Oracle's RPO Backlog Soars 438% to $523 Billion, Fuels Multi-Year AI Growth

Oracle has announced a planned increase in its capital expenditures on cloud infrastructure to $50 billion over the next several years, a 67 percent rise from its previous five-year investment plan. The company is targeting GPU-rich data centers to support surging enterprise AI and cloud demand, with plans to deploy at least 20 hyperscale facilities by 2028. Oracle estimates that these investments will boost cloud revenue growth rates by 30 percent annually once deployments reach full capacity, positioning the company to challenge larger rivals in the high-performance AI compute market.

In its second quarter of fiscal 2026, Oracle reported that remaining performance obligations—its backlog—had ballooned 438 percent year-over-year to $523 billion. Quarterly cloud and license support revenue grew 14 percent to $16.06 billion, with EPS of $2.26 exceeding consensus estimates by 38 percent. The outsized backlog reflects large commitments from strategic customers such as major AI developers, locking in revenue streams that Oracle expects to recognize over the next three to five years.

Oracle is pursuing FedRAMP Moderate authorization for its Primavera Cloud project management suite, aiming to qualify for multi-year U.S. federal contracts. The company has already engaged with three federal agencies for pilot deployments, targeting initial contract awards totaling $1.2 billion over five years. Securing FedRAMP certification would expand Oracle’s addressable market in regulated sectors—estimated at $15 billion annually—and reinforce its position in government cloud services.