Oracle Stock Rockets 40% on $455B AI Backlog, Margin Doubts

💡 TL;DR - The 30 Seconds Version

🚀 Oracle stock surged 40% Wednesday after revealing $455 billion in AI contract backlog, up 359% year-over-year from major customers including OpenAI, xAI, and Meta.

📊 The company projects cloud infrastructure revenue will explode from $18 billion this year to $144 billion by fiscal 2030, backed by signed contracts already in the books.

💰 The rally briefly made Oracle co-founder Larry Ellison the world's richest person at $393 billion, surpassing Elon Musk's $385 billion fortune.

🏭 Oracle is ramping capital spending to $35 billion in 2026 to build data centers and secure GPU capacity for AI training and inference workloads.

⚠️ Analysts warn that AI compute rental business operates at single-digit margins or potential losses, marking a dramatic shift from Oracle's 50% legacy software margins.

🌍 The surge signals investor belief in AI infrastructure as the next profit center, but Oracle must prove it can convert massive bookings into sustainable cash flow.

Stock rockets as Safra Catz sketches a $144B OCI by 2030; margins and capex test the story.

Oracle missed on revenue and EPS. Its stock still exploded, powered by a new five-year cloud outlook and a booked backlog that dwarfs the company’s past. In Oracle’s fiscal Q1 results and five-year cloud outlook, CEO Safra Catz said Remaining Performance Obligations (RPO) hit $455 billion—up 359% year over year—and previewed a path to $144 billion in annual cloud-infrastructure revenue by fiscal 2030.

What changed, exactly

Two forces converged. First, Oracle signed four multibillion-dollar deals with three customers in the quarter, and says more are imminent. Second, management translated that surge into a concrete revenue ramp, asserting that OCI will grow 77% to $18 billion this year and then step to $32 billion, $73 billion, $114 billion, and $144 billion in the four fiscal years after. That’s the headline.

The company also expects RPO to top $500 billion “in the next few months,” signaling long-duration commitments for compute and storage tied to AI training and inference. That’s the thesis.

The numbers that moved the market

The quarter itself was mixed: revenue rose 12% to $14.93 billion and adjusted EPS came in a penny light. Cloud revenue rose 28% to $7.2 billion, while software revenue slipped 1% to $5.7 billion. None of that justified a 40% move by itself.

Guidance did. Catz’s five-year OCI glidepath reframed Oracle as a capacity broker to AI’s big spenders—OpenAI, xAI, Meta, and others—backed by a backlog that management says now captures the bulk of that plan. Oracle also says multi-cloud deals with Amazon, Microsoft, and Google are compounding, with dozens more data centers slated for those partners, according to remarks summarized by wire services. Momentum is visible.

Why investors cheered this hard

Because the claim is simple and audacious: booked demand today underwrites growth tomorrow. Analysts called out the scale and composition of RPO—citing contributions from the Stargate AI-infrastructure venture (SoftBank/OpenAI) and other hyperscaler spillover—as a key reason to treat the five-year plan as more than marketing. If correct, Oracle has engineered a rare second act: from database king to AI compute landlord.

There was also an unmistakable signaling effect. A 359% jump in RPO tells boards and CFOs that GPU scarcity is now a contractual problem solved with capacity reservations, not spot pricing. That favors vendors willing to pre-buy chips, stand up power-dense sites, and sign long-term offtake. Oracle is doing exactly that.

The catch: margins and cash

The economics remain the rub. Several analysts warned that AI compute sold as VM or GPU “rental” can run at single-digit operating margins, even losses, depending on utilization and power costs. Oracle is leaning in anyway, lifting fiscal-2026 capex guidance to roughly $35 billion to build data centers and network fabric at speed. That is a dramatic pivot from the company’s 50%-margin legacy software profile.

Said differently: the revenue trajectory looks steep; the profit slope is the open question. That tension won’t resolve in a quarter.

Competitive positioning

Oracle’s pitch rests on three planks. First, it offers bare-metal GPU instances and dense local storage configurations that AI builders value. Second, it has become a friendly “safety valve” for rivals’ excess demand—multi-cloud as business model, not just architecture. Third, it pairs capacity with tight executive relationships in AI’s inner circle. None is a magic wand. Together, they explain the bookings spike.

Still, the hyperscalers control distribution, developer mindshare, and massive in-house capex. If they catch up on supply or pull demand back on-platform, Oracle’s role as overflow provider could compress. Price matters.

The market signal beyond Oracle

The rally’s collateral story is capital. A 40% one-day surge in a near-trillion-dollar company is a referendum on the AI capex supercycle. It also crowned Larry Ellison the world’s richest person by mid-day Wednesday, according to multiple outlets—an anecdote that underscores how much investor belief has swung toward AI infrastructure as the “picks and shovels” trade. Sentiment is now firmly risk-on for anything tied to GPUs, power, and data centers.

But sentiment won’t pay electricity bills. Earnings quality and free-cash-flow conversion will be scrutinized as contracts start to recognize into revenue, especially if power constraints, chip supply, or unit economics wobble. Patience will be required. So will precision.

Bottom line

Oracle just gave Wall Street something it rarely gets in AI: a dated, quantified, multi-year revenue map backed by signed paper. The company now has to turn capex into durable margins, not just breathtaking bookings. That is harder than ringing the bell at the NYSE.

Why this matters

AI’s infrastructure boom is spilling beyond the Big Three clouds, with long-term contracts rewiring who supplies—and profits from—GPU capacity.
The next test isn’t demand; it’s profitability under power, chip, and utilization constraints as capex surges.

❓ Frequently Asked Questions

Q: What exactly is "Remaining Performance Obligations" and why does it matter?

A: RPO represents the total value of signed contracts that Oracle will deliver in the future. At $455 billion, it's like a guaranteed revenue pipeline—money already committed by customers but not yet earned. This backlog jumped 359% in one year, signaling massive pre-committed AI spending.

Q: How is Oracle's cloud business different from Amazon or Microsoft's?

A: Oracle focuses on "bare-metal" GPU instances and dense storage for AI training, while also serving as overflow capacity for other clouds. When AWS or Azure run out of GPUs, customers can use Oracle's infrastructure. It's positioning itself as the AI industry's "safety valve."

Q: Why are AI compute margins so much lower than Oracle's traditional software business?

A: Oracle's legacy database software has 50% operating margins because it's mostly code with minimal ongoing costs. AI compute requires expensive GPUs, massive electricity bills, and data center construction. Analysts estimate Oracle may operate at single-digit margins or even losses on GPU rentals.

Q: What's the Stargate venture that's contributing to Oracle's backlog?

A: Stargate is a $500 billion joint venture between SoftBank, OpenAI, and Oracle to build massive data center capacity across the US. It's designed to support OpenAI's future AI models and represents one of the largest infrastructure investments in tech history.

Q: How realistic is Oracle's projection to hit $144 billion in cloud revenue by 2030?

A: Oracle's current cloud infrastructure revenue is around $10 billion, so reaching $144 billion requires 14x growth in six years. The company claims most of this growth is already "booked" in signed contracts, but executing the buildout and maintaining customer relationships at that scale remains unproven.