Oracle Loads Up on AI Infrastructure as OCI Backlog, Data Center Commitments Surge

Key Takeaways

Oracle's Q2 FY2026 cloud revenue reached $8 billion, growing 33% YoY, with OCI revenue up 66% to $4.1 billion and GPU-related revenue skyrocketing 177%. The cloud now constitutes approximately half of Oracle's total quarterly revenue.

The company disclosed significant financial commitments, including $248 billion in long-term data center leases, indicating a substantial investment in AI infrastructure and cloud capacity, which raises questions about debt management and profitability over the coming years.

Oracle is strategically embedding its AI capabilities and multi-cloud database offerings into all major cloud platforms, positioning itself as a competitive hybrid solution in the ERP market, and enabling advanced data unification and reasoning across applications.

Oracle’s Q2 FY2026 numbers, released on December 10, showed how quickly its cloud infrastructure and AI bets are scaling. Remaining performance obligations reached $523.3 billion, up some 433% year over year (YoY) and $68 billion since August, driven by contracts with Meta, Nvidia, and other hyperscale buyers.

Total cloud revenue, including applications and infrastructure, hit $8 billion, up 33%, with cloud now accounting for around half of Oracle’s overall $16.1 billion in quarterly revenue, itself up 13%. Oracle Cloud Infrastructure (OCI) revenue grew 66% to $4.1 billion, while GPU-related revenue jumped 177% and multi-cloud database consumption surged 817% YoY.​

OCI runs various customer-facing regions, with more planned, and delivered close to 400 megawatts of data center capacity and 50% more GPU capacity than in Q1, TechRadar Pro December 15 reports. Oracle expects Q3 total cloud revenue to grow 37% to 41% in constant currency, with total revenue up 16% to 18%. The company now estimates $4 billion of additional revenue in FY2027 on the back of the newly added backlog.​

Data Center Leases, Rising Debt

In an SEC filing, Oracle disclosed $248 billion in additional lease commitments, “substantially all related to data centers and cloud capacity arrangements,” per the outlet. The lease commitments are expected to start between Q3 FY2026 and FY2028, with 15- to 19-year terms that are not yet on the balance sheet. That figure represents a 148% increase versus the August 2025 quarter.

Oracle recently lifted its Capex plan to $50 billion for the fiscal year, up from $35 billion, as it races to build out Project Stargate and other AI facilities.​​

New contracts with Meta and Nvidia, plus a reported $300 billion deal with OpenAI to add up to 4.5 GW of capacity across multiple US sites, have materially reshaped Oracle’s balance sheet. Total debt obligations reportedly now exceed $124 billion, up from $89 billion a year earlier, and free cash flow in Q2 was negative $10 billion on $12 billion of Capex, per the article. Oracle reportedly stresses that most of its Capex is for revenue-generating equipment, while land, buildings, and power are leased, and that the equipment spend late in the build cycle is timed to align cash outlays with contracted revenue.

Multi-Cloud Database, AI Data, Fusion Momentum

Beyond infrastructure, Oracle is leaning hard into database and application integration as a differentiator in a three-step strategy: making Oracle Database available in all major clouds via embedded OCI regions; adding vector capabilities to turn it into an “AI database”; and launching an AI data platform that can “vectorize” and catalog data across Oracle databases, Oracle applications, other databases, object stores, and bespoke systems so large language models can reason across all enterprise data in one shot. This unlocks “all the value in all [the] data,” said Oracle’s Chairman and CTO Lawrence Ellison, enabling multistep reasoning on private enterprise data while keeping it secure and breaking down fragmentation across applications and stores.​

On the SaaS side, cloud applications revenue reached a roughly $6 billion annualized run rate, up 11%, with Fusion ERP up 17%, Fusion SCM up 18%, Fusion HCM up 14%, and NetSuite up 13%. Oracle’s combined “industry cloud” portfolio for sectors such as hospitality, construction, retail, banking, restaurants, local government, and communications grew 21% in the quarter. Oracle has also reorganized its sales force to unify industry and Fusion app selling, aiming for “One Oracle” deals that bundle applications with the AI data platform, and reports growing cross-sell as industry suites pull in Fusion and vice versa.​

What This Means for ERP Insiders

AI infrastructure commitments will increasingly shape ERP platform risk and opportunity. Oracle’s $523 billion remaining performance obligations and $248 billion in long-term data center leases indicate that OCI is being built out as a long-horizon AI capacity utility, not a side business. For ERP buyers and architects, that scale signals both durability of the platform and concentration of risk around Oracle’s ability to manage debt, utilization, and profitability over 15- to 19-year terms.​​

Multi-cloud database and AI data platform strategy raises the bar for data-unification narratives. Oracle’s push to embed its database in AWS, Azure, and Google Cloud; vectorize all data; and expose a unified AI data platform that spans Oracle and non-Oracle sources sets a high benchmark for what “AI-ready ERP data” looks like. ERP leaders evaluating roadmaps will need to consider whether their platforms can support similar cross-cloud, cross-application reasoning patterns without compromising security or governance.​

Fusion and industry cloud growth reinforce Oracle as a full-stack alternative in the ERP race. With Fusion ERP, SCM, and HCM growing double digits and industry clouds up 21%, Oracle is positioning itself as a combined SaaS-plus-infrastructure competitor rather than a pure-play ERP vendor. For system integrators, product teams, and transformation leaders, this integrated stack model may shift buying conversations from single-suite ERP replacement to broader decisions about cloud standardization, AI data platforms, and multi-cloud strategy.​