Elon Musk’s ‘5x Value’ Self-Driving Math Has a Fatal Economic Flaw

I keep seeing Elon Musk’s 2022 Baron Conference clip resurface on X, where he explains why Tesla is a better investment than Mercedes-Benz using “net present value of future cash flows.” The same argument he made at the March 2025 All-Hands meeting when Tesla stock was down 40%:

“It’s very difficult for people in the stock market to imagine a future where suddenly a 10 million-vehicle fleet has five to 10 times the usefulness.”

I’ve been covering Tesla’s autonomy promises for years, and I’m genuinely surprised nobody has pointed out the fundamental economic flaw in this argument. So let me be the one to say it: Musk’s “5-10x more valuable” thesis contains basic errors that any first-year economics student should catch.

And here’s the kicker: His theory doesn’t just break down because of competitors like Waymo or BYD. It breaks down because of Tesla’s own fleet.

• What: Musk claims self-driving software will make Tesla vehicles 5-10x more valuable
• When: He’s made this argument consistently since 2019, most recently at the March 20, 2025 All-Hands meeting
• The problem: He’s pricing autonomy against today’s human-driver costs, ignoring that supply explosion crashes prices
• Why it matters: This narrative underpins Tesla’s trillion-dollar valuation, and it’s economically incoherent

A Pattern of Over-Promising

Before diving into the economics, let’s establish the track record. Musk’s autonomy predictions have a long history of optimism bias:

• 2019: “1 million robotaxis by 2020”
• 2020-2024: Multiple claims of unsupervised FSD arriving “next year”
• Late 2025: “Unsupervised robotaxis in Austin by year-end” with safety monitors removed “in approximately 3 weeks”
• January 8, 2026 (today): That year-end deadline was missed. No public, commercial unsupervised robotaxi service has launched.

Testing has advanced. Internal “perfect” drives with empty driver seats occurred in Austin in late December 2025. But operations remain limited and supervised, with broader rollout now pushed to later in 2026, including the Cybercab production ramp starting around April.

This pattern of delays matters because every missed deadline is another quarter where competitors like Waymo (now at ~450,000 weekly driverless rides) extend their operational lead. The “5-10x value” thesis requires Tesla to win the autonomy race. The scoreboard suggests otherwise.

The Transitional Value Fallacy

Musk’s math works like this: A human-driven Uber costs roughly $2-3 per mile because you’re paying for the driver’s time. A self-driving Tesla robotaxi could charge similar rates while eliminating the driver, capturing that labor value as pure profit. Multiply that by a 10-million-vehicle fleet operating 24/7, and the numbers look astronomical.

But here’s what this analysis misses: Musk is using transitional arbitrage as the basis for permanent valuation. He’s comparing autonomous transport costs to today’s human-driver costs and assuming Tesla captures that entire spread forever.

That’s not how markets work.

Tesla vs. Tesla: The Internal Competition Problem

Here’s what makes this argument particularly weak: Even if Tesla achieved a complete monopoly on autonomy, with zero competition from Waymo, BYD, or Chinese manufacturers, Musk’s theory still self-destructs due to Tesla’s own fleet.

Consider the supply dynamics:

• Personal cars today average roughly 1 hour of use per day. They sit idle the other 23 hours.
• Human-driven rides are limited by driver availability. People need sleep, breaks, and lives. An average Uber/Lyft driver might do 100-150 miles per day profitably.
• Autonomous Tesla vehicles could theoretically operate 20+ hours per day, delivering 5x+ the miles per vehicle. Musk himself has cited this utilization increase as the value driver.

If Tesla enables millions of owner-opted-in vehicles plus company-owned Cybercabs onto the network, supply floods the market. Basic supply and demand: Massive increase in available rides means prices plummet toward marginal cost.

What’s the marginal cost? Per Musk’s own estimates: electricity at roughly $0.05-0.10 per mile, plus depreciation, maintenance, and insurance totaling approximately $0.10-0.20 per mile. Call it $0.20-0.30 per mile total operating cost.

Tesla might charge $0.30-0.50 per mile to riders, taking a 20-30% platform cut like Uber today. Great for consumers who currently pay $2+ per mile. But that “arbitrage” profit spread Musk extrapolates from current human-driver costs? It evaporates almost immediately once scale hits.

It’s like Airbnb: Early hosts made bank when supply was scarce. Now, in saturated markets, nightly rates are competitive and margins are thin. The same dynamic will hit robotaxi yields the moment Tesla’s own fleet scales.

External Competition Makes It Worse

Add external competition and the picture gets even bleaker for the “5x value” thesis.

Once autonomy is solved, whether by Tesla, Waymo, BYD, Baidu, Pony.ai, or others, the technology gets competed away. The economic value flows to consumers in the form of lower transport costs, not to producers like Tesla in the form of sustained margins.

Waymo is already at ~450,000+ weekly paid driverless rides as of late 2025, expanding aggressively. Chinese players like Baidu and Pony.ai are scaling cheaply. As these competitors grow, margins compress further, and value flows to consumers via cheap rides rather than sustained producer rents.

