Oracle expands Nvidia Vera Rubin deal to multi-gigawatt scale and adds Google TPUs, Cerebras and SambaNova chips
Oracle said it will deploy Nvidia's newest AI servers at multi-gigawatt scale while also offering Google's tensor chips for the first time, plus inference chips from Cerebras and SambaNova.
Oracle (ORCL) announced Tuesday that it is widening the AI hardware in its cloud: more Nvidia (NVDA) systems, and for the first time chips from several Nvidia rivals.
| Item | Detail |
|---|---|
| Nvidia | Vera Rubin servers at “multi-gigawatt” scale; deal value not disclosed |
| Google (GOOGL) | Tensor processing units offered in Oracle’s cloud for the first time |
| Inference chips | Cerebras and Intel-backed SambaNova |
| Networking | Some Nvidia-based systems use Oracle’s in-house Acceleron chips |
| FY2027 capex target | about $90-95B, unchanged |
Why it matters: it is a large new outside customer for Google’s TPUs and the AI chip startups, and a sign that big cloud buyers want options beyond Nvidia for inference, the part of AI work that is easier to move between chip types. Nvidia still anchors the training side of the deal. Oracle did not say how the purchases will be financed.
Market reaction: chip stocks led the Nasdaq up about 1% on the day; no company-specific move was attributed to the announcement in the reports reviewed.
Next: Oracle’s investor day later this week, and Alphabet’s Q3 results on Wednesday, Oct. 28.
Sources
All content on this site is for informational purposes only and is not investment advice. You are solely responsible for your investment decisions. Past performance does not guarantee future results. Editorial & Corrections Policy