Source: Nvidia is scaling back DGX Cloud to primarily internal R&D use; DGX Cloud was initially envisioned to compete with major cloud providers like AWS
Nvidia is stepping back from its nascent cloud computing business, which had put it in quasi-competition with Amazon Web Services.
Context & Ripple Effects
DGX Cloud began as Nvidia’s hosted route for companies scaling AI workloads, but its cloud-provider positioning faced an early constraint when AWS declined to use DGX Cloud chips while considering AMD’s MI300. That left Nvidia balancing a direct-service ambition against its role as a supplier to major clouds.
The reported retrenchment is consistent with the later move to fold DGX Cloud into Nvidia engineering, turning the service’s infrastructure and expertise toward product development rather than enterprise cloud sales.
First-order effects
- Nvidia reduces DGX Cloud’s role as an enterprise-facing cloud service and redirects it primarily toward internal R&D.
- AWS faces less direct competition from Nvidia’s own hosted offering, while prospective DGX Cloud customers have fewer reasons to treat Nvidia as a standalone cloud alternative.
Second-order effects
- Nvidia’s cloud partners and hyperscalers gain relative importance as routes for customers to access Nvidia infrastructure, concentrating customer relationships outside DGX Cloud.
- The pullback reinforces the incentive for cloud providers to differentiate their AI offerings through their own chips, software, and managed services rather than assume Nvidia will remain only a component supplier.
Third-order effects
- If replicated across the sector, AI chip vendors may find it harder to operate neutral cloud services without conflicting with the large providers that buy and distribute their hardware.
- The episode points to an AI infrastructure market in which control of compute supply and customer access is split among chipmakers, hyperscalers, and specialized operators rather than consolidated in one vendor.
The trend: AI infrastructure is shifting toward selective vertical integration, with chipmakers testing services but preserving the cloud-provider channels that scale their platforms.