Eight days ago, Oracle and Amazon signed a long-term strategic collaboration agreement. Oracle AI Database@AWS is now live across 22 AWS regions. Oracle Exadata hardware sits physically inside Amazon’s data centers. The two companies just became each other’s most important enterprise partners, and one of them is priced like it is losing.
Oracle (ORCL) closed Wednesday at roughly $142, against a 52-week high of $345.72. The stock has lost more than half its value since September 2025 despite posting its strongest organic revenue quarter in 15 years just three months ago. The disconnect between what the business is doing and what the stock is pricing is the investment question.
Why This Stock Now
Two facts landed in the same week that change the risk calculus. First, Oracle and Amazon Web Services signed an expanded, long-term strategic collaboration agreement focused on accelerating customer migration to Oracle AI Database@AWS. Second, Amazon Web Services grew 37% year over year in Q2 2026, its fastest rate in 18 quarters, and Amazon’s total revenue crossed $200 billion for the first time in a single quarter. The two data points belong together. A healthier AWS means more enterprise workloads on Amazon infrastructure, and those workloads increasingly run on Oracle’s database layer. Oracle’s revenue follows Amazon’s traffic growth now, not competes with it.
The stock market has not priced that in.
The Business
Oracle Cloud Infrastructure, known as OCI, was built as a second-generation cloud. While its global market share sits at roughly 2% to 3%, Oracle has purposefully targeted a specific, highly lucrative demographic: large enterprises with complex, mission-critical legacy applications. That focus is now paying off in a way that the headline market-share number obscures.
Total quarterly revenues increased 21% to $19.2 billion in the most recent quarter, reflecting broad-based demand for Oracle’s cloud technology and applications suites. Cloud revenues increased 47% to $9.9 billion, driven by 93% growth in Cloud Infrastructure. IaaS growing 93% means Oracle’s pure compute business, the piece that competes most directly with AWS, is the fastest-growing segment in the company by a wide margin.
The database franchise is what makes this defensible. Oracle and AWS are now running a collaboration that has Oracle’s database infrastructure inside Amazon’s data centers, with high-performance networking between OCI and AWS designed to reduce the friction enterprises face when running database and AI workloads across both platforms. That is not a temporary partnership. It is an architectural decision by enterprise customers who refuse to migrate their Oracle databases to native AWS equivalents, and who now don’t have to choose.
Why Wall Street Is Paying Attention
The backlog is the most debated number in enterprise tech right now. Oracle’s remaining performance obligation reached $638 billion as of May 31, up 363% year over year. The OpenAI contract, reported by Reuters as amounting to $300 billion of computing power over five years, is by far the largest individual cloud deal ever reported. Oracle has also named Meta, xAI, Nvidia, and others as major AI customers tied to its infrastructure demand.
The recognition schedule is what matters for near-term revenue. Approximately 12% of the $638 billion, roughly $76.6 billion, is expected to be recognized within the next 12 months, and approximately 34%, or $216.9 billion, in months 13 through 36. Oracle’s own FY2027 revenue guidance is $90 billion. The contracted backlog converting over the next 12 months alone is almost equal to that full-year target. The guidance looks conservative against the backlog math.
The August 13 AWS deal adds a second revenue engine. The expansion builds on momentum for Oracle AI Database@AWS, giving enterprises a lower-cost path to migrate Oracle databases and integrate them with AWS analytics and AI services. Enterprises running regulated workloads in financial services, transportation, and healthcare now have a path to cloud that does not require abandoning 20 years of Oracle investment. That is a migration unlock, not a margin dilutant.
What’s Driving the Opportunity
Oracle is being priced as a capital-destruction story. The market is not wrong to flag the cost. For fiscal 2026, Oracle reported $23.7 billion in negative free cash flow, with depreciation at $7.623 billion. Capital expenditures jumped 162% to $55.663 billion. That is an enormous cash outflow for a company carrying significant debt.
But the AWS deal changes the financing math at the margin. Oracle has said a meaningful portion of the RPO increase came from large-scale AI contracts where the customer prepaid Oracle for GPU purchases, or where the customer bought and supplied the GPUs themselves. Those two mechanisms lower the sum Oracle will need for data center construction. The backlog is not entirely Oracle’s capital at risk. A material portion of the buildout is being funded by the customers themselves.
Meanwhile, the AWS collaboration pulls Oracle’s database revenue into Amazon’s ecosystem rather than forcing Oracle to build competing distribution. Oracle has said it delivered more than 1.2 gigawatts of capacity to customers in FY2026 and expects Q1 FY2027 delivery to approach 1 gigawatt, nearly matching the prior four quarters combined. The capacity ramp is real. The AWS deal expands the addressable customer base for that capacity without requiring Oracle to win new cloud regions from scratch.
FY2027 guidance, if achieved, represents a structural step change. Oracle projects a $90 billion revenue target. For the current quarter, Oracle called for $1.72 to $1.76 in adjusted earnings per share, with 27% to 29% revenue growth. Analysts had expected $1.68.
What Could Go Wrong
The bear case on Oracle is not irrational, and investors who have held it down 59% from the high are not simply wrong on the facts.
The debt is the primary risk. Moody’s has warned that Oracle’s debt would grow faster than earnings, pushing leverage toward 4x EBITDA due to the capital spending required to fulfill large AI infrastructure commitments. A debt load exceeding $100 billion and negative free cash flow create significant financial fragility compared to cash-rich competitors. Amazon has indicated it plans to spend about $220 billion in 2026 capex from a position of operating strength. Oracle is doing it from a position of leverage.
Concentration is the second risk. OpenAI is widely reported as the largest single driver of Oracle’s recent AI-contract surge, and the primary unanswered question is whether Oracle’s customers can reasonably fulfill their contracted obligations, especially OpenAI, which is still a private company. If OpenAI’s capital access tightens or its model strategy shifts toward on-premise compute, a portion of that backlog becomes renegotiable.
Execution on capacity delivery is the third. Oracle itself has highlighted that capacity buildouts depend on securing large volumes of GPUs, and the strategy is vulnerable to industry-wide power and supply chain bottlenecks. Every quarter that Oracle runs behind on megawatt delivery is a quarter where customers gain optionality to redirect spending toward AWS, Azure, or other infrastructure providers.
The Bottom Line
Oracle and Amazon are not competing for the same enterprise customers anymore. They are sharing them. The August 13 deal makes Oracle’s database layer a dependency inside AWS, not a challenger to it. That is a durable position, and it changes the revenue risk profile in ways the stock price does not yet reflect.
In the fiscal year ending May 31, 2026, Oracle posted $67.357 billion in revenue, up 17%. The company then guided for 27% to 29% growth in the current quarter. AWS just posted 37% cloud growth. Both businesses are accelerating. One is priced at all-time highs and one is down 59% from its peak.
The market is pricing ORCL as though the backlog will not convert, the debt will overwhelm cash flow, and OpenAI will cancel. All three are risks worth monitoring. But a stock that has lost more than half its value while the underlying business posted its strongest organic growth quarter in 15 years, signed one of the largest cloud deals ever reported, and just formalized a deeper partnership with the world’s dominant cloud provider is not a crowded trade.
The catalyst that closes the gap is Q1 FY2027 earnings, due in September. If Oracle delivers on its 27% to 29% revenue guide while showing that GPU utilization, reported near 97.5%, is holding, the free cash flow trajectory becomes the argument, not the debt. The gigawatt delivery figure is the metric to watch. That is what converts backlog skeptics into believers.