Nvidia may be set to aggressively expand into the cloud, and is in talks to lease space from a data center operator.

The Information reports that Nvidia has held discussions with at least one data center owner about leasing its own space for its DGX Cloud service.

nvidia-dgx-superpod.jpg
– Nvidia

With Nvidia's GPUs at the heart of the current generative AI boom, the company has unprecedented power over the market. The company reportedly has already provided preferential access to its supply limited H100s to smaller cloud providers, and then invested in them.

With the hyperscalers, the company has tried to use its position to convince them to adopt DGX Cloud - which would essentially see the companies lease Nvidia's servers, deploying them as a cloud within their cloud that Nvidia can market and sell to enterprises looking for large GPU supercomputers.

While Google, Microsoft, and Oracle agreed to the proposal, The Information previously reported that Amazon Web Services - the world's largest cloud provider - did not. The hyperscalers that did deploy DGX Cloud have not promoted the service on their own websites.

Now, Nvidia is considering cutting out the cloud providers completely, and becoming a wholesale data center customer. Discussions are believed to be still at an early stage, and the company has not committed to the expensive task of becoming a hyperscaler of its own.

At the same time, Google and AWS have developed their own competing AI chips, while Microsoft has long been rumored to be developing its own hardware.

If Nvidia has approached your company about leasing space, get in touch anonymously here.