Oracle's Q1 FY2027: OCI Revenue More Than Doubles, Backlog Hits $664 Billion, Guidance Raised — Then the Market Changed Its Mind 오라클 2027 회계연도 1분기: OCI 매출 2배 이상, 수주잔고 $664B 기록, 가이던스 상향 — 그런데 시장은 마음을 바꿨다

The Print: A Clean Beat, a Bigger Backlog

On September 10, after market close, Oracle reported fiscal Q1 2027 results (quarter ended August 31, 2026): total revenue of roughly $19.35 billion, up about 30% year-over-year – Oracle’s own investor-relations release titled the quarter “Driven by Triple-Digit Growth in Cloud Infrastructure Revenues.” Oracle Cloud Infrastructure (OCI) revenue more than doubled, up 121% year-over-year, according to multiple outlets (TradingKey, finance.biggo.com, MEXC) citing the earnings call. Remaining Performance Obligations (RPO) – Oracle’s forward-looking contracted backlog, and the metric this blog has watched across Oracle’s last several quarters as evidence of AI-demand durability – hit a record $664 billion, up from $638 billion the prior quarter (itself a record at the time). Oracle also disclosed it delivered 850 megawatts of AI data-center compute capacity in the quarter alone, per finance.biggo.com’s earnings-call coverage. Capital expenditure for the quarter came in at $28.5 billion. Multiple outlets (Yahoo Finance, 24/7 Wall St., TradingKey) reported Oracle raised its full-year fiscal 2027 revenue and profit guidance following the print, with TradingKey specifically flagging “record RPO, full-year cloud guidance raised substantially” as the reason shares moved.

Why This Belongs in the AI CapEx Conversation

Oracle is one of the five companies this blog’s AI CapEx coverage explicitly tracks (alongside Microsoft, Alphabet, Amazon, and Meta), and it sits in a different seat than the other four: rather than spending its own balance sheet on GPUs, Oracle is the cloud landlord renting out AI compute capacity to tenants like OpenAI – making its backlog a read on how much AI-infrastructure demand from others is actually getting contracted and delivered, not just announced. A $664B RPO figure, growing from $553B (Q3 FY2026, reported March) to $638B (Q4 FY2026, reported June) to $664B now, is a steadily compounding forward order book rather than a one-quarter spike. The 850 MW delivered this quarter is a direct, physical-capacity data point on how fast that backlog is converting into usable data centers – the conversion question this blog flagged as the open risk in Broadcom’s September 7 post and Qualcomm’s September 9 post.

The Pattern This Blog Has Tracked Since July — And Where Oracle Fits

Oracle’s case is the most mixed entry in this table. Thursday’s after-hours and Friday premarket reaction was unambiguously positive – TradingKey and stockstotrade.com both described a sharp premarket jump driven specifically by the backlog record and guidance raise. But multiple outlets (FXLeaders, Onlinekhabar) reported that by Friday’s session, Oracle’s shares had given back much of that move as part of a broader fourth-consecutive-day market selloff tied to oil prices and rate-hike bets – not anything specific to Oracle’s numbers. That’s a different failure mode than Broadcom’s (where the stock’s own guidance disappointed) or Alphabet/Meta’s (where the capex raise itself was the problem): here the fundamentals landed cleanly positive, and a macro move overran it anyway.

Investment Read

Oracle’s quarter is the cleanest “fundamentals beat expectations” data point this blog has seen yet in the AI capex thread – a backlog that keeps compounding quarter over quarter, OCI revenue more than doubling, and management raising guidance rather than defending a miss. That the stock still couldn’t hold its initial pop past one trading session says less about Oracle’s AI demand (which every number here says is real and growing) and more about how little room the broader market currently has to treat good AI-infrastructure news as good news in isolation, with rate and inflation concerns pulling in the opposite direction the same week. Watch two things next: whether Oracle’s 850 MW-per-quarter delivery pace holds or accelerates (the backlog-to-delivered-capacity conversion rate is the real constraint, not contract signings), and whether the next round of hyperscaler earnings in October-November treats a beat-and-raise the way this market treated Amazon’s (rewarded) or the way it treated Oracle’s second trading day (shrugged off by macro).

Disclaimer: This post is for informational purposes only and does not constitute investment advice. All investment decisions are the reader’s sole responsibility.

Sources: Oracle Corporation fiscal Q1 2027 earnings release and call (2026-09-10); CNBC “Oracle’s stock edges up on earnings beat as cloud infrastructure revenue more than doubles” (2026-09-10); Reuters “Oracle tops estimates as AI demand tempers cash-burn fears” (2026-09-10); Yahoo Finance “Oracle FQ1 Revenue Reaches $19.35 Billion as FY2027 Profit Guidance Rises” (2026-09-11); finance.biggo.com “Oracle Delivers First-Ever Sequential Q1 Revenue Jump to $19.3B as OCI Surges 121%” and “Oracle Delivers 850 Megawatts of AI Compute in a Single Quarter, Raises Full-Year Revenue Guidance to $90 Billion” (2026-09-10/11); TradingKey “Oracle Q1 FY2027 Earnings: Cloud Infrastructure Revenue More Than Doubles” and “Why Oracle Shares Surged 7% Pre-Market? Record RPO, Full-Year Cloud Guidance Raised Substantially” (2026-09-11); 24/7 Wall St. “Oracle Weakens Bear Case With Broader AI Customer Base and $664 Billion Backlog” (2026-09-11); MEXC “Oracle Earnings Analysis: AI Demand Drives 121% Infrastructure Growth as Backlog Hits $664 Billion” (2026-09-11); FXLeaders “ORCL Sinks to $150s levels as Oracle’s AI Spending Strains Balance Sheet” (2026-09-11); this blog’s July 29, July 30, September 7, and September 9 posts on hyperscaler and custom-silicon AI capex.