This is textbook producer surplus versus consumer surplus dynamics. In competitive equilibrium, price converges toward marginal cost. The “5x value” Musk cites would only hold if Tesla maintained a permanent monopoly on autonomy, which is implausible given the competitive landscape.

Consider the historical analogies: Air conditioning was once a luxury car feature commanding thousands in premium. Now it’s standard, with zero pricing power. Automatic transmission used to be expensive. Now manual is the oddity. GPS navigation commanded hundreds. Now it’s expected.

Self-driving will follow the same trajectory. Once ubiquitous, it becomes the baseline, not the differentiator. You won’t pay a premium for autonomy. You’ll pay a discount penalty for NOT having it.

Induced Demand Has Limits

Tesla bulls might counter: “Cheap rides will induce massive new demand! People will take way more trips if transportation is nearly free!”

There’s some truth here. Cheaper rides will encourage more trips, potentially replacing car ownership for some and enabling new use cases for leisure and errands. The total addressable market grows.

But induced demand hits real-world caps:

• Time budgets: People still need to sleep, work, and live. There are only so many hours for travel.
• Congestion: More cheap rides means more vehicles on roads. Infrastructure has physical limits.
• Practicality: Not everyone wants endless rides. Many trips aren’t discretionary.

The total addressable market grows, but not infinitely. And growing the pie doesn’t automatically translate to trillions in added valuation without durable moats to capture that growth.

The Optimus Problem Is Even Worse

Musk’s robot argument suffers from the same flaw, perhaps more severely. At the same March 2025 All-Hands, he claimed Optimus robots would usher in an era of “sustainable abundance” where “you can literally just have anything you want.”

If Optimus robots truly “unconstrain labor” as Musk claims, what happens to prices in that economy? They collapse.

If labor costs approach zero across the economy, the prices of goods and services fall accordingly. The robots aren’t creating “unlimited value.” They’re creating unlimited output at collapsing prices. GDP measured in units might explode. GDP measured in dollars might not.

This is essentially a deflationary scenario, and it’s unclear how Tesla captures outsize returns when everyone has access to cheap robotic labor.

Where Musk Has a Partial Point

I’ll give partial credit where it’s due. The transition window matters. If Tesla achieves L4/L5 autonomy 2-3+ years before competitors, they can extract significant rents during that period. First-mover advantage in a winner-take-most market with network effects from data could extend this window.

Tesla bulls argue that network effects and data moats from a 10-million-vehicle fleet could create durable advantages, similar to how Apple captured value from commoditized features through ecosystem lock-in. Recent progress with FSD v14 improvements and internal unsupervised tests shows momentum, and Tesla’s vision-only approach could scale cheaper long-term than Waymo’s expensive sensor suite.

But that’s a timing argument, not a permanent valuation argument. And timing arguments require execution. Tesla’s track record of missed deadlines, documented above, makes the “we’ll be years ahead” thesis harder to defend. As of today, Waymo has commercial driverless operations in multiple cities. Tesla has supervised pilots in Austin.

The ‘Appreciating Asset’ Track Record

This isn’t new territory for Musk. Back in April 2019, he made an even bolder claim: “If you buy a Tesla today, I believe you are buying an appreciating asset, not a depreciating asset.”

Nearly seven years later, the data is in. According to Cox Automotive, used Tesla values have depreciated significantly, not appreciated. The Model 3, priced at $35,000 in early 2019, saw used values surge during the pandemic, then fall to around $29,000. Tesla’s own FSD price went from continuous increases to cuts and free trials, as we covered with the 1.5 million owner trial.

The appreciation thesis died. Now Musk is recycling the same logic with bigger numbers.

EVXL’s Take

Here’s what I expect: Tesla will eventually deploy unsupervised FSD more broadly, likely sometime in 2026. Musk will declare this proves his “5-10x value” thesis. But the moment Tesla scales its robotaxi network with millions of vehicles, the very success of that deployment will crash per-mile economics toward marginal cost.

Early Tesla owners who opt their cars into the network might see decent yields initially, like early Airbnb hosts. But as more vehicles join, supply floods the market, and everyone’s margins compress. Tesla’s 20-30% platform cut helps but doesn’t preserve the massive arbitrage spread versus today’s $2+/mile human-driven rides indefinitely.

Musk is confusing social value creation (huge) with private value capture (much smaller in competitive equilibrium, and smaller still when your own fleet is the competition). Self-driving and robots will be enormously beneficial to humanity. Cheap, abundant transportation will transform how we live. That doesn’t mean Tesla shareholders capture that benefit at 5-10x current levels.

The robotaxi narrative drives much of Tesla’s premium valuation today. But execution risks and pricing dynamics make the promises feel stretched. The “5x value” pitch works in investor presentations. It doesn’t survive contact with Economics 101, the history of missed deadlines, or even the logical implications of Tesla’s own success.

Do you think Musk’s self-driving valuation math holds up? Let us know in the comments below.