Oracle’s AI Engine Roars: OCI Growth Nears 85% as Demand Outpaces Supply

Oracle delivered one of its strongest quarters in years on Tuesday, posting fiscal third-quarter results that exceeded Wall Street expectations and offered fresh evidence that demand for artificial intelligence infrastructure remains exceptionally strong. The company reported revenue of $17.2 billion, comfortably ahead of the roughly $16.9 billion analysts had been expecting. Adjusted earnings per share came in at $1.79, also above consensus estimates of about $1.69 to $1.70. The results landed above the high end of Oracle’s own guidance as well, with revenue growing 22% year over year in U.S. dollars and 18% in constant currency. The report arrives at a critical moment for the stock and the broader AI infrastructure trade, as investors increasingly view Oracle not as a legacy software company but as a capital-intensive cloud infrastructure provider whose future depends on successfully monetizing the massive AI buildout now underway.

Oracle had previously guided fiscal third-quarter revenue growth of 16% to 18% in constant currency, or 19% to 21% in U.S. dollars, alongside non-GAAP EPS between $1.70 and $1.74. The actual results surpassed those targets, with revenue hitting $17.2 billion and non-GAAP earnings reaching $1.79 per share. Management noted that this marked the first time in more than fifteen years that both organic revenue growth and adjusted earnings per share expanded by more than 20% in the same quarter. That performance suggests Oracle’s pivot toward AI-driven cloud infrastructure is beginning to translate into meaningful financial acceleration rather than remaining simply a story about future demand.

The most closely watched segment of the report was Oracle Cloud Infrastructure, which sits at the center of the company’s AI strategy. OCI revenue climbed to $4.9 billion during the quarter, up 84% year over year and well ahead of the $4.1 billion reported in the previous quarter. Investors had been watching for growth in the mid-80% range, with many analysts pointing to roughly 86% as the key benchmark. While the 84% figure came in slightly below that informal target, it still represents an extraordinary rate of expansion for a business of this size. Total cloud revenue, which includes both infrastructure and software services, rose to $8.9 billion, representing growth of 44% year over year and landing at the high end of the company’s guidance range of 40% to 44%.

Several indicators pointed to accelerating AI demand within Oracle’s cloud ecosystem. Multicloud database revenue surged an astonishing 531%, reflecting growing adoption of Oracle databases across multiple hyperscaler environments. Oracle Cloud Database infrastructure revenue increased 35%, while overall cloud infrastructure growth continued to benefit from rising demand for GPU-powered compute capacity used for AI training and inference workloads. Management said demand for AI computing resources continues to grow faster than supply, reinforcing the argument that Oracle is emerging as a critical provider of infrastructure for next-generation AI models.

If OCI growth confirmed the strength of demand, the company’s remaining performance obligations offered perhaps the most dramatic signal of future revenue potential. RPO ended the quarter at $553 billion, up $29 billion sequentially and far above the roughly $470 billion analysts had been expecting. Although the year-over-year increase slowed to 325% from the extraordinary 433% growth reported in the prior quarter, the absolute size of the backlog remains one of the most striking figures in the AI infrastructure ecosystem. The increase was largely driven by large-scale AI contracts, many of which involve arrangements where customers either prepay for equipment or supply their own GPUs. That structure significantly reduces Oracle’s need to raise incremental capital to support the projects, a key concern that has weighed on the stock in recent months.

Financing and capital expenditures have been at the center of investor debate surrounding Oracle’s AI strategy. The company reaffirmed its fiscal 2026 capital expenditure target of approximately $50 billion, providing some relief to investors who had worried the figure might climb even higher. Earlier this year Oracle announced plans to raise up to $50 billion through a mix of debt and equity financing and quickly secured $30 billion through a heavily oversubscribed issuance of investment-grade bonds and mandatory convertible preferred stock. Management also reiterated that it does not expect to issue additional bonds beyond that amount during calendar year 2026, suggesting the company believes its funding requirements are largely addressed for the near term.

The financing plan has been controversial because of the sheer scale of the spending required to build AI data-center capacity. However, Oracle emphasized that many of its largest AI contracts involve structures where customers fund or supply the GPUs required for training clusters. That arrangement shifts some of the capital burden away from Oracle while allowing the company to monetize the infrastructure through cloud services. Investors will likely seek additional clarity on these structures during the conference call, but the company’s comments suggest management believes it can expand capacity without dramatically increasing leverage.

Profitability metrics also held up relatively well despite the massive investment cycle underway. Oracle reported non-GAAP operating income of $7.4 billion, up 19% year over year, while operating margins came in around 32% on a GAAP basis. Although that sits toward the lower end of the company’s historical margin profile, it remains within management’s expected range given the heavy spending required to scale AI infrastructure. Over time Oracle has suggested that AI contracts should ultimately deliver gross margins between 30% and 40%, though investors remain focused on whether the near-term timing mismatch between capital expenditures and revenue recognition will temporarily compress profitability.

Beyond the infrastructure buildout itself, Oracle also highlighted how AI is reshaping its internal operations and software business. The company said new AI code-generation tools are allowing development teams to produce more software with fewer engineers, enabling Oracle to build additional SaaS applications while reducing development costs. That dynamic could ultimately improve the competitiveness and profitability of Oracle’s cloud application suites, which include products like Fusion Cloud ERP and NetSuite. During the quarter, Fusion Cloud ERP revenue rose 17% year over year while NetSuite grew 14%.

Looking ahead, Oracle offered a confident outlook for the current quarter and beyond. For fiscal fourth quarter 2026, the company expects total revenue growth between 19% and 21% in U.S. dollars, while cloud revenue is projected to increase between 46% and 50%. Non-GAAP earnings per share are expected to reach $1.96 to $2.00. Management also raised its fiscal 2027 revenue target to $90 billion, underscoring its belief that the AI infrastructure cycle is still in its early stages.

The bigger question facing investors is whether Oracle can maintain this pace of expansion while managing the financial demands of building out global AI capacity. The company’s enormous backlog suggests demand is real and growing rapidly. What the market still wants to see is consistent evidence that the backlog is converting into revenue and that the resulting revenue will generate attractive long-term returns.

For now, the fiscal third-quarter results appear to move Oracle closer to answering those concerns. Strong cloud growth, accelerating infrastructure demand, and a massive pipeline of AI contracts all point to a company that has successfully repositioned itself at the center of one of the most important technology shifts of the decade. Whether that transformation ultimately delivers sustainable profitability will determine whether Oracle’s AI infrastructure gamble becomes one of the defining success stories of the current technology cycle—or a cautionary tale about the risks of building too much, too quickly.