Elon Musk claims Tesla is investing in digital infrastructure to help train its humanoid robot, Optimus.

Tesla CEO Musk told investors on Wednesday that training the robot, which the company intends to market to businesses and consumers, is much more complex than training one of the firm’s vehicles.

“We continue to invest in training infrastructure out of our Texas headquarters,” Musk said. “The training needs for the Optimus humanoid robot are probably at least 10x of what is needed for the car to get to the full range of useful roles.”

A humanoid robot has “probably 1,000 times” more functions than a car, he added. “That doesn't mean the training scales by 1,000, but it's probably at 10x,” Musk said.

Musk was speaking on the company’s earnings call, which followed a set of disappointing financial results that did not match market expectations. He did not go into detail on what sort of infrastructure the company required, but said working on Optimus “doesn’t mean Tesla's going to spend $500 billion on training computers,” referencing Stargate.

He continued: “The cost of training is dropping dramatically with time. But it's one of those things where I think, long-term, Optimus has the potential to be north of $10 trillion in revenue, it's really bananas.

“You can obviously afford a lot of training compute in that situation. In fact, even $500 billion [of] training compute in that situation will be quite a good deal.”

It should be noted that Musk has been talking up the potential of Tesla’s humanoid robot, formerly known as Teslabot, for several years, but the machine has yet to make it to market. When it was revealed in 2021, the billionaire claimed it would be with customers the following year.

Last July, in a post on X, he said a small number of the robots would be in internal use at Tesla this year, and ready for market in 2026.

Tesla spent more than $10 billion on capex in 2024, and expects a similar outlay in each of the next two years, though much of this relates to infrastructure for its electric vehicles. On the earnings call, CFO Vaibhav Taneja said the company’s total AI-related capex outlay, including infrastructure, is now more than $5 billion.

In October, Tesla said it was bringing a 50,000-strong cluster of Nvidia H100 GPUs online at its Texas headquarters. The project was reportedly delivered behind schedule, with Musk firing the construction manager in April.

The company also has its own D1 custom chips for autonomous driving training. Tesla's fully autonomous efforts are also several years behind schedule.

It is perhaps no surprise that Musk is hoping that robots will boost Tesla’s bottom line given that it saw year-on-year revenue from its vehicle business decline eight percent. Overall revenue was up two percent year-on-year, to $25.71 billion, less than the $27.26bn predicted by analysts.