Microsoft’s AI Offensive Accelerates: From KPMG to Own Models, but the Market Remains Unimpressed

Microsoft partners with KPMG and Mayo Clinic while launching MAI models, but $190B capex and OpenAI losses weigh on shares, down 13.65% YTD.

Microsoft is sprinting in two directions at once, signing blockbuster enterprise partnerships while simultaneously building the technology to go it alone. The result is a company firing on all operational cylinders—yet the stock continues to slide, caught between staggering capital commitments and investor impatience.

The software giant this week unveiled one of the year’s largest corporate AI pacts, with KPMG rolling out Microsoft 365 Copilot to its global workforce of more than 276,000 employees. The accounting and consulting firm will also weave Agent 365 into its internal KPMG Workbench, a hub that coordinates AI tools for client service. “We’re helping customers embed AI deeper into their workflows—moving from experimentation to enterprise-wide impact,” said Deb Cupp, Microsoft’s Global Enterprise Chief Revenue Officer. KPMG has been classified as a “Frontier Firm”, a designation for organizations fundamentally rebuilding their operating model around artificial intelligence.

The KPMG deal came hot on the heels of a separate collaboration with the Mayo Clinic to develop a clinical AI model trained on anonymized patient data, research findings, and medical expertise. While the Mayo Clinic retains ownership of the model, Microsoft plans to make it available to other healthcare organizations through the Azure Foundry platform.

On the product side, Microsoft is also accelerating its push toward self-reliance. At the Build 2026 developer conference, the company introduced seven proprietary MAI models, led by MAI-Thinking-1, which targets complex mathematics and software development. A faster companion, MAI-Code-1-Flash, is already integrated into GitHub Copilot and VS Code. Specialized image and transcription models round out the suite. The strategic shift is clear: AI will no longer be a passive assistant but an active agent embedded across the entire operating system. By developing its own models, Microsoft reduces dependence on external partners and gains tighter control over margins.

That independence comes at a price. Microsoft’s AI business is now generating $37 billion in annual revenue, up 123% from a year earlier. Azure and other cloud services grew 40% last quarter, beating analyst expectations. But the cost of building infrastructure is skyrocketing. The company plans roughly $190 billion in capital expenditure for calendar 2026—a 61% jump from 2025. CFO Amy Hood attributed $25 billion of that increase to pricier components, particularly scarce storage chips. Additionally, losses from the OpenAI investment reached $3.1 billion in the first quarter of fiscal 2026, though royalty-free rights to OpenAI technology through 2032 and an Azure growth target of 39–40% for the fourth quarter are meant to offset the drag.

Despite the operational momentum, the stock has been unresponsive. Shares recently traded at €348.50, down 13.65% year-to-date and roughly 27% below the 52-week high of €478.10 set in October 2025. The relative strength index sits at 42.8, indicating a neutral to slightly oversold condition. UBS reiterated a buy rating, citing the combined cloud revenue of Microsoft, Amazon, and Oracle—$84.8 billion in the first quarter of 2026, up 39% year-over-year. Investors will collect a dividend of $0.91 per share in June, yielding about 0.9%, but the bigger catalyst will be whether the capex burden eases meaningfully in the second half of 2026.

For now, Microsoft is fighting a two-front war: deploying enormous sums to build its AI empire while the market waits for visible returns. The pieces are all in place—partnerships, proprietary models, a booming cloud business—but the stock’s trajectory will depend on when those pieces start delivering margin growth instead of just revenue.