Analysis: Musk is deploying too much computing power, and competitors are rushing to grab the bargains.

According to Beating Monitoring, senior research analyst Harrison Rolfes from PitchBook said in an Axios interview that Musk repeatedly overexerts himself on computing infrastructure: building at maximum scale but the products can’t handle it, and the excess capacity is eventually taken by competitors. The earliest instance was in 2024. xAI discussed a roughly $10 billion server leasing deal with Oracle to train Grok 3. Musk found Oracle’s cluster building too slow, the deal fell through, and he turned to build the Memphis data center himself. The GPU capacity freed up by Oracle was signed over to OpenAI. Rolfes’s exact words: “xAI’s Colossus 1 ultimately ended up with more capacity than Grok users could ever fill.” Musk not only relies on procurement to stockpile chips for xAI. CNBC obtained internal Nvidia emails in 2024, showing he instructed Nvidia to prioritize reallocating the 12k H100s originally meant for Tesla to X and xAI. Tesla’s over $500 million worth of chips were delayed by several months, slowing down autonomous driving and robot training. The chips taken were ultimately not fully used. Two years later, the same thing happened again. All 220k+ Nvidia GPUs of Colossus 1 were leased out to Anthropic. SpaceXAI claims that training has moved to a larger Colossus 2, so Colossus 1 was cleared out. However, The Information previously reported that xAI’s model compute power utilization with tens of thousands of GPUs was only 11%, and Grok’s user base couldn’t support this scale of computing power. It wasn’t an upgrade that was actively vacated; it was capacity that was never fully used in the first place.