By TradingView
Publication Date: 2026-10-27 19:23:00
Cloud infrastructure and corporate software firm Oracle Corp. (ORCL, Financials) is diversifying its AI chip supply strategy, lowering dependency on one source.
Oracle and Nvidia stated the cloud provider will grow installations of Nvidia’s latest Vera Rubin systems to multi-gigawatt levels.
The firms did not reveal the amount of Oracle’s arrangement, however Nvidia CEO Jensen Huang has previously suggested Vera Rubin systems could earn up to $40 billion in revenue per gigawatt.
At the same time, Oracle is expanding its chip portfolio.
For the first time, the firm will provide Google tensor processing units, and it will also include inference processors from Cerebras and SambaNova, which is funded by Intel.
Even some of the Nvidia-based systems will employ Oracle’s own networking technology, Acceleron, instead of networking processors from Nvidia.
And that important because AI workloads are increasingly demanding varied mixes of training, inference and network gear.
For Oracle, more options might make its cloud platform more attractive to clients, as well as increase its bargaining power with suppliers.
Up next is Oracle’s investor day and management’s aim to fund a fiscal 2027 capital-spending spree of about $90 billion to $95 billion.
