
Elon Musk provided a specific reason for redirecting Nvidia AI chips from Tesla to X Corp and xAI Corp. He stated that Tesla had no place to put the Nvidia chips where they could be turned on, suggesting that the chips would have just sat in a warehouse if they were not redirected. Musk made this statement in a series of posts on X, responding to a report by CNBC. The report had cited an Nvidia memo from December, which mentioned that 12,000 of the chipmaker's H100 graphics processing units originally slated for Tesla were redirected to X instead.

According to the Nvidia memo, 12,000 Nvidia H100 graphics processing units were originally designated for Tesla but were redirected to X.

Nvidia staff noticed inconsistencies between Elon Musk's statements and Tesla's bookings with the chipmaker regarding the procurement and usage of Nvidia's H100 graphics processing units (GPUs). In an email from late April, Nvidia employees mentioned that Musk's comment on Tesla's first-quarter earnings call and his post on X about spending $10 billion on AI conflicted with the company's bookings and forecasts for the fiscal year 2025.
During the earnings call, Musk said that Tesla had around 35,000 active H100 GPUs and expected to have approximately 85,000 working by the end of the year. However, Nvidia staff pointed out that this statement conflicted with their internal records. Additionally, Musk's claim about Tesla spending $10 billion on AI in 2024 was also questioned by Nvidia employees, as it did not align with their forecasts4.
Furthermore, it was revealed that Musk diverted a significant shipment of AI processors, which were initially reserved for Tesla, to his social media company X. This decision resulted in Tesla's receipt of over $500 million worth of graphics processing units being delayed by months, potentially affecting the development of autonomous vehicles and humanoid robots.