Microsoft was sued on June 12, 2026, in federal court in Seattle by a Michigan public pension fund, alleging the tech giant misled investors about slowing growth in its Azure cloud division and the spiraling costs of its AI infrastructure build-out. The lawsuit, filed in the Western District of Washington, claims Microsoft executives made materially false and misleading statements regarding the company's capital expenditures, GPU availability, and the economic viability of its Copilot AI assistant—ultimately harming shareholders when the truth began to emerge.

The complaint, brought by the Michigan Municipal Employees' Retirement System, seeks class-action status on behalf of investors who purchased Microsoft stock between January 2024 and May 2026. It details a series of alleged misrepresentations about Azure's growth trajectory, the scalability of its AI data centers, and the return on investment for Copilot, Microsoft's flagship AI product integrated into Windows and Office.

At the heart of the suit is the contention that Microsoft knew, or should have known, that its aggressive AI spending was not translating into proportionate revenue gains, and that constraints in the GPU supply chain were severely limiting the deployment of AI services—even as the company projected confidence to Wall Street.

Azure Growth: The Deceleration That Wasn't Disclosed?

For years, Azure has been Microsoft's growth engine, consistently delivering revenue increases above 25% year-over-year. But in early 2025, signs of a slowdown became apparent. The lawsuit alleges that Microsoft attributed any deceleration to "macroeconomic headwinds" and "customer optimization cycles," while downplaying the internal projections that showed a structural plateau in cloud demand.

According to the complaint, internal documents reveal that Microsoft's own forecasts for Azure growth in fiscal 2026 had been revised downward multiple times by late 2024, yet public guidance remained optimistic. The pension fund argues that this disconnect between private forecasts and public statements artificially inflated Microsoft's stock price.

The Capex Explosion: Billions Poