American multinational technology company Oracle, founded by Larry Ellison, is executing the largest layoffs in its history, cutting up to 30,000 jobs or about 19% of its 162,000 global workforce. In India, where it employs around 50,000 people, about 10,000-12,000 roles are being eliminated.
At the same time, the company has raised its restructuring budget by $500 million to $2.1 billion in FY26 (June-May), mainly for severance. Oracle attributes part of the reduction to AI coding tools that allow smaller teams to build software faster. However, the primary driver is financial pressure from its aggressive investment in AI data centres, which has strained liquidity and forced cost-cutting at scale.
Infrastructure bet
Oracle started as an asset-light seller of database software and enterprise applications, generating strong cash flows with minimal capital spending. In the early 2010s, it moved into cloud computing with Oracle Cloud Infrastructure, gradually building data centres for enterprise workloads.
Demand surged after 2022 with generative AI, which requires thousands of GPUs for training and inference. This has since shifted into a capital-intensive AI strategy. In 2025, this intensified. Oracle signed a $300 billion cloud deal with OpenAI and joined the $500 billion Stargate project with SoftBank Group and others. It began building dedicated AI data centres across multiple US states.
The AI bet drove a sharp rerating. Oracle’s stock rose from around $140 in early 2025 to an all-time high of $345.72 on 10 September 2025 (before closing at $328.33), an increase of about 2.5 times. This surge briefly lifted Ellison’s net worth to about $393 billion, making him the world’s richest person at the time.
The rally reversed as concerns grew over the cost and execution risks of AI infrastructure. The stock fell more than 50% to $144.23 by 8 April 2026. Despite the decline, it still trades at roughly 18 times forward earnings, indicating that investors continue to expect strong growth from Oracle’s AI-driven cloud business.
Capex carnival
Oracle’s shift towards AI has driven up its capital expenditure from $6.9 billion in FY24 to $21.2 billion in FY25. For fiscal year 2026, Oracle has guided for $50 billion, revised up from $35 billion. Quarterly spending has also surged from $8.5 billion to $12 billion in Sep-Nov 2025 (Q2 FY26), and further to $18.6 billion in Dec 2025-Feb 2026 (Q3 FY26), well above expectations. Analysts estimate it could exceed $85 billion annually by 2029.
Oracle’s approach differs from that of its traditional rival SAP, which has chosen to partner with cloud providers. Oracle instead built its own end-to-end cloud infrastructure stack to compete directly, increasing both risk and capital intensity.
In absolute terms, Oracle’s capex is smaller than those of Big Tech firms. Amazon, Alphabet, Microsoft, and Meta platforms are expected to spend a combined $630 billion in 2026 alone. These four are funding their expansion through strong cash flows. Oracle generates less than a quarter of the operating cash flow of these giants, making it the only hyperscaler funding its expansion primarily through debt. Its debt-to-equity ratio has risen to around 500%, compared with roughly 50% for Amazon and 30% for Microsoft.
Margin pressure
Oracle’s growth is being powered by AI infrastructure demand, but much of that momentum is still deferred. In Q3 FY26, revenue rose 22% year-on-year to $17.19 billion, ahead of expectations. Oracle Cloud Infrastructure (OCI) remains the key driver, with revenue surging 84% to $4.9 billion, accelerating from 68% in the prior quarter. Total cloud revenue grew 44% to $8.9 billion, partly offset by a continued decline in legacy software, with on-premise revenue down 1% in constant currency.
However, a significant portion of this AI-driven growth sits in the future. Oracle’s contracted revenue (RPO) has touched $553 billion, up more than four-fold year-on-year, but only about 10% is expected to convert within the next 12 months.
Margins also reflect the cost of competing in AI. While operating margins stood at 30.8% end of FY25, many AI deals are priced aggressively, with some as low as around 14% gross margins. Even at scale, management expects about 30-40%, below peers such as Microsoft and Amazon Web Services. Meanwhile, heavy AI capex has pushed down free cash flow to negative $24.7 billion over the past year.
Debt overload
Oracle’s AI push is being funded by a mix of heavy borrowing and long-term commitments that sit outside its balance sheet. To support more than $500 billion in contracted future revenue, the company plans about $50 billion in capex this year. Its reported debt has already risen to roughly $105–$124 billion.
But the bigger issue is off-balance-sheet obligations: around $248 billion in long-term data center leases, taking total commitments close to $380 billion. Much of this comes from how Oracle builds data centers. Instead of owning them, it relies on partners to finance and construct facilities, then signs long-term leases that lock in costs for 15-19 years.
Credit markets are growing cautious. The cost of insuring Oracle’s debt has risen sharply, nearing levels last seen during the financial crisis, and rating agencies have flagged its rising leverage and customer concentration. With a credit rating just above junk, investors are demanding higher returns to hold its debt.
On the positive side, Oracle is shifting some costs to customers. New AI deals often require upfront payments or customer-supplied chips from Nvidia, easing pressure on its balance sheet.
Concentrated bet
Oracle’s AI strategy is increasingly tied to a single customer: OpenAI. The company has committed massive infrastructure, including data centers, long-term leases, and tens of billions in chips from Nvidia, to support OpenAI’s needs. In return, it has secured a roughly $300 billion five-year contract that now makes up the bulk of its future revenue.
The upside is clear. If OpenAI scales as expected, Oracle positions itself as a top-tier AI cloud provider with long-term, high-volume demand locked in. But the risks are just as large.
Oracle is effectively betting its balance sheet on a single, unprofitable customer. OpenAI’s current revenue is far below its commitments, and its ability to pay depends on continued fundraising. At the same time, Oracle’s obligations run much longer. It is signing 15-19 year leases to build capacity for a contract that lasts about five years, creating a mismatch that could leave it with expensive, underused infrastructure.
This concentration risk is already straining Oracle’s finances. Heavy upfront investment, rising debt, and negative cash flow have forced aggressive cost cuts, including large-scale layoffs. In effect, Oracle’s AI pivot is not only driving its growth, but also its financial stress.
www.howindialives.com is a database and search engine for public data
