Nvidia Repositions DGX Cloud, Avoids Cloud War With Hyperscaler Partners
Nvidia's DGX Cloud quietly pivots from direct cloud competition to internal R&D, protecting its crucial hyperscaler relationships.
September 13, 2025

Nvidia is reportedly recalibrating the strategy for its DGX Cloud service, shifting its focus from directly competing with major cloud providers to prioritizing its own internal research and development needs. This move, first reported by The Information, suggests a significant change in direction for the AI chip giant, which had initially positioned DGX Cloud as a premium, direct-to-enterprise offering for high-performance computing. The service, which provides access to clusters of Nvidia's powerful GPUs, is now said to be primarily utilized by the company's own teams for designing new processors and developing advanced artificial intelligence models.[1][2][3] This strategic pivot is believed to be driven by a combination of lower-than-expected customer demand, a pricing structure that struggled to compete with established cloud vendors, and a desire to avoid antagonizing its largest customers—the very cloud hyperscalers that purchase billions of dollars worth of its chips.[4][5]
Launched in March 2023, DGX Cloud was marketed as a way for businesses in sectors like finance and pharmaceuticals to rent AI supercomputing power directly from the source.[1][4] The service came with a premium price tag, such as a monthly charge of $36,999 for a single H100 instance, a cost that was justifiable during a period of intense GPU shortages.[6] However, as the availability of high-performance chips has improved, major cloud providers like Amazon Web Services have significantly cut prices for their own Nvidia-powered instances, with reductions of as much as 45% for H100 and A100 GPUs.[6] This aggressive pricing from competitors has reportedly made DGX Cloud a less attractive option for many potential customers, thereby limiting its market traction.[5] Underscoring this shift, recent Nvidia financial reports have omitted previous disclosures that linked its substantial cloud spending commitments specifically to DGX Cloud, suggesting a change in its role.[6]
The decision to redirect DGX Cloud resources internally also appears to be a calculated move to manage relationships with its most important partners. Major cloud providers such as AWS, Microsoft Azure, and Google Cloud are collectively responsible for about half of Nvidia's revenue.[5] By offering a competing cloud service, Nvidia risked alienating these critical customers, who might be incentivized to accelerate the development of their own proprietary AI chips to reduce their dependency on Nvidia.[7][5] The reported strategy change mitigates this channel conflict, repositioning Nvidia as an enabler and infrastructure architect for the broader AI ecosystem, rather than a direct competitor in the cloud services market.[8] This allows the company to focus on its core business of selling chips and the underlying software stack, ensuring its technology remains the backbone of AI development across various platforms.[6]
Despite numerous reports detailing this strategic change, Nvidia has publicly denied that it is altering its course for DGX Cloud. Alex Black Bjorlin, the executive in charge of the service, has stated there has been no change in strategy, emphasizing that both internal research teams and external customers require significant computing power.[1][9] While the company maintains this public stance, it has also expanded its cloud initiatives through efforts like DGX Cloud Lepton, a marketplace designed to connect AI developers with a global network of GPU cloud providers.[6][10] This platform strategy allows Nvidia to facilitate access to its technology through partners, effectively directing workloads across the ecosystem without directly owning the customer relationship in the same way a traditional cloud provider would. This approach helps Nvidia maintain its central role in the AI industry, tightening its grip on the flow of GPU workloads worldwide while empowering a wider range of cloud partners.[6] The quiet shift for DGX Cloud, coupled with the promotion of partner-focused marketplaces, suggests a nuanced strategy aimed at maximizing chip sales and industry influence without provoking a direct war with the cloud titans that form the foundation of its customer base.[8]