Oracle Boosts Lease Commitments by Nearly 150% to Meet AI Demand

On Wednesday, Oracle announced that it will increase capital expenditures for the current fiscal year to $50 billion, up from an original forecast of $35 billion. The boost is driven by new contracts with major AI players such as Meta and Nvidia.

Oracle is also accelerating its lease strategy. As of Nov. 30, the company reported $248 billion in lease commitments for data‑center space and cloud capacity that run for 15 to 19 years—an increase of 148 % since the end of August. Those commitments include $10 billion in cloud‑capacity arrangements secured during the quarter.

Over the past decade, Oracle has diversified its business model from enterprise software to cloud infrastructure, positioning itself against Amazon, Microsoft and Google in the race to rent out servers, storage and Nvidia GPUs for AI workloads. OpenAI has become a marquee Oracle cloud customer, announcing a multiyear commitment valued at more than $300 billion in September.

Microsoft, which once served as OpenAI’s exclusive cloud provider, has also been expanding its lease portfolio to supplement its internal data‑center footprint, striking deals with emerging “neocloud” providers such as CoreWeave and Lambda.

Oracle collaborated with startup Crusoe to launch the first phase of the Stargate data‑center in Abilene, Texas, reinforcing its push into high‑performance AI infrastructure.

Investors are scrutinizing how Oracle will finance these massive AI‑data‑center projects. RBC analyst Rishi Jaluria, who maintains a neutral stance on Oracle, asked during the earnings call how the company plans to fund the rapid expansion.

In September, Oracle raised $18 billion in new debt. By the end of November, total debt—including operating‑lease liabilities—had grown to more than $124 billion, up from roughly $89 billion a year earlier.

“We have a range of financing options across public bonds, banks and private‑debt markets,” said Doug Kehring, Oracle’s chief financial officer, during the conference call.

Magouyrk noted that allowing customers to bring their own hardware—particularly custom AI accelerators—could help offset some of the company’s capital outlays.

Despite strong demand for AI infrastructure, Oracle reported revenue that fell short of analysts’ expectations in its latest earnings release. The share price closed down nearly 11 % on Thursday.

What This Means for the Cloud Market

The escalation in Oracle’s capital spending underscores a broader industry shift: hyperscale providers are racing to secure “AI‑first” infrastructure to meet exploding demand for generative‑AI services. By leveraging long‑term lease agreements, Oracle can lock in real‑estate at predictable costs while scaling capacity in line with customer demand.

However, the financing strategy carries risk. The company’s debt load is now approaching the upper end of what credit analysts typically deem sustainable for a technology firm of Oracle’s size. If AI spending slows or pricing pressure intensifies, Oracle could face margin compression.

From a technology standpoint, Oracle’s partnership with Crusoe signals an embrace of energy‑efficient, low‑latency architectures—critical for large‑scale model inference. The Stargate facilities are being engineered to support next‑generation GPUs, custom ASICs and emerging optical‑interconnect technologies, giving Oracle a potential edge in offering differentiated performance‑per‑dollar metrics.

Overall, Oracle’s aggressive push into AI‑focused data centers could reshape competitive dynamics, offering enterprises an alternative to the dominant Amazon, Microsoft and Google clouds. The success of this strategy will hinge on Oracle’s ability to balance rapid capacity growth with disciplined capital management.