric vehicle brand owned by Geely Holding Group -- the Chinese automotive conglomerate that also owns Volvo, Polestar, and Lotus. 

The Ojai is built on Zeekr's platform at a Geely facility in Ningbo, China, then shipped to Waymo's Mesa, Arizona facility as a stripped-down glider chassis for autonomous hardware installation.

The Commerce Department's Bureau of Industry and Security finalized a rule in January 2025, effective March 2025, prohibiting the import and sale of connected vehicle hardware and software supplied by entities subject to Chinese government influence.

Waymo's position is that the Ojai qualifies as a base vehicle, imported as a stripped-down chassis without Chinese-origin connected systems. Waymo says their proprietary technology is retrofitted at its Mesa, Arizona facility. That legal argument is being challenged, however.

At a February Senate Commerce Committee hearing, Senator Bernie Moreno called the arrangement a loophole and a backdoor. 

Tesla carries none of this exposure. Its Cybercab is manufactured at Gigafactory Texas. That supply chain advantage is entirely separate from the software capability question, and it is durable.

Even if Tesla solves vision-only autonomy, that does not mean Tesla captures the majority of robotaxi revenue and profit. Those are two separate outcomes. The bull case treats them as one. They are not.

Ride-hailing has never produced a winner-take-all market structure. Uber did not eliminate Lyft. Didi did not eliminate its competition in China. The global ride-hailing market fragmented across geographies, regulatory regimes, and consumer preferences. There is no structural reason to assume robotaxi economics will be different.

Waymo is already operating commercially across 10 U.S. cities. It is expanding to more than 20 additional cities this year, including its first international markets. WeRide, Baidu's Apollo Go, and others are expanding internationally in markets Tesla has not entered.

Tesla entering the robotaxi market with functional vision-only autonomy is meaningful. Tesla entering that market and capturing the majority of the economics requires believing that all other operators fail or retreat. That is a much stronger claim.

These conditions are worth stating precisely. Each is independently plausible. All three holding simultaneously, on a required timeline, is what Tesla’s valuation requires.

Not FSD with occasional teleoperator intervention. Fully unsupervised operation across diverse environments, adverse weather conditions, and edge cases at a safety record comparable to or better than Waymo.

If Tesla's vision-only architecture gets there - and demonstrates it through disclosed operational data, not press releases - the technology thesis resolves.

A Cybercab retail price under $30,000 versus an Ojai at somewhere between $50,000 and $125,000 is a real gap today. If that gap holds as Waymo's costs continue falling, and if Tesla's per-mile operating economics are demonstrably better than Waymo's at equivalent fleet utilization, then the margin structure argument has real weight.

Manufacturing scale is the condition that converts the cost advantage into market capture. You cannot claim a cost-per-vehicle advantage without the ability to produce at the volume the market share thesis requires.

This condition is structurally different from the first two. Tesla can influence whether it solves the technology. Tesla can influence its manufacturing output. Tesla cannot determine when European or Asian regulators approve vision-only unsupervised commercial deployment.

No major international market has yet approved vision-only architecture for unsupervised robotaxi operation. 

These are not permanent barriers. They are friction points requiring market-by-market resolution on timelines that are not Tesla's to control. What the current operational record does not support is treating those conditions as already resolved.

The robotaxi market will likely exist at meaningful scale. The technology will likely get there.

The more important question is who gets there, how fast, at what cost, and under what regulatory conditions.

Tesla may be one of the companies that gets there. What I do not think is established - analytically, not emotionally - is why Tesla captures the dominant share of the economics in a market where Waymo is already operating at scale, lowering costs, and expanding geographically.

I held Tesla shares from late 2018 to mid-2022. That reflects both my belief in the company's early automotive execution and why I'm comfortable questioning the valuation logic today. I no longer hold Tesla shares. I hold Google shares, which means I have a financial interest in Alphabet, Waymo's parent company. 

I am not telling you that to qualify what I have written here. You should weigh my position. You should also weigh whether the argument holds regardless of it.

The bet inside the bet is not whether Tesla solves autonomy. It is whether solving autonomy, in this competitive landscape, on this timeline, translates into the market dominance the valuation requires. The fandom variable, in both directions, has made it harder to see that clearly.

If you have a perspective or disagreement, reply directly. I read every response.