Q2 2025 Summary
Published Feb 7, 2025, 7:58 PM UTC- Microsoft's commercial bookings increased 75% in constant currency, significantly ahead of expectations. This surge was driven by strong Azure commitments from OpenAI and robust execution across core sales motions, including renewals and upsells like E5 and Microsoft 365 Copilot, indicating robust demand across the portfolio.
- Strong adoption and expansion of Microsoft 365 Copilot, with customers expanding seats by more than 10x over the past 18 months, and usage intensity increasing over 60% quarter-over-quarter, suggests significant future revenue growth from AI-powered productivity tools. This reflects growing customer engagement and the potential for increased ARPU growth. ,
- Microsoft is achieving significant cost efficiencies in AI scaling, with inference optimizations improving by 10x per cycle. This reduction in AI costs allows the company to expand AI services to more customers, driving profitability and market share expansion as AI becomes more ubiquitous across their product offerings. ,
- Microsoft is experiencing execution challenges in its non-AI Azure services, particularly with customers reached through scale motions (indirect sales methods), leading to lower-than-expected growth in these areas and potential impact on near-term Azure growth.
- Microsoft's AI infrastructure is currently capacity constrained, with shortages in power and space, potentially limiting the ability to meet high demand for AI services until infrastructure investments are realized, which may affect revenue growth in the near term.
- Free cash flow decreased by 29% year-over-year to $6.5 billion, reflecting higher capital expenditures on AI infrastructure, which could impact Microsoft's financials if high capital spending continues.
| Metric | YoY Change | Reason |
|---|---|---|
Total Revenue | +12% | Strong performance in Microsoft Cloud underpinned overall growth. Continued demand for Azure and other cloud services, along with focus on AI-driven solutions, contributed to higher bookings and consumption throughout the year. |
Productivity & Business | +53% | Growth was driven by robust increases in Office 365 Commercial (seat expansion and higher ARPU), LinkedIn, and Dynamics 365, reflecting both underlying demand and a favorable comparison to prior-year revenue mix changes. |
More Personal Computing | -13% | The decline was primarily due to lower device sales and a cyclical softness in Windows OEM revenue. These factors offset gains in search and news advertising and Xbox services and content. |
United States Revenue | +10% | The steady US enterprise and consumer demand for cloud and productivity solutions helped drive revenue. Ongoing Azure adoption and elevated renewal rates for Microsoft 365 in both commercial and consumer segments contributed. |
Other Countries Revenue | +15% | International growth was spurred by global expansion of Microsoft Cloud, notably Azure, and further boosted by Activision Blizzard content integration in key overseas markets, positioning Microsoft for broader global reach. |
| Metric | Period | Previous Guidance | Current Guidance | Change |
|---|---|---|---|---|
Azure Revenue Growth | Q3 2025 | no prior guidance | 31%–32% in constant currency | no prior guidance |
On-Premises Server Business Revenue | Q3 2025 | no prior guidance | decline in the mid-single digits | no prior guidance |
Enterprise and Partner Services Revenue Growth | Q3 2025 | no prior guidance | low- to mid-single digits | no prior guidance |
More Personal Computing Revenue | Q3 2025 | no prior guidance | USD 12.4B to 12.8B | no prior guidance |
Windows OEM and Devices Revenue | Q3 2025 | no prior guidance | decline in the low- to mid-single digits | no prior guidance |
Search and News Advertising ex TAC Revenue Growth | Q3 2025 | no prior guidance | mid-teens | no prior guidance |
Overall Search and News Advertising Revenue Growth | Q3 2025 | no prior guidance | mid- to high single digits | no prior guidance |
Gaming Revenue Growth | Q3 2025 | no prior guidance | low single digits | no prior guidance |
COGS Growth | Q3 2025 | no prior guidance | 19%–20% in CC or USD 21.65B–21.85B | no prior guidance |
Operating Expense Growth | Q3 2025 | no prior guidance | 5%–6% in CC or USD 16.4B–16.5B | no prior guidance |
