Mark Moerdler

Direct on both parts: short-lived asset mix and late binding give downside flexibility; component pricing answered via efficiency work and newer contracts reflecting pricing.

Amy pointed to $600B+ RPO as evidence of revenue visibility. She noted short-lived assets correlate with revenue while 1/3 goes to 15-year long-lived assets. Satya added conviction about structural TAM in knowledge work, coding, and security.

Hood said GPU contracts for largest customers are sold for the full useful life; Nadella added software re-writes Moore's Law to keep the fleet efficient over its lifespan.

Hood tied short-lived GPU/CPU spend to contract duration matching; Nadella stressed fleet fungibility and continuous modernization as the bubble hedge.

Nadella framed AI SaaS monetization as an expansion of infra/data/app layers; Hood added per-user, tiered, and consumption models will blend over time.

Nadella argued software is the most malleable resource for customers fighting inflationary/growth pressure, implying Microsoft would play offense and take share.

Hood cited Azure AI strength plus broad-based Copilot outperformance in new and expansion seats, usage, and price per seat.

Amy said CapEx growth slows and revenue growth rises as demand persists, converging over time as in the prior cloud transition, with pace set by adoption.

Clayton clarified the distinction between CapEx and Oracle cash, then walked through supplier-financing, BYO-hardware and customer-prepay funding models plus sovereign demand.

Clayton emphasized customer focus, high utilization and renewal rates, and that demand still vastly exceeds supply despite new market entrants.

Clay confirmed 30-40% gross margin range holding at 32% for AI, with adjacent services adding 10-20% at higher margins. Database services at 60-80% margins. Only drag on profitability is under-construction capacity during hypergrowth.

Clay detailed three funding models (customer bring-own-hardware, vendor leasing, Oracle capital). No cash outflow until delivery. Cash flows layer linearly across data centers.

Larry's one-sentence answer: faster networks = faster training = half the cost when paying by the hour. Safra affirmed the simplicity of the competitive moat.

Larry focused on Oracle's database installed base as the key AI enabler rather than providing specific AI revenue or margin numbers. Directional but not quantitative.

Larry explained Oracle starts data centers smaller with higher utilization, then scales with demand. Automation reduces labor costs and human error, improving margins and reliability.

Larry explained OCI's modular rack architecture from 50kW to 1.6GW, emphasizing standardization enables automation across 100+ regions. Safra added that smaller landing footprints allow CapEx to match revenue.

Shantanu reframed the comparison in absolute dollar terms, noting the $2.6B net new ARR guide starts at the same level achieved in FY2025. Emphasized confidence across all customer groups and product lines including Express, Acrobat, Firefly, CC Pro, AEP, and GenStudio.

Shantanu said value of AI will be delivered through software. Adobe wins in both seats and automation. Enterprise still expanding seats while automation adoption accelerates.

Dan explained cross-cloud offerings are growing and the new groupings better reflect innovation at intersections of productivity/creativity and creativity/marketing.