Musk is unusually qualified to understand those relationships. Nvidia’s processors power the AI ambitions of Tesla, SpaceX and xAI. Huang personally delivered the first DGX-1 AI supercomputer to Musk and OpenAI in 2016. Nvidia later helped satisfy Musk’s vast appetite for GPUs as he raced to build xAI’s computing clusters.
Musk is also an interested witness. If Nvidia loses access, allocation or negotiating power without Huang, Musk’s own companies could bear the consequences. His campaign can, therefor,e be read as admiration, alarm and self-preservation at the same time.
The question is no longer why Musk likes Jensen Huang. It is whether Musk’s AI empire still needs him.
Nvidia sits at the centre of AI, but it does not manufacture its most advanced chips by itself. It depends upon foundries including TSMC and Samsung, buys memory from SK Hynix, Micron and Samsung, and relies on specialist packaging, assembly and systems partners. Its annual report names Hon Hai, Wistron and Fabrinet among the companies involved in the chain.
Capacity must be reserved long before a finished AI system reaches a customer. Nvidia disclosed in July that its supply and capacity commitments had increased from $119 billion to $279 billion in a single quarter as it attempted to secure inventory for future demand.