🖥️ Microsoft FY26 Q1 Earnings — Core Brief Edition

Microsoft opened FY26 with AI-driven outperformance across Cloud, Copilot and Azure, a record commercial bookings surge (boosted by OpenAI), and a capacity-constrained AI buildout that will keep capex high as demand arrives faster than supply.

1) Big Numbers

• Revenue: $77.7B (+18% YoY, +17% cc)
• Microsoft Cloud: $49.1B (+26% YoY, +25% cc) — well ahead of plan
• Operating income: +24% YoY
• EPS: $4.13 (+23% YoY, +21% cc; excl. OpenAI investment effects)
• Commercial bookings: +112% YoY 🤯 — driven by large Azure commitments incl. OpenAI
• Commercial RPO: $392B (+51% YoY) — ~2 years average duration → it’s short-dated, real usage
• Cloud GM%: ~68–69%, slightly lower YoY because of AI infra scaling
• CapEx: $34.9B in the quarter (!) — about half = short-lived (GPU/CPU), half = long-lived DC sites; cash PP&E $19.4B
• FCF: $25.7B (+33%)
• Capital returned: $10.7B (div + buybacks)

2) OpenAI — “next chapter”

• New definitive agreement signed “yesterday” (after quarter) — management calls it one of the most successful partnerships in the industry.
• Microsoft has already 10x’d its OpenAI investment as contracted and now has incremental $250B of Azure services tied to OpenAI over time.
• Rev-share, exclusive IP, and Azure API exclusivity now run until AGI or through 2030; model/product IP extended to 2032.
• None of this hit Q1 — but it drove bookings and will sit in RPO / future capacity.
• Amy: future OIE will be more volatile because of OpenAI accounting.

3) Azure & AI Platform

• Azure & other cloud services: +40% YoY (+39% cc) — took share again.
• But: demand > capacity across AI workloads — even as MSFT brings GPUs online.
  • They will be capacity constrained through at least the rest of FY26.
  • Q2 Azure guide: ~+37% cc — still very strong, but gated by supply and internal priorities.
• They’re building a “planet-scale cloud & AI factory”:
  • AI capacity +80% this year
  • DC footprint ~2x in 2 years
  • New 2 GW AI DC in Wisconsin (Fairwater) coming.
  • First large-scale cluster of Nvidia GB300s already deployed.
• Fleet is built to be fungible (pre-train → post-train → inference → R&D → 1P products → 3P customers) and continuously modernized (they boosted GPT-4.1 & GPT-5 token throughput +30% per GPU just by software).
• Digital sovereignty: Azure AI in 33 countries; OpenAI + SAP will use Azure to deliver AI to German public sector.

4) Copilot, agents, 1P apps

• 900M monthly active users of AI features across MSFT products.
• 150M+ monthly active users of the family of Copilots (info work, coding, security, health, consumer).
• Microsoft 365 Copilot:
  • Adopted faster than any other M365 suite add-on.
  • 90% of the Fortune 500 now using M365 Copilot.
  • Big seat buys: PwC 200K+, Lloyds 30K (saving 46 minutes/day per employee), Accenture, BMS, EY, UK tax authority all at 15K+.
  • New agent mode: turns a single prompt into a full Word/Excel/PPT and then iterates “on-rails”.
  • Teams mode makes Copilot multi-player; 3rd-party agents (Adobe, SAP, ServiceNow, Snowflake, Workday…) can plug in.
• GitHub:
  • 26M Copilot users; 180M total developers (growing fastest ever; 1 dev/sec).
  • 80% of new devs use Copilot in 1st week.
  • New Agent HQ = mission control for multi-agent coding (OpenAI, Anthropic, Google, xAI, OSS, in-house) with governance.
• Security: 3 dozen agents in Copilot across Defender/Entra/Purview/Intune → phishing triage agent makes analysts up to 6.5x more efficient.
• Health: Dragon Copilot documented 17M patient encounters in the quarter (~5x YoY); 650+ health orgs using ambient.

5) Why capex is exploding (and why they’re ok with it)

• Booked business is already huge (RPO $392B; up 51%; short duration). They’re not building DCs “on spec.”
• Half of capex is short-lived GPUs/CPUs → lifetimes match the contract duration → risk is lower than headline number.
• Long-lived assets (sites, power) are 15–20 yr → needed for multi-refresh AI cycles.
• Amy: “we thought we would catch up; we are not. Demand is increasing… across many places.”
• Q2 and FY26: capex goes higher than FY25.

6) Other segments (quick)

• Productivity & Business Processes: $33B, +17%
  • M365 commercial cloud +17% (1 pt in-period rev)
  • M365 consumer cloud +26%; subs 90M+
  • LinkedIn +10% (hiring still soft)
  • Dynamics 365 +18%
• Intelligent Cloud: $30.9B, +28%
  • SQL DB Hyperscale +75%, Cosmos DB +50%
  • Fabric revenue +60%; 28K paying customers
• More Personal Computing: $13.8B, +4%
  • Windows OEM/devices +6% (Win10 EOS pull-ahead)
  • Search & news X-TAC +16%
  • Gaming –2% vs tough comp, but content/services +1%; 155M MAU in Minecraft.

7) Guidance / color

• Q2 revenue: $79.5–$80.6B (+14–16%), FX tailwind.
• Azure Q2: ~+37% cc — still capacity constrained.
• Cloud GM% to be ~66% (down YoY) because of AI + Azure mix.
• Expect bookings volatility because of very large, multi-year AI deals (like OpenAI).

8) What the street is worried about (and what they said)

• “Are we in an AI bubble?”
  • MSFT says: demand is real, booked, short-dated, and broad (not just one customer).
  • Fleet is fungible → lowers concentration risk.
  • They’ll say no to shapes of demand that don’t fit the long-term margin/footprint.
• AGI clause in OpenAI deal?
  • Satya: with our definition of AGI “not anytime soon.”
  • Real value will come from systems of agents that smooth “jagged intelligence,” not just bigger models.

9) Core takeaways

1. AI pull is stronger than supply → expect high capex + capacity-limited Azure growth for a few quarters.
2. Cloud is re-accelerating (Azure +40%, guide +37%) despite those limits.
3. Copilot is working at scale (150M+ MAU, 90% F500) → AI is already a P&L item, not a demo.
4. OpenAI deal locks in $250B Azure and IP/exclusivity through 2030/2032 → moat got deeper.
5. RPO $392B, 2 yrs → a lot of this turns into revenue fast.

MSFT 1Q26

0:00

/3520.416

1×