A high-stakes race for artificial intelligence infrastructure is intensifying, with Nvidia pouring billions into OpenAI and Elon Musk pledging to scale compute to levels never before seen.
The chipmaker’s up to $100 billion commitment to OpenAI has reshaped expectations across Silicon Valley and Wall Street, accelerating pressure on rivals to match spending or risk falling behind.
The deal involves roughly 10 gigawatts of new capacity of Nvidia systems for OpenAI that could power the needs of over eight million US households.
In response, xAi founder Elon Musk says on X that his company will be the first to deliver unprecedented scale.
“Just as we will be the first to bring a Gigawatt of coherent training compute online, we will also be the first to 10GW, 100GW, 1TW, …”
Deepwater Asset Management co-founder Gene Munster says the deal will add pressure to other big AI firms and quicken the tempo of competition.
“And my sense is that the reason why this is such an important announcement is that by Nvidia investing in OpenAI and effectively creating more demand for data centers, more demand for cloud compute, more demand for AI inference and training, that basically quickens the pace that all the other players are going to have to respond to. You saw the response that Elon had to a comment on X today that this was concerning, that Nvidia was making this investment in OpenAI. And I think it speaks to kind of this bar inching up.
And if you look at Nvidia’s core customers, they have six core customers. One of them, Meta, has been very clear that they’re going to be spending a lot. Call it 45% plus in each of the next three years when it comes to CapEx. The other companies have had more modest expectations: Google, Amazon, Microsoft, talking about CapEx increasing at 7% next year. And so when I think about this announcement with Nvidia and OpenAI, and it’s staged over several years, but I think ultimately it speaks to this race, this arms race, this bet that companies need to make, either they need to play and meet each other more or less one for one when it comes to the AI race, or make an incredibly difficult bet to dial back some of that spending.”