Oracle targets 10GW of AI compute by FY28 as its backlog hits $664 billion

The database giant plans a roughly tenfold capacity jump and says it can fund most of it without heavy new borrowing Oracle plans to grow its compute capacity from about 1 GW in FY25 to 10 GW by FY28. That is roughly a tenfold expansion in three fiscal years. The timing matters because Oracle is now near the most expensive stretch of the plan. The company expects capital spending to peak over the next two fiscal years. The numbers behind the buildout Oracle laid out the roadmap at its October 2026 investor day. The goal is approximately 10 GW of total AI data center capacity by the end of fiscal 2028. Around 7 GW of that is expected to come from incremental additions over the coming years. The rest reflects capacity Oracle already has in place or under way. Oracle’s remaining performance obligations, or RPO, reached $664 billion by the end of the first quarter of FY27. RPO is contracted revenue that customers have committed to but Oracle has not yet booked. That figure grew 3 to 4-fold year over year, with the company attributing the surge to AI contracts. In Q1 FY27, Oracle brought online 850 MW of capacity, including more than 300,000 GPUs. Those deployments spanned 38 projects a...