Microsoft (MSFT) reported fiscal 2026 fourth-quarter (April-June) results on July 29, highlighting the accelerating expansion of its AI-related businesses. Revenue rose 18% year-over-year to $90.07 billion, while net profit surged 31% to $35.766 billion, marking substantial top- and bottom-line growth. The core driver was the cloud platform Azure, where quarterly revenue growth accelerated to 43%, pushing full-year revenue past the $100 billion threshold for the first time.
Paid subscriptions for the generative AI assistant Microsoft 365 Copilot surpassed 30 million. CEO Satya Nadella stated during the earnings call that this reflects “the trust customers place in us as they drive AI transformation.” However, amid intensifying AI development competition, quarterly capital expenditures swelled 70% year-over-year to $41 billion, with the company adjusting its calendar 2026 investment plan to approximately $175 billion. Balancing growth investment with profitability now becomes the key focus.
Cloud and AI Drive Structural Business Transformation
By business segment, Intelligent Cloud revenue grew 32% to $39.306 billion, leading overall performance. Azure’s accelerating growth was particularly notable, with CFO Amy Hood explaining that “customer demand continues to outpace available compute capacity.” Efficiency improvements in CPU/GPU clusters and the rapid ramp-up of new compute capacity supported the high 43% growth rate.
Productivity and Business Processes revenue rose 14% to $37.847 billion. Office products and LinkedIn performed solidly, with Copilot’s expanding paid subscription base beginning to contribute to earnings. According to Nadella, the number of large customers with over 50,000 seats grew more than sevenfold year-over-year, while per-user AI interactions roughly doubled. The time required to reach over 80% monthly active usage post-deployment has shortened “from months to days,” with weekly usage frequency now matching that of Outlook and Teams.
In contrast, the More Personal Computing segment declined 4% to $12.854 billion, weighed down by weakness in Windows OEM and Xbox. In early July, Microsoft announced 3,200 job cuts and studio closures within its Xbox division, signaling ongoing restructuring of non-core operations.
“Measure by Business Metrics, Not Lines of Code”
As AI adoption scales up, Microsoft is articulating a clear philosophy on measuring return on investment (ROI). Charles Lamanna, Corporate Vice President overseeing the Copilot and AI agent platform, stated in an interview with Impress Watch that “having AI-specific metrics should be avoided.” He emphasized that AI adoption outcomes should be judged by existing business metrics such as revenue, costs, customer satisfaction, and market share.
Lamanna cited the management of his own engineering team as an example: “I don’t look at how many thousands of lines of code were written with AI or how many AI pull requests were made. What matters is driving adoption, reducing support tickets, and improving product quality and performance.” This philosophy aligns with Microsoft’s strategy of evolving AI from a mere efficiency tool into “agentic AI” that fundamentally redesigns organizational workflows.
Indeed, enterprise AI utilization is shifting from “time savings” to “autonomous” applications. Microsoft is accelerating the rollout of products enabling more advanced autonomous task execution, including Scout, an always-on personal agent announced in June, and Copilot Cowork, which allows multiple agents to collaborate on report generation and analysis. The number of registered agents on Agent 365 reached nearly 40 million within roughly two months of launch, while AI agents are now involved in approximately one-third of code merge requests on GitHub Copilot.
Multi-Model Strategy and the Role of Proprietary AI “MAI”
A distinctive feature of Microsoft’s AI strategy is its model-agnostic, multi-model approach. Copilot currently employs a system that automatically selects among OpenAI’s GPT-5.6, Anthropic’s Claude, and Microsoft’s proprietary MAI model based on the task. Lamanna explained: “Over the past several months, we’ve automatically switched to superior models as they’ve been released. Customers don’t need to evaluate models themselves.”
Regarding MAI, the company’s proprietary frontier model, Lamanna positioned it as offering “the best value in terms of cost, performance, and latency” in areas such as image and speech recognition, while explicitly stating that MAI receives no preferential treatment within Copilot. He would only say it is “used when it is the best option for the customer.”
Microsoft is also actively embracing open-source models. Addressing regulatory debates around models originating from China, Lamanna commented: “We are big believers in open-source models and open-source weights. The country of origin doesn’t matter.” He further emphasized data sovereignty, noting that by hosting open models on Microsoft’s cloud infrastructure, “data never goes back to the company that created the model; it stays within Microsoft.”
Anthropic Investment Yields $3.2 Billion Gain, Revealing Multi-Layered Financial Structure
Special factors also influenced this quarter’s net profit. Microsoft recorded a $3.2 billion valuation gain from its investment in AI startup Anthropic, while its investment in OpenAI resulted in a net loss of $480 million. This was primarily due to accounting treatment changes following OpenAI’s conversion to a for-profit entity at the end of 2025.
Additionally, impairment charges related to the Xbox division and costs associated with early retirement programs were incurred. On the personnel front, Microsoft’s global full-time headcount stood at approximately 223,000 as of the end of June 2026, down roughly 5,000 year-over-year. The company announced approximately 4,800 additional job cuts in July.
The rapid expansion of capital expenditures is also a key focus for the market. Full-year fiscal 2026 capex totaled $145.3 billion, with the quarterly figure reaching $41 billion, a 70% year-over-year increase. CFO Hood explained that roughly two-thirds of this is in short-lived assets such as CPUs and GPUs, where “procurement can be slowed relatively quickly if demand changes.” She also noted that an accounting policy change extending data center useful life from 15 to 25 years adjusted the calendar 2026 capex outlook to approximately $175 billion, but stated that “the actual investment plan has not changed.”
The Dilemma of Internal Compute Resource Allocation
Behind the strong earnings, Microsoft faces difficult management decisions regarding compute resource allocation. According to a Business Insider report, the company is already experiencing severe compute capacity shortages and has been forced to seek assistance from competitors Amazon and Google.
The dilemma: allocate limited compute resources to Azure customers to drive immediate revenue growth, or prioritize them for developing proprietary products like Copilot and AI agents. Choosing the former would contribute to short-term revenue but risks falling further behind in the AI development race, where Microsoft already lags in some areas. The latter would strengthen long-term competitiveness, but if Azure’s growth rate falls short of expectations, it could put downward pressure on the stock price.
The article reported that difficult discussions are occurring internally, with one source framing the dilemma as: “Satya would never prioritize Adobe’s creative tool growth over Microsoft 365 growth, so even if the conclusion is that compute resources should prioritize internal use, how do you communicate that message to enterprise customers?”
Market Assessment and Future Focus
Following the earnings release, Microsoft shares rose approximately 8.8% in after-hours trading to $425.01. While this helped recover from a roughly 19% year-to-date decline, some believe a full re-rating will still take time.
Analysts offered mixed assessments. Brad Reback of Stifel noted that “Azure’s accelerating growth and Copilot’s expanding seat count are evidence that AI investments are beginning to bear fruit.” However, others are closely watching the 23% year-over-year decline in free cash flow to $19.639 billion for the April-June quarter, as well as the decline in cloud gross margins from 68% to 65%.
For the first quarter of fiscal 2027 (July-September), CFO Hood provided an outlook for Azure revenue growth to accelerate to approximately 45%, with growth expected to continue through the first half of the fiscal year. Capital expenditures are projected to exceed $50 billion. As AI demand continues to expand, the key question going forward is whether Microsoft can sustain a virtuous cycle where front-loaded investment translates into earnings growth.