Nvida emails: Elon Musk diverting Tesla GPUs to his other companies

Elon Musk is yet again being accused of diverting Tesla resources to his other companies. This time, it’s high-end H100 GPU clusters from Nvidia. CNBC’s Lora Kolodny reports that while Tesla ordered these pricey computers, emails from Nvidia staff show that Musk instead redirected 12,000 GPUs to be delivered to his social media company X.

It’s almost unheard of for a profitable automaker to pivot its business into another sector, but that appears to be the plan at Tesla as Musk continues to say that the electric car company is instead destined to be an AI and robotics firm instead.

Does Tesla make cars or AI?

That explains why Musk told investors in April that Tesla had spent $1 billion on GPUs in the first three months of this year, almost as much as it spent on R&D, despite being desperate for new models to add to what is now an old and very limited product lineup that is suffering rapidly declining sales in the US and China.

Despite increasing federal scrutiny here in the US, Tesla has reduced the price of its controversial “full-self driving” assist, and the automaker is said to be close to rolling out the feature in China. (Questions remain about how many Chinese Teslas would be able to utilize this feature given that a critical chip was left out of 1.2 million cars built there during the chip shortage.)

Perfecting this driver assist would be very valuable to Tesla, which offers FSD as a monthly subscription as an alternative to a one-off payment. The profit margins for subscription software services vastly outstrip the margins Tesla can make selling physical cars, which dropped to just 5.5 percent for Q1 2024. And Tesla says that massive GPU clusters are needed to develop FSD’s software.

Isn’t Tesla desperate for Nvidia GPUs?

Tesla has been developing its own in-house supercomputer for AI, called Dojo. But Musk has previously said that computer could be redundant if Tesla could source more H100s. “If they could deliver us enough GPUs, we might not need Dojo, but they can’t because they’ve got so many customers,” Musk said during a July 2023 investor day.

Which makes his decision to have his other companies jump all the more notable. In December, an internal Nvidia memo seen by CNBC said, “Elon prioritizing X H100 GPU cluster deployment at X versus Tesla by redirecting 12k of shipped H100 GPUs originally slated for Tesla to X instead. In exchange, original X orders of 12k H100 slated for Jan and June to be redirected to Tesla.”

X and the affiliated xAi are developing generative AI products like large language models.

Not the first time

This is not the first time that Musk has been accused of diverting resources (and his time) from publicly held Tesla to his other privately owned enterprises. In December 2022, US Sen. Elizabeth Warren (D-Mass.) wrote to Tesla asking Tesla to explain whether Musk was diverting Tesla resources to X (then called Twitter):

This use of Tesla employees raises obvious questions about whether Mr. Musk is appropriating resources from a publicly traded firm, Tesla, to benefit his own private company, Twitter. This, of course, would violate Mr. Musk’s legal duty of loyalty to Tesla and trigger questions about the Tesla Board’s responsibility to prevent such actions, and may also run afoul other “anti-tunneling rules that aim to prevent corporate insiders from extracting resources from their firms.”

Musk giving time meant (and compensated) for by Tesla to SpaceX, X, and his other ventures was also highlighted as a problem by the plaintiffs in a successful lawsuit to overturn a $56 billion stock compensation package.

And last summer, the US Department of Justice opened an investigation into whether Musk used Tesla resources to build a mansion for the CEO in Texas; the probe has since expanded to cover behavior stretching back to 2017.

These latest accusations of misuse of Tesla resources come at a time when Musk is asking shareholders to reapprove what is now a $46 billion stock compensation plan.