Other Income and Expense | Q3 2025 | no prior guidance | negative USD 1B | no prior guidance |
Effective Tax Rate | Q3 2025 | no prior guidance | approximately 18% | no prior guidance |
FX Impact | Q3 2025 | no prior guidance | expected to decrease Q4 revenue & COGS growth by >1 pt | no prior guidance |
Capital Expenditures | Q3 2025 | no prior guidance | remain at similar levels as Q2 | no prior guidance |
Productivity and Business Processes Revenue Growth | Q3 2025 | no prior guidance | 11%–12% in CC or USD 29.4B–29.7B | no prior guidance |
Microsoft 365 Commercial Cloud Revenue Growth | Q3 2025 | no prior guidance | 14%–15% in constant currency | no prior guidance |
M365 Consumer Cloud Revenue Growth | Q3 2025 | no prior guidance | mid- to high single digits | no prior guidance |
LinkedIn Revenue Growth | Q3 2025 | no prior guidance | low- to mid-single digits | no prior guidance |
Dynamics 365 Revenue Growth | Q3 2025 | no prior guidance | mid-teens | no prior guidance |
Intelligent Cloud Revenue Growth | Q3 2025 | no prior guidance | 19%–20% in CC or USD 25.9B–26.2B | no prior guidance |
| Metric | Period | Guidance | Actual | Performance |
|---|---|---|---|---|
Productivity & Business Processes Revenue | Q2 2025 | Expected to be between $28.7B and $29.0B | $29,437 million | Beat |
Intelligent Cloud Revenue | Q2 2025 | Expected to be between $25.55B and $25.85B | $25,544 million | Missed |
More Personal Computing Revenue | Q2 2025 | Expected to be between $13.85B and $14.25B | $14,651 million | Beat |
LinkedIn Revenue Growth | Q2 2025 | Expected ~10% year-over-year | 9.35% year-over-year (from $4,195 millionIn Q2 2024 to $4,587 millionIn Q2 2025) | Missed |
Dynamics 365 Revenue Growth | Q2 2025 | Expected mid- to high-teens year-over-year | 21.4% year-over-year (from $1,576 millionIn Q2 2024 to $1,913 millionIn Q2 2025) | Beat |
Windows OEM Revenue Growth | Q2 2025 | Expected to decline in the low to mid-single digits year-over-year | –14.25% year-over-year (from $5,262 millionIn Q2 2024 to $4,512 millionIn Q2 2025) | Missed |
Search and News Advertising Growth | Q2 2025 | Expected to grow in the high teens year-over-year | 10.5% year-over-year (from $3,220 millionIn Q2 2024 to $3,558 millionIn Q2 2025) | Missed |
Gaming Revenue Growth | Q2 2025 | Expected to decline in the high single digits year-over-year | –7.45% year-over-year (from $7,111 millionIn Q2 2024 to $6,581 millionIn Q2 2025) | Met |
| Topic | Previous Mentions | Current Period | Trend |
|---|---|---|---|
AI infrastructure capacity constraints | Q1 2025, Q4 2024, and Q3 2024 all discuss short-term constraints with ongoing investments. | Q2 2025: Still facing capacity constraints; investing heavily in data centers and expecting better balance by end of FY 2025. | Recurring; sentiment remains focused on gradual improvement. |
High AI-related capital expenditures & uncertain ROI | Q1 2025, Q4 2024, and Q3 2024 mention high AI CapEx but with growing confidence in revenue correlation. | Q2 2025: Not explicitly framed as uncertain ROI; executives express confidence in long-term monetization. | Recurring; concern shifting to optimism about ROI. |
Strong adoption of AI products | Q1 2025, Q4 2024, and Q3 2024 also emphasize rapid Copilot and Azure AI uptake. | Q2 2025: M365 Copilot, GitHub Copilot, and Azure AI show significant growth (e.g., Copilot seats up 10x, AI biz at $13B run rate). | Recurring; momentum continues to build. |
OpenAI partnership | Q1 2025: Mutually beneficial; noted equity-method losses offset by strong AI revenue; minimal detail in Q4 & Q3 2024 beyond usage. | Q2 2025: Large driver of Azure bookings; ongoing commitments fuel cloud growth. | Recurring; consistent focus on revenue contribution. |
Slowing growth in non-AI Azure segments | Q1 2025 was in line with expectations; Q4 2024 saw softness in some regions; no slowdown noted in Q3 2024. | Q2 2025: Slower growth attributed to execution challenges in go-to-market for scale motions. | Sentiment turned cautious; some headwinds continue. |
Activision acquisition | Q1 2025, Q4 2024, and Q3 2024 detail significant impacts on gaming revenue and expenses. | Q2 2025: No mention. | No longer discussed post-Q4 2024. |
Potential reallocation of IT budgets to AI | Only explicitly addressed in Q3 2024, where existing budgets were shifting to AI solutions. | Q2 2025: Not discussed. | No further mentions after Q3 2024. |
Cost efficiencies in AI scaling | Q1 2025, Q4 2024, Q3 2024 all highlight cost discipline, unified builds, and margin improvements. | Q2 2025: Emphasized 10x+ software-driven efficiency gains and fungible fleet management. | Recurring; consistently targeting lower AI serving costs. |
Generative AI expansion with large future impact | Q1 2025, Q4 2024, and Q3 2024 also cite major investments in Copilot, Azure AI, and global infrastructure. | Q2 2025: AI revenue run rate of $13B, up 175% YoY, broadening enterprise adoption. | Recurring; viewed as a key long-term growth driver. |
-
Strong Commercial Bookings
Q: What drove the surge in commercial bookings?
A: Microsoft saw a 75% increase in commercial bookings this quarter, the highest in a decade. This was driven by significant Azure commitments from OpenAI, ongoing renewals, upsells like Copilot and GitHub Copilot, and strong suite momentum with E5 licenses. The growth was broad-based, with both existing customers increasing commitments and new customers coming on board. -
OpenAI Partnership and CapEx
Q: How is the OpenAI relationship affecting CapEx needs?
A: Microsoft remains committed to its partnership with OpenAI, which has made significant Azure commitments. The investments are focused on building a fungible fleet that balances training and inference, optimizing costs through Moore's Law and software improvements. CapEx growth is pivoting towards more CPU and GPU investments directly correlated to revenue, supporting both OpenAI and broader customer contracts. -
Azure Execution Issues
Q: What caused the execution challenges in Azure?
A: Non-AI Azure consumption faced challenges due to adjustments in scale motions, particularly with customers reached indirectly through partners. Changes in sales strategies to balance AI workloads and migrations caused short-term impacts, but adjustments are underway to restore balance. Azure AI results were better than expected, and confidence in AI growth remains strong despite capacity constraints. -
AI Revenue Drivers
Q: What contributed to the AI revenue beat?
A: AI revenue exceeded expectations due to strong performance in Azure AI services and Microsoft Copilot, with significant growth in new and expanded seats. Increased usage adds value and supports a strong price per seat, even though usage doesn't directly impact revenue. -
Copilot Adoption
Q: How is Copilot adoption progressing?
A: Copilot is gaining traction in departments seeking immediate productivity gains, such as sales, finance, and supply chain. Collaboration across functions is driving broader enterprise adoption, with network effects leading to wider deployment. Microsoft is simplifying adoption with Copilot Chat, offering flexibility for widespread use across enterprises. -
Copilot Go-to-Market Strategy
Q: How will Copilot evolve to reach more customers?
A: Microsoft is expanding Copilot's reach by introducing Copilot Chat in M365, enabling deployment across the entire installed base. By leveraging inference optimizations that reduce costs, premium features are becoming more accessible. Similar expansions are happening with GitHub Copilot and Security Copilot, enhancing availability across the portfolio. -
AI Scaling and Cost Reduction
Q: Are AI costs decreasing as it scales?
A: Innovations like DeepSeek are driving significant 10x improvements per cycle through software optimizations in AI inference. This reduces token and inference computing prices, enabling broader consumption and more application development. AI models are becoming more efficient, even running on PCs, making AI more ubiquitous and beneficial for Microsoft. -
Proprietary vs. Open Models
Q: Does it matter if inference uses open or proprietary models?
A: Microsoft anticipates a combination of proprietary and open-source models in applications. It's crucial to start with the best models and continuously optimize for cost and latency, regardless of their origin. Investments in platforms like Foundry help developers manage the complexity of multiple models, ensuring operational efficiency.