Musk Says AI5 Goes to Optimus, Tesla Cars Stay on AI4

Tesla’s next self-driving chip, AI5, is going to its Optimus humanoid robot rather than its cars, Elon Musk said on X on Sunday: “It’s going to Optimus. Cars will remain at AI4. That’s enough to solve superhuman driving.”

He was answering a Tesla owner who asked when AI5 would be in new cars. Tesla’s research spending rose 69% to $2.76 billion in the third quarter, with Optimus and silicon development among the drivers, its 10-Q shows. For background, see our Tesla business strategy guide.

Tesla has not published a separate statement on AI5’s use.

Business Pill · The Model and the Harness

An evergreen Business Engineer explainer. It teaches the general idea only and does not discuss this report.

The key insight: Elon Musk said Tesla’s AI5 chip is going to Optimus and cars will stay on AI4, as Tesla’s R&D expense rose 69% to $2,761 million in Q3 2026 with Optimus and silicon development among the drivers.

What Musk Said

The question came from an owner on X: “When will AI5 be in new cars? I feel like I should wait to upgrade from my AI3”, the post read. Musk replied about seven minutes later.

AI3 and AI4 are earlier generations of Tesla’s in-car computer. Musk’s reply means a buyer waiting for AI5 in a car would, on his account, be waiting for a chip the cars will not get.

Tesla's R&D expense rose to $2,761 million in Q3 2026 from $1,630 million a year earlier, up 69%, with Optimus
Tesla’s R&D expense rose to $2,761 million in Q3 2026 from $1,630 million a year earlier, up 69%, with Optimus and silicon development among the drivers, its 10-Q says.

What Tesla’s Filings Show

Tesla’s Q3 2026 update says “Optimus production began in Fremont on our high-volume lines”, and that the bots coming off the line are the first production-quality bots it has made. It lists Optimus in Texas as under construction.

Research and development expense was $2,761 million in the quarter, up from $1,630 million a year earlier, or 10% of revenue against 6%. Tesla’s 10-Q says the rise came “primarily due to increases in costs related to SI and other programs, such as Optimus, silicon development, battery technologies, and new vehicle programs”.

The Compute Behind Both

Tesla lists more than 105 MW at Cortex 1 and more than 120 MW at Cortex 2 in Texas, at least 225 MW of training compute in production. It says: “We plan to reach 400 MW of SI training compute by the end of the year, supporting the development of both vehicle and humanoid robot autonomy software.”

By our arithmetic, the listed capacity is about 56% of that 400 MW goal, leaving up to about 175 MW to add in the last weeks of the year.

Bar chart: Tesla lists more than 105 MW at Cortex 1 and more than 120 MW at Cortex 2 in production, against a 400 MW end-2026 training compute target. Data: Tesla Q3 2026 Update
Tesla lists at least 225 MW of training compute in production against a 400 MW year-end goal, about 56% by our arithmetic.

Tesla also says it continued building its Austin Fab, “where we plan to start commissioning equipment in 2027.” The FSD software those cars run is in our FSD v15 report; Tesla’s factory spending is in our Grünheide piece.

The Structural Read

As we read it, Musk is ranking the robot ahead of the car for Tesla’s newest chip.

As we read it, the reply also tells AI3 and AI4 owners not to wait for AI5 in a car.

Elon Musk on X, 22 Nov 2026

“It’s going to Optimus. Cars will remain at AI4. That’s enough to solve superhuman driving.”

Three Implications

FOR TESLA OWNERS Musk said cars will remain at AI4, so AI5 is not coming to cars on his account.

FOR INVESTORS Tesla’s Q3 R&D rose 69% to $2,761 million, 10% of revenue, its 10-Q shows.

FOR ROBOTICS WATCHERS Tesla says Optimus production began in Fremont on its high-volume lines.

What Is Not Established

Musk’s reply is a post on X, not a company filing. Tesla has not said when AI5 will ship, how many chips it plans to make or what they cost.

The claim that AI4 is enough for superhuman driving is Musk’s; Tesla has not published data showing it.

This Week In The AI Supercycle

A World Without an Order, and an AI Central Bank Without a Successor

This week’s issue: the week AI memory stayed largely sold out while the people holding the shares paid: why the AI Supercycle is not a dot-com replay, read through geopolitics, macro and technology, with the gap between building and financing the widest we’ve recorded.

Read the full issue →

Business Engineer Framework

The AI Supercycle

The Business Engineer map of who builds, finances and captures value in AI.

Read the Map of AI →

The Bottom Line

Musk says Tesla’s next chip goes to the robot, not the car, while research spending rose 69% in a quarter with Optimus and silicon among the drivers. For car buyers waiting on new hardware, his answer is that AI4 is the end of the line for now.

97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. Musk’s statement and the question it answered are posts on X, read on 22 November. Production, compute and R&D figures are from Tesla’s Q3 2026 update and 10-Q. The 225 MW total, the 56% share and the 175 MW gap are our arithmetic; Tesla lists its Cortex figures as minimums. Tesla has not issued a statement on AI5.

Sources: Elon Musk on X, reply on AI5, 22 Nov 2026 · @JBrunner8484 on X, question on AI5 in new cars, 22 Nov 2026 · Tesla, Q3 2026 Update (Form 8-K Exhibit 99.1), 21 Oct 2026 · Tesla, Form 10-Q for the quarter ended 30 Sep 2026

Scroll to Top
FourWeekMBA