Tesla’s AI teams brace for their ‘hardest year’ yet: 2026

For a company that sells self-driving cars, Tesla still relies on human endurance. The automaker’s AI boss just told the engineers behind Autopilot, Robotaxis and the Optimus robot to prepare for the “hardest year” of their lives in 2026, as the company strives to deliver the autonomous future it has been promising for a decade.

Within Tesla’s AI division, this line landed with the force of a deadline. According to a report from Business Insider, Ashok Elluswamy, the company’s vice president of AI software, delivered the message during a meeting between the Autopilot and Optimus teams — a meeting that lasted nearly two hours. One participant described it as a “rallying cry.”

The timelines outlined on the slides appear to leave no ambiguity: 2026 is the year Tesla needs to turn its decade-long promises of autonomy into something commercial, scalable and visible to investors who have already heard these claims.

The pressure is structural and not motivational. CEO Elon Musk has now tied the company’s next era — and its new $1 trillion pay package — to two products that are far from being deployed at scale: a nationwide fleet of robotaxis and a humanoid robot, which Musk says could become “the biggest product in history.”

During Tesla’s latest earnings conference call, he pushed the stakes even higher. “We anticipate that in the coming months, Austin’s autonomous taxis will operate without a safety driver,” Musk said, adding that by the end of the year, Tesla aims to have these robot taxis operating in “about eight to 10 major metropolitan areas.” This goal would be ambitious for a company that already offers a fully driverless service; for Tesla, which still uses supervised drivers, it’s a moonshot.

Wedbush analyst (and longtime Tesla bull) Dan Ives said earlier this year that he believes “the AI ​​and autonomy opportunity alone is worth at least $1 trillion to Tesla” and that he has never “considered Tesla simply as an automaker” but rather “has always viewed Musk and Tesla as a leading global player in disruptive technology.” The deadlines have not yet lived up to these predictions.

Optimus carries his own weight. Production is scheduled for late 2026, a timeline that Musk said will “take a while” to come together because the robot contains “10,000 unique items.” Elluswamy now oversees more of that program following the departure of Optimus vice president Milan Kovac. Musk said on the latest earnings conference call that he believes “Optimus could become the most important product in history,” presenting the humanoid robot less as a science project and more as Tesla’s eventual center of gravity.

Tesla has promised 1,000 robo-taxis on the roads by the end of the year. He promised robotaxis in eight to ten cities. He promised the deployment of Optimus on an industrial scale. He promised – implicitly, explicitly and repeatedly – ​​that Tesla should be seen as a real-world AI company first, and then as an automaker. Meanwhile, margins tightened, talent churn increased, and the company quietly dismantled its internal Dojo supercomputer team, choosing instead to buy the same Nvidia and AMD hardware used across the industry.

All of this puts enormous weight on the people within the division that Musk claims to meet with every week, often late into Friday night. If Tesla wants to bridge the gap between aspiration and reality – between a supervised driver assistance system and a national network of robo-taxis, between organized robot demonstrations and a production goal of one million units of humanoids – 2026 is not just a difficult year. This is the year on which the entire valuation is based. And now, next year, Tesla will have to exceed its own scenario.