Tesla Stock After Q3 2026 Earnings: Gross Margin Came Within Half a Point of My Test

Tesla reported its third quarter after the close on October 21. Revenue was $28.05 billion, flat on a year earlier, and free cash flow was $1.53 billion after a negative Q2. The CFO also cut 2026 capex guidance to approximately $20 billion, from more than $25 billion. The stock fell 3.1% on October 22 to $374.58.

After the delivery report I argued that a delivery beat feeds Tesla's one strong pillar, the driving data from its consumer fleet, and says nothing about a moat in the car itself. I said I'd be wrong if Tesla reported GAAP gross margin of 20% or more in Q3 or Q4 2026 with deliveries above 450,000, because a car business with pricing power is exactly what the moat file says Tesla doesn't have. Q3 came in at 19.6% on 486,532 deliveries. That's the closest the claim has come to breaking, and Q4 can still break it.

The gross margin recovery is real. Tesla added 2.8 points in a quarter while regulatory credits shrank to $113 million, so it didn't come from credits. Some of it came from energy, where gross margin jumped to 33.1% from 20.4% in Q2, helped, Tesla said on the call, by tariff refunds and warranty true-downs. Automotive gross margin excluding credits was 18.1%, up from 16.3% in Q2 and 15.4% a year ago. That's a better car business than in the spring.

None of that reached the bottom line year over year. Gross profit rose by about $440 million from Q3 2025, and operating expenses rose by $1.35 billion, so income from operations fell 56%. The CFO said opex will keep rising quarter after quarter until the new products are commercial. So the car business is earning more, and that money is paying for the AI build rather than showing up as profit.

Tesla added 260,000 FSD subscriptions in Q3, more than in any of the four quarters before it, and year-over-year growth sped up from 56% in Q2 to 67%. The CFO said more than 75% of North American deliveries signed up for paid FSD at delivery, and that about half of the 1.7 million or so paying customers are on monthly plans. Two caveats. Every North American delivery comes with a free month, so a sign-up at delivery isn't the same as a customer who keeps paying, and Tesla doesn't report retention. And Tesla doesn't break out FSD revenue, so I can count the subscribers but not what they're worth.

Still, this is the first line in the report where the fleet-data pillar shows up as something people pay for, and it's growing faster than anything else Tesla sells. Robotaxi is the other half of the story and it's still reported in miles. Tesla said on the call it operates in six cities, it put Cybercab into the Austin fleet on September 3, and it passed nearly 2 million cumulative unsupervised miles, up eightfold in the quarter. It didn't give a paid ride count or a fleet size, so my October test for the demand side, 250,000 paid rides a week by the Q2 2027 report, can't be checked yet.

The strongest case against me is in the table. A year ago Tesla booked $417 million of regulatory credits and made 18.0% gross margin. This quarter it booked $113 million and made 19.6%. Strip the credits out and the car business improved by more than the headline shows, with automotive margin ex-credits up 2.7 points in a year to 18.1%. Q4 2025 came in at 20.1% and Q1 2026 at 21.1%, so Tesla has cleared 20% recently. If Q4 2026 does it with deliveries above 450,000, my October claim breaks and I'll say so.

The bull would push further. The opex is a choice, they'd say. Tesla could stop spending on compute and Optimus and the operating margin would look like 2023 again. The capex cut and positive free cash flow show the build is funded from operations, with $45.4 billion in cash and nothing drawn on its new credit lines. Tesla's own LFP cells have also started shipping in Megapacks. And 1.74 million people paying for FSD is a customer base no other carmaker has, which could feed the robotaxi network if owners add their cars to it.

Here's where I come down. The Q4 test is live and close, and I won't move the bar. But both recent quarters above 20% came on fewer than 450,000 deliveries, Q4 2025's 20.1% also had $542 million of credits, and Q3's energy margin had one-offs in it, so a repeat at full volume isn't automatic. The capex cut looks like a delay: Musk said spending becomes more meaningful next year and reaches "tens of billions in '28" for the Terafab chip plant. And an owner paying to have their own car drive them isn't the same as a rider choosing Tesla's app over Waymo's or Uber's. I read Q3 as the data pillar starting to earn, with the car margin still short of the bar.

Tesla reports Q4 in late January. Gross margin against 20%, with deliveries above 450,000, settles the October claim either way. Then the FSD count: if it's under 2.2 million in the Q2 2027 report in July, the line where the data pillar earns is slowing and I've misread where the story sits. I'll also look for the first paid-ride figure, and if Tesla publishes one it gets a chart here.