Tesla|Robotaxi Miles Flatline as 25B AI Bet Triggers First Cash Burn in Two Years

The 14% Crash Behind a Revenue Beat

Tesla shares plunged as much as fourteen percent Thursday morning, the steepest single-day drop the stock has taken in more than a year. The strange part is what triggered it: revenue of twenty-eight point two billion dollars beat every analyst estimate on the board, and deliveries jumped twenty-five percent to just over four hundred eighty thousand vehicles. A record quarter by the numbers that matter to a car company sent the stock into its worst session since June of last year.

The market wasn't reacting to the top line. Adjusted earnings came in at thirty-three cents a share against a Wall Street estimate near fifty-one cents, and free cash flow turned negative for the first time in two years, landing at a loss of one point one billion dollars. Capital expenditures more than doubled from a year earlier to five point eight billion dollars in the quarter alone. The bottleneck sitting underneath that number is Tesla's Robotaxi buildout, and by sentence three of Musk's own earnings call, that buildout was already the thing analysts were pressing him on.

Musk didn't walk any of it back. On the call he told analysts this is a massive capex year and that Tesla should be spending as fast as it can without being wasteful, reiterating a full-year capital expenditure target above twenty-five billion dollars. That spending funds Optimus humanoid robot production, Cybercab manufacturing, AI compute expansion, and the Robotaxi fleet all at once. Operating income fell fifty-seven percent to three hundred ninety-eight million dollars, and operating margin compressed from four point one percent to just one point four percent.

JPMorgan, Cantor Fitzgerald, and Mizuho all cut their price targets within hours of the report. Canaccord Genuity's George Gianarikas put it bluntly, calling Tesla currently a story of stagnant margins, negative free cash flow, and breakthrough promises anchored to timelines that are hard to measure and model. That last phrase is the hinge. The promise Musk is asking investors to fund isn't the car business anymore. It's Robotaxi, and Robotaxi is exactly where the next chapter's numbers stop cooperating with the story.

The Chart That Hides a Stall

Tesla's shareholder letter led with a chart showing cumulative paid Robotaxi miles climbing past two point four million, offered as proof the service is expanding. A cumulative chart, by construction, only ever slopes upward no matter what happened in any given quarter. Break the same numbers into quarter-over-quarter additions instead of a running total, and the picture Tesla chose not to draw comes into view.

Tesla added roughly one hundred fifty thousand paid Robotaxi miles in the third quarter of last year, four hundred fifty thousand in the fourth, and nine hundred thousand in the first quarter of this year. Then it added nine hundred thousand again in the second quarter. The ramp that ran through the end of last year simply stopped. Growth didn't slow down gently, it went flat, and April alone accounted for roughly five hundred thousand of that quarter's total before May and June each added only about two hundred thousand miles.

The active unsupervised fleet across Austin, Dallas, Tampa, and Orlando is roughly twenty-one vehicles combined. Austin itself peaked near twenty-five cars in late April and has since drifted down to about seventeen. Tesla's response to the plateau was to add service areas in Tampa and Orlando the day before earnings without disclosing vehicle counts in either city, spreading the same small fleet across more dots on a map rather than adding cars to the road.

The disagreement is now in the record, not just in the data. Barclays analysts wrote after the call that more is needed to show Tesla's capabilities are inflecting, noting the company offered no new expansion targets and hasn't met the ones it previously laid out. Tesla's own VP of AI, Ashok Elluswamy, countered on the same call that growth in Robotaxi miles driven is, quote, literally exponential, just early in the exponential curve. One side is reading the disclosed numbers. The other is reading the chart Tesla chose to publish.

A Merger Rumor Reframes the Sell-Off

While Tesla investors were digesting the earnings miss, a separate conversation was building around a possible Tesla-SpaceX combination. SpaceX has fallen roughly thirty-five percent from its post-IPO peak after a record eighty-six billion dollar listing in June. On the earnings call, Musk again declined to rule out a merger, pointing to growing overlap between the two companies through the shared Terafab chip project and Starlink's role in autonomous vehicle connectivity. Deepwater co-founder Gene Munster raised his estimate of merger odds within the next few years from eighty to ninety percent immediately after hearing Musk's answer.

Musk's actual words left the door open without confirming anything. He told analysts that, as they can tell from all the collaborations with SpaceX, there's more and more overlap, but that combining companies isn't something that can be discussed on an earnings call and has to go through the appropriate process. Any real transaction would draw regulatory scrutiny and raise governance questions for both sets of shareholders, and no timeline or structure has been offered by either company.

The genuine risk in this quarter isn't the miss itself, it's whether the Robotaxi plateau proves structural before the spending that depends on it comes due. If the next monthly fleet and paid-mile disclosures out of Austin, Dallas, Tampa, and Orlando show the count climbing again rather than holding flat, the sell-off looks like a mispriced entry into the same capex story Musk is defending. If those same numbers stay flat into the release of Full Self-Driving version fifteen, which Musk has tied to any large-scale unsupervised rollout, the plateau becomes the evidence that the twenty-five billion dollar capex case was ahead of the technology it was funding. Watch the fleet count city by city before the next quarterly print arrives, that is the number that resolves this, not the merger talk around it.