News

Oracle to deploy 10GW of AI data center capacity and quadruples rack manufacturing output

Tuesday, March 10, 2026 at 10:33 PM

During an earnings call, Oracle leadership revealed plans to bring over 10GW of AI data center capacity online within the next three years, primarily through partner funding. The company has tripled its manufacturing sites and quadrupled rack output, leading to a 60% reduction in the cycle time from rack delivery to revenue generation. While AI accelerator margins are lower, the overall AI infrastructure gross margin is cited at 30-40% due to adjacent services.

Context

Oracle recently reported a surge in AI infrastructure demand, with CEO Safra Catz stating, "Growth in the AI segment of our Infrastructure business was extraordinary—GPU consumption was up 336% in the quarter." During its fiscal 2025 Q2 and subsequent updates, the company revealed a massive expansion plan to deploy 10GW of AI data center capacity over the next three years. This scale-up is supported by a significant manufacturing pivot, where Oracle has tripled its manufacturing sites and quadrupled its rack output, effectively reducing the "rack-delivery-to-revenue" cycle time by 60% to meet what management describes as demand that currently exceeds supply. Financial results highlight this momentum, with Cloud Infrastructure (IaaS) revenue growing 52% to $2.4 billion and the total remaining performance obligation reaching $97 billion. Beyond raw compute, Oracle is positioning itself as the critical "data plane" for agentic AI through multi-cloud partnerships with Microsoft, Google, and Amazon, while leveraging its Alloy platform to capture the growing global demand for sovereign AI and localized data control.

Related Companies

Oracle
Oracle
ORCL
US