Elon Musk’s Nvidia AI chip order is too big to handle

Elon Musk has told Nvidia to prioritize sending their powerful AI chips to his social media company, X, and his AI startup, xAI, rather than to Tesla. This decision, revealed through internal emails obtained by CNBC, shows Musk’s challenging act of balancing his different companies.

Busy days for Elon Musk: He just announced that Tesla plans to spend $3 billion to $4 billion this year on Nvidia hardware too. However, it turns out Tesla can wait or have to wait. Here is why.

Elon Musk’s hard quest with Nvidia’s AI chips

Musk’s decision to redirect Nvidia’s AI chip, specifically the highly sought-after H100 chips, from Tesla to X and xAI has significant implications. Tesla, which plans to spend $3 billion to $4 billion this year on Nvidia hardware as part of its $10 billion AI budget, now faces delays. These delays could slow down Tesla’s efforts to enhance its supercomputing systems, which are crucial for developing self-driving cars and robots.

Of the roughly $10B in AI-related expenditures I said Tesla would make this year, about half is internal, primarily the Tesla-designed AI inference computer and sensors present in all of our cars, plus Dojo.

For building the AI training superclusters, NVidia hardware is about…

— Elon Musk (@elonmusk) June 4, 2024

Of course, Musk has big dreams for Tesla. In a recent earnings call, he announced plans to increase the number of active H100 chips at Tesla from 35,000 to 85,000 by the end of the year. Surprisingly, he also admitted that it is a hard job to do. More good news for Nvidia!

Correct.

Also, I can’t overstate the difficulty of making 50k H100s train as a coherent system. No company on Earth has been able to achieve this yet.

Once we have that system working, we will order more hardware.

— Elon Musk (@elonmusk) June 4, 2024

That’s probably why Elon Musk thinks he has to prioritize his company’s needs. Also, Tesla has one more problem with AI chips right now: there is no room for them yet!

Tesla’s Texas gigafactory, which is undergoing expansion, will eventually house 50,000 H100 chips for training AI models. Until the expansion is complete, there simply isn’t enough space to deploy all the GPUs. This logistical challenge has prompted Musk to temporarily prioritize X and xAI, which also have significant AI needs.

xAI’s need for AI chips

Elon Musk wants AI chips for xAI because he believes in the power of artificial intelligence to change the world. xAI, his new AI startup, aims to create smart computer programs that can do things like have conversations or solve problems.

Musk also wants xAI to be a place where people can make AI programs that aren’t afraid to say controversial things. He believes this kind of freedom can lead to new and exciting ideas in the AI world. To do so, they have to be really clever, that’s where Nvidia involved the story.

The firm you have called can not be reached at the moment

Nvidia is also grappling with the high demand for its AI chips, a situation intensified by Musk’s recent directives. An internal email from December underscores Musk’s instruction to prioritize X’s GPU needs over Tesla’s, a move that has disrupted Nvidia’s distribution plans and created internal friction. Another email from April highlights significant discrepancies between Musk’s public statements about Tesla’s AI chip procurement and the actual bookings recorded by Nvidia, further causing concern within the company.

Nvidia’s CEO, Jensen Huang, has publicly acknowledged the immense challenge of meeting the soaring demand for GPUs.

The unprecedented growth in AI development across industries has left the company struggling to allocate resources fairly and effectively. Major tech players are all competing for a limited supply of Nvidia’s advanced processors. This intense competition has made it difficult for Nvidia to fulfill orders promptly, underscoring the critical importance of these chips in the race to advance AI technologies.

All images are generated by Eray Eliaçık/Bing Image Creator