Oracle’s $455B Mic Drop: The Backlog That Blew Past the Bears

Oracle just posted a rare kind of quarter: a miss that doesn’t matter. Revenue of $14.9 billion (up 12% y/y) and non-GAAP EPS of $1.47 both landed a touch below consensus, and yet the stock vaulted more than 30% pre-market and ripped through $300. Why? Because the company’s remaining performance obligations (RPO)—a proxy for contracted backlog—exploded to $455 billion, up 359% year over year and $317 billion sequentially. Management says several additional multi-billion-dollar deals are lining up, pushing RPO toward (and likely beyond) the half-trillion mark. In a market obsessed with AI infrastructure scarcity, Oracle essentially unveiled the industry’s largest AI cloud revenue pipeline—and investors recalibrated the story from “nice cloud transition” to “secular AI utility.”

On the prints versus expectations, the top and bottom line were marginally light: revenue missed by about $100 million versus ~$15.0 billion consensus, and non-GAAP EPS missed by a penny. Mix was the tell. Cloud revenue reached $7.2 billion (+28% y/y), with IaaS at $3.3 billion (+55%) and applications at $3.8 billion (+11%). Gross margins are digesting the capital build, but operating income still grew, with non-GAAP op income at $6.2 billion and operating margin roughly in line. Free cash flow dipped negative as capex scaled, a function of building capacity just-in-time for booked demand. Guidance implies total revenue growth re-accelerating into mid-teens in constant currency in FY26, with non-GAAP EPS up high single to low double digits despite the investment cycle.

Capex and bookings are two sides of the same flywheel. With RPO surging, Oracle lifted FY26 capex to roughly $35 billion, from “$25B+” previously. Management emphasized a deploy-to-revenue cadence: hardware is stood up rapidly and turned over to customers to start metering. That’s key for ROI and for keeping unit economics on track as GPU clusters scale. On the bookings side, the company signed four multi-billion deals with three customers in the quarter and highlighted a “who’s who” AI roster—OpenAI, xAI, Meta, Nvidia, AMD, among others. Multi-cloud database demand is inflecting as enterprises run LLMs directly on Oracle data, with multi-cloud database revenue up more than 1,500% in Q1. The strategic pitch is full-stack: GPU-dense, high-speed networking superclusters; a vectorized Oracle database; and multi-cloud footprint embedded with AWS, Azure, and GCP.

Now to the star of the show—the backlog. At $455 billion, Oracle’s RPO is no longer a comforting cushion; it’s a multi-year operating plan. Management framed a five-year OCI revenue glide path that effectively converts a large portion of today’s RPO into tomorrow’s reported revenue: OCI is expected to grow 77% to $18B in FY26, then to $32B, $73B, $114B, and $144B over the subsequent four years. While those outer-year figures extend beyond FY2028, the implication through FY2028 is profound: a step-function increase in recognized revenue anchored by contracted demand that already exists on the balance of obligations. Sell-side previews and post-mortems converged on the same point: the P&L trajectory has been structurally reset by mega-contracts that lock in capacity utilization for years. In practical terms, this backlog provides unusually high revenue visibility, de-risks a substantial capex ramp, and supports management’s view that operating income growth can accelerate even as the company builds data centers at gigawatt scale.

Price action is doing its part to underscore the shift in narrative. With shares through $300, Oracle’s market cap is now discounting not only rapid OCI growth in FY26 but meaningful scale by FY27-FY28. Multiple expansion is riding shotgun with the backlog; several brokers moved price targets into the $330–$350 range, with some framing valuation around ~30x FY28 EPS estimates near $11–$12. That is a rich multiple for a historically “mature” software name, but investors are paying for duration: a half-trillion in contracted work, a mix tilting toward cloud (IaaS plus SaaS approaching or surpassing 50% of revenue next year), and a technology stack positioned for both training and the even larger inferencing opportunity.

Of course, transforming backlog into revenue isn’t automatic; it’s an execution marathon. The capital intensity is real—capex at $35B in FY26 will suppress free cash flow in the near term, and buybacks are already throttled. Supply constraints remain a variable (demand still “dramatically outstrips supply,” per management), and any hiccups in GPU availability, networking throughput, or data center build timelines could push revenue recognition to the right. There’s also the competitive backdrop: hyperscalers are scaling their own AI capacity, and differentiation—on speed, cost per token, data gravity, and multi-cloud flexibility—must remain evident to sustain pricing and share. Oracle argues its superclusters train faster (therefore cheaper per hour) and its vectorized database plus multi-cloud footprint create a moat; the next several quarters will test that assertion in live workloads.

For FY26–FY28 specifically, the RPO-to-revenue bridge is the thesis. With cloud RPO up nearly 500% and total RPO up 359%, the company has line-of-sight to deliver accelerating OCI consumption and rising applications attach. Management’s five-year OCI targets imply an S-curve that bends sharply up through FY28, while the broader business benefits from mix shift, operating leverage in services, and cross-sell into database and ERP. If execution stays on plan, the rich valuation can be worked down by growth; if not, the market will be quick to remind that backlog quality (cancellability, pacing, and pricing) matters as much as headline size.

Bottom line: Oracle’s quarter reframed the investment case from “cloud transition” to “AI infrastructure utility with contractual visibility.” The $455 billion RPO—and management’s expectation to surpass $500 billion—anchors multi-year revenue growth into and beyond 2028, even as the company shoulders a heavy but purposeful capex load. With shares above $300, the bar is higher, the scrutiny tighter, and the onus on conversion unmistakable. But if Oracle continues to stand up capacity on schedule and monetize its backlog at the speed suggested, the “miss that didn’t matter” will be remembered as the moment the story changed lanes.