MICROSOFT CORP Q1 FY2027 earnings call

• AI Agent Adoption: Copilot daily users increased nearly 50% quarter-over-quarter, with automated tasks nearly doubling since July. Customer deployments are scaling rapidly, evidenced by Volkswagen’s plan to deploy to 330,000 employees.
• Product Strategy: Copilot is positioned as the 'new OS for work,' integrating chat, cowork, autopilot, and code. The E7 suite (Copilot E5, Entra, Agent 365) is gaining traction across industries.
• Monetization Model Shift: Microsoft is expanding from per-seat subscriptions to include consumption-based billing. Over 75% more customers are actively consuming credits through usage-based models, particularly in GitHub.
• GitHub & Development: Agent-generated pull requests now account for over one-third of all PRs on the platform. GitHub Advanced Security revenue grew over 50% year-over-year.
• Infrastructure Efficiency: Engineering teams delivered significant efficiency gains, allowing faster allocation of capacity to high-demand areas like Foundry. Dock-to-live times for new GB300s were reduced by approximately 20%.
• First-Party Models: Microsoft continues to advance its own models (MAI series), aiming to deliver capabilities comparable to leading frontier models within three to six months at a fraction of the training cost.
• Security & Governance: Agent 365 registry has grown to over 50 million agents. Microsoft Purview has audited over 7.6 billion agent activities, emphasizing the separation of intelligence from authority.

Revenue from Agents and Infrastructure was $76.8 billion, growing 25% year-over-year. Microsoft Cloud revenue reached $63.9 billion, with accelerated growth of 30% (31% in constant currency). Azure revenue increased 48%, outperforming expectations. M365 Commercial Cloud revenue grew 18% (19% in constant currency), while M365 Consumer Cloud revenue rose 17% (16% in constant currency). Productivity and Server Licensing revenue decreased 3% (2% in constant currency). Industry Solutions Cloud revenue increased 10%. Frontier and Support Services revenue grew 4% (5% in constant currency). Devices and Consumer revenue was $15.5 billion, declining 3%. Search and Advertising revenue grew 12%, ahead of expectations. Xbox content and services revenue decreased 2% (1% in constant currency). Windows OEM and devices revenue fell 22% (27% in constant currency).

• Agents and Infra Revenue: Expected to be between $80.1 billion and $80.7 billion, representing 24-25% growth.
• Azure Growth: Projected to grow 49-50% in constant currency for Q2, driven by efficiency gains and strong demand exceeding supply.
• M365 Commercial Cloud: Expected to grow approximately 19% in constant currency, with continued momentum in premium seat-based and usage-based offerings.
• M365 Consumer Cloud: Anticipated to grow in the low teens, down sequentially due to lapping prior-year price increases.
• Devices and Consumer Revenue: Forecasted between $15.9 billion and $16.4 billion. Windows OEM and devices revenue expected to decline in the mid-30s due to elevated inventory levels and weaker PC market demand.
• Total Company Revenue: Guidance set between $96.0 billion and $97.1 billion, implying 18-19% growth.
• Capital Expenditures: Q2 CapEx expected to exceed $50 billion. Full-year spend is projected at approximately $170 billion, reflecting delayed third-party GPU deliveries shifting to the second half.
• Operating Margins: Expected to decrease slightly year-over-year, impacted by sales mix shifts to Azure and higher COGS growth.

• Capacity Delays: Third-party GPU capacity deliveries intended for Q1 were delayed, impacting capital expenditure timing and potentially affecting near-term supply constraints.
• PC Market Weakness: Elevated inventory levels among OEMs and channel partners, combined with weaker general PC market demand, continue to negatively impact Windows OEM revenue.
• Sales Execution Variability: Weaker sales execution in specific M365 SKUs (E3 and standalone) offset stronger performance in premium offerings, highlighting variability in sales motions during the transition to new monetization models.
• Forward-Looking Statement Uncertainty: Actual results may differ materially from projections due to risks, uncertainties, and assumptions inherent in forward-looking statements regarding AI adoption, infrastructure build-outs, and market conditions.

Q: Adam Wood (Morgan Stanley) asked whether returns will be driven by improving Azure utilization/unit costs or monetizing layers across Fabric, Foundry, and Copilot.

A: Satya Nadella stated that leverage comes from the entire integrated stack rather than isolated components. He emphasized treating infrastructure as a fungible pool where efficiencies in data center operations allow optimized workload allocation. Amy Hood added that engineering efficiencies enable better capacity allocation, reinforcing that value compounds when working across the unified platform.

Q: Karl Keirstead (UBS) sought clarification on the drivers behind the strong Azure growth guidance of 49-50% in Q2.

A: Amy Hood attributed this to significant engineering efficiency gains across the platform layer, including CPU, GPU, storage, and memory optimization. She noted that these efficiencies allowed greater allocation to Foundry despite some third-party supply delays. Additionally, favorable token pricing and the ability to quickly convert existing fleet capacity contributed to the robust outlook.

Q: Mark Moerdler (Bernstein) asked about the customer mix driving Azure AI strength, specifically enterprise vs. digitally native workloads.

A: Satya Nadella highlighted that enterprise workloads are increasingly deploying sophisticated agentic systems for complex tasks like supply chain management and banking processes. He observed a shift away from 'shiny object' experiments toward practical, reliable, and secure enterprise-grade solutions. This indicates a broadening enterprise motion, though he cautioned it remains a long-term process.

Q: Gabriela Borges (Goldman Sachs) inquired about the evolving token mix between frontier OpenAI models and Microsoft's first-party MAI models.

A: Satya Nadella described a three-tier toolbox: using frontier models for peak needs, leveraging multi-model routing (like Copilot Fusion) for cost/performance optimization, and training custom open-weight or MAI models for proprietary use cases. He explained that customers dynamically choose based on pricing and requirements, often starting with frontier models before moving to optimized or custom alternatives.

Q: Brad Reback (Stifel) asked if unlocking consumption meters for core M365 customers would be fast without significant sales effort, similar to GitHub.

A: Satya Nadella noted that coding adoption via consumption may be faster due to market familiarity with pay-as-you-go models. However, broader M365 consumption requires careful education to balance subscription value with usage-based flexibility. Amy Hood confirmed that investments in the Frontier team and updated sales motions are critical to helping customers understand optimal usage across both licensing models.