Oracle’s once-soaring valuation has collapsed even as its artificial-intelligence infrastructure business posts record growth and secures landmark government deals, exposing a widening gap between operational momentum and investor sentiment.
The divergence is striking. Shares have fallen more than 60 percent from their September peak, recently touching a 52-week low near $114.75, while Oracle Cloud Infrastructure revenue jumped 93 percent year over year to $5.8 billion in the latest quarter. The company’s remaining performance obligations stand at $638 billion, up 363 percent from a year earlier. Yet the market has priced the stock at roughly 22 times earnings, well below its five-year average of 34.
This tension raises a central question for technology investors: whether Oracle’s heavy spending on AI capacity represents a durable competitive advantage or an unsustainable leverage bet that could pressure future cash flows.
The Sharp Decline in Oracle Shares
Oracle’s stock trajectory has reversed sharply from its 2025 peak. After trading as one of the market’s strongest AI-related names, the shares have lost nearly two-thirds of their value in ten months. The drop has occurred even as reported earnings continued to rise, compressing the price-to-earnings multiple from 76 at the September high to the current level near 22.
Market participants appear to be discounting future growth rather than celebrating recent results. The sell-off accelerated after Oracle disclosed its fiscal-2026 capital expenditure run rate and the associated negative free-cash-flow figure of $24.5 billion. Investors have shifted their focus from contract wins to the balance-sheet cost of fulfilling those contracts, a re-pricing that has left the equity near multi-year lows despite continued revenue expansion.
AI-Driven Revenue Surge and Massive Backlog
Oracle’s infrastructure business has become one of the fastest-growing segments in the cloud industry. The most recent quarter showed total revenue of $19.2 billion, a 21 percent increase, driven almost entirely by demand for GPU clusters used in training and inference workloads. Customers including OpenAI, Meta Platforms, xAI, and SpaceX have committed to multi-year leases of Oracle’s high-density GPU capacity, some exceeding 131,000 accelerators per cluster.
The backlog figure of $638 billion underscores the scale of contracted demand. A single quarter added $85 billion to remaining performance obligations, illustrating both the velocity of new bookings and the long-term nature of AI infrastructure commitments. Oracle’s proprietary RDMA networking fabric and automated data-center provisioning have allowed it to bring capacity online faster than many rivals, translating into measurable cost and speed advantages for hyperscale AI developers.
Financing the Expansion: Debt and Cash Flow Pressures
The capital intensity required to meet this demand has produced a markedly different financial profile from Oracle’s traditional software operations. The company has accumulated nearly $130 billion in debt against only $43 billion in stockholders’ equity, resulting in a debt-to-equity ratio around 300 percent. Comparable figures for Amazon and Microsoft stand at 50 percent and 30 percent, respectively.
Fiscal-2026 capital expenditures reached $55.7 billion, outstripping operating cash flow and producing the negative free-cash-flow outcome noted above. Management has indicated plans to raise an additional $40 billion to $50 billion, signaling that further leverage or equity issuance may be required. While the resulting capacity positions Oracle as a credible alternative to the largest cloud providers in the AI segment, the carrying cost of that capacity introduces execution and refinancing risks that the market is currently capitalizing at a steep discount.
Government Contract Highlights Steady Enterprise Demand
On July 24, the Department of Defense awarded Oracle a consolidated enterprise software agreement covering the entire U.S. military, Coast Guard, and intelligence community. The initial five-year term is valued at $3.31 billion, with a second five-year option that could bring the total to $6.99 billion. The deal consolidates previously fragmented purchases and is projected to save taxpayers at least $441 million.
Despite the award’s size, the stock fell 4.2 percent the following session. The muted reaction reflects the contract’s modest contribution relative to Oracle’s overall revenue base of $67.4 billion in fiscal 2026; even at the full ceiling, the agreement represents roughly 1 percent of annual sales. More importantly, the market interpreted the win as evidence of steady but not incremental demand, reinforcing the view that Oracle’s valuation reset is driven by infrastructure financing concerns rather than a loss of enterprise traction.
Valuation, Analyst Views, and Competitive Context
After the sell-off, Oracle trades below its historical earnings multiple and well below the premiums commanded by pure-play AI infrastructure names. Wall Street remains largely constructive, with 84 percent of covering analysts maintaining buy ratings. The depressed multiple offers a margin of safety if the company can convert its backlog into sustained free-cash-flow growth once new data-center capacity stabilizes.
Competitively, Oracle’s position rests on its ability to deliver large-scale GPU clusters at lower operating cost than general-purpose clouds. Its software-defined automation and RDMA fabric provide measurable latency and efficiency gains for AI workloads. Whether these technical advantages translate into durable pricing power will determine if the current valuation represents an attractive entry point or an accurate reflection of elevated financial risk.
The coming quarters will test whether Oracle can moderate its capital intensity while still capturing the multi-year AI infrastructure wave it has contracted. Success would narrow the gap between operational performance and equity valuation; continued cash burn would likely keep pressure on the multiple regardless of revenue growth.
