"We had a tremendous quarter that exceeded expectations across the board." — Doug Caring (CFO)
"Oracle Corporation has the fastest growing, most complete suite of cloud applications in the market, full stop." — Mike Cecilia (CEO)
"Multicloud database revenue grew 531% year over year. AI infrastructure revenue grew 243% year over year." — Clay Magouyrk (CEO)
"Our numbers speak for themselves. We are overdelivering on FY26 revenue and earnings, and we are constantly raising our FY27 forecast." — Clay Magouyrk (CEO)
"AI-powered end-to-end ecosystem automation platforms are quite unique to Oracle Corporation." — Mike Cecilia (CEO)
Prepared Metrics
指标
数值
来源/说明
有机总收入同比增长（美元）
20%+
15年来首次（CFO）
有机非GAAP EPS同比增长（美元）
20%+
15年来首次（CFO）
云应用收入（固定汇率）同比增长
11%
Mike
云应用年化收入（固定汇率）
161 亿美元
Mike
多雲数据库收入同比增长
531%
Clay
AI基础设施收入同比增长
243%
Clay
剩余履约义务（RPO）
5530 亿美元
Clay
已交付AI容量（Q3）
>400 兆瓦
Clay
AI容量毛利率
32%
高于30%指引（Clay）
新签合同（新商业模式）
>290 亿美元
自上次财报后（Clay）
已筹集融资
300 亿美元
债券+可转换优先股（CFO）
多雲数据库区域（Q3末）
MS 33 / GCP 14 / AWS 8
Clay
Q4末AWS区域目标
22
Clay
Q&A Batch (1-5 of 5)
Q1 — John DiFucci
Topic: Halo effect on non-AI cloud businesses and fiscal 2027 CapEx
Key points:
RPO increase driven by large-scale AI contracts; pipeline for traditional cloud workloads (Dedicated Region, sovereign clouds, Alloy) is “up materially”
Halo effect includes: embedding AI into apps (training models on OCI + private data); OCI as budget creator (faster/cheaper than competitors); Sovereign AI pipeline increasing globally
CapEx for fiscal 2027: Company will discuss after fiscal year-end; additional funding mechanisms (e.g., $50 billion announced this calendar year) allow CapEx without out-of-pocket cash from Oracle Corp; committed to maintaining investment-grade rating
Mgmt stance: Bullish — halo effect described as “pretty big,” with concrete examples of apps, OCI, and sovereign momentum; CapEx approach flexible but disciplined
Q2 — Mark Murphy
Topic: Data center location strategy for AI inferencing
Key points:
Inferencing demand growing rapidly due to higher model utilization and new use cases (e.g., Claude, Codex)
Latency is proportional: for most business queries (seconds-long), 40ms extra latency from Wyoming to New York is negligible; latency problem is hardware architecture, not location
Data centers can be placed where power is abundant and land is plentiful; innovation in AI accelerators (GROQ, Cerebras, Positron, NVIDIA GTC) will reduce latency
Mgmt stance: Neutral/confident — data center location is a “very tiny part” of latency; flexibility to optimize for power and land availability
Q3 — Siti Panigrahi
Topic: AI Database and AI Data Platform growth; private LLM adoption
Key points:
Customers are not building private LLMs; instead, they combine best models with private data (retrieval-augmented generation)
Oracle AI Database added MPT servers and natural language SQL; AI Data Platform solves private data + model integration
Multicloud database growth (mentioned in prepared remarks) driven by customers moving private data to cloud to use AI
Analyst Day (October) inflection for AI Database growth: confidence implied by “a lot of momentum” across the stack
Mgmt stance: Bullish — strong adoption of private data + AI models; multicloud database accelerating as prerequisite for AI
Q4 — Mark Moerdler
Topic: AI data center profitability, capital costs, and sovereign cloud opportunity
Key points:
AI data center gross margin on accelerators: 30%–40% (guidance unchanged); costs improving (networking, hardware, power)
Adjacent services (compute, storage, security) account for 10%–20% of spend, with higher margins; multicloud database margins 60%–80%
Profitability limitation: not capacity, but under-construction expenses (multiple data halls being built); capacity already contracted at profitable rates
Sovereign cloud: Alloy model delivers full-stack OCI (all services, apps, AI Data Platform) in sovereign zones of any size; customers can define zone across multiple countries (e.g., Europe/Africa); flexibility in contract and location
Mgmt stance: Bullish — “extremely confident” in profitability and capacity; sovereign cloud positioned as “most flexible” and unique (full OCI, not subset)
Q5 — Raimo Lenschow
Topic: Will AI kill SaaS/application software?
Key points:
Customers are not replacing core systems (retail, banking, health records) with cobbled AI features; they want to consume AI out of the box from Oracle apps
Oracle embeds AI (1,000 AI agents live in Fusion; hundreds in banking suite) at no additional cost as part of quarterly upgrades
AI seen as disruptive, but Oracle positions as disruptor by embedding AI into mission-critical apps with decades of industry/regulatory experience
Mgmt stance: Bullish — AI strengthens SaaS position, does not kill it; “thrilled with results” and expects more color forward