Nvidia’s graphics processing units have become the most coveted resources in the artificial intelligence sector, driving the company’s stock to new highs and pushing its valuation close to $6 trillion. With anticipated revenue of $108 billion for the upcoming October quarter—a 89% increase year-over-year—demand for computing power has surged. However, this expansion has created a complex procurement landscape for businesses seeking immediate access to these chips.
While cloud infrastructure giants like Amazon, Microsoft, and Google remain primary vendors, the ecosystem is rapidly diversifying. According to SemiAnalysis, the number of Nvidia GPU providers grew to 323 by September, up from 209 less than a year earlier. In July, five clients accounted for at least 10% of Nvidia’s accounts receivable, a rise from just three in January. This shift highlights how companies are navigating a “paradox of choice” to secure the infrastructure they need urgently.
Jeff Hawkins, Nvidia’s chief executive, noted at a recent Goldman Sachs conference that a new wave of neoclouds is emerging with hundreds of billions of dollars in backlog. He stated, “You’re going to see a whole new crop of really, really exciting neoclouds.”
Hawkins mentioning billions in backlog is interesting. It sounds like demand will stay insane even if supply catches up eventually.
Has anyone actually tried the neocloud route? My team heard rumors of massive price gouging from these new providers.
I work at a mid-size AI firm. The ‘paradox of choice’ feels more like a nightmare when you can’t get any chips at all.
Wait, are we sure $6 trillion is sustainable? That revenue projection seems incredibly optimistic for a single quarter.
The rise in GPU providers is wild. It really shows how desperate enterprises are becoming for compute access.