Oracle announced record financial results for both its fourth quarter and fiscal year 2026, showcasing significant growth in its Cloud Infrastructure (IaaS) and Cloud Applications (SaaS) businesses. Total quarterly revenues increased by 21% to $19.2 billion, while total cloud revenues (IaaS + SaaS) surged by 47% to $9.9 billion. This cloud growth was primarily driven by a 93% increase in Cloud Infrastructure revenue to $5.8 billion, reflecting robust demand for AI training and inferencing capabilities. For the full fiscal year 2026, total revenues reached a record $67.4 billion, up 17%, with cloud revenues increasing 39% to $34.0 billion.
Remaining Performance Obligations (RPO), a measure of contracted future revenue, reached an impressive $638 billion at the end of Q4, marking a 363% year-over-year increase and an $85 billion sequential jump from Q3. A significant portion of this RPO increase, totaling $75 billion, comes from large-scale AI contracts where customers either prepaid for GPUs or supplied their own hardware. This strategy substantially reduces the capital Oracle needs to raise for its AI datacenter build-out. Oracle also reported signing $67 billion in AI infrastructure contracts in Q4, primarily with bring-your-own hardware or prepaid terms.
Despite the strong revenue and earnings beat, Oracle's stock experienced a decline of approximately 7-10% in after-hours trading. This dip was attributed to investor concerns over the company's escalating capital expenditure. Oracle spent $55.7 billion on datacenters in FY2026, exceeding its own guidance of $50 billion. The company now expects to spend roughly $70 billion in net capital expenditure in fiscal year 2027, with the actual capital expenditure potentially $20-$25 billion higher due to prepayments for components. Oracle plans to raise another $40 billion in equity and debt in FY2027, adding to its already substantial $117 billion debt in the Bloomberg US high-grade corporate bond index.
Oracle's CEO Clay Magouyrk noted a strong 97.5% utilization rate of GPUs globally and mentioned that the company brought online 1.2GW of datacenter capacity in 2026, with an anticipated 1GW in Q1 FY2027. Four customers each contracted for more than $8 billion during the quarter. While the $638 billion RPO represents contracted revenue, only 12% is expected to convert within 12 months and 34% within three years, raising questions among investors about the sustained demand for AI to justify such significant annual capital spending and increased debt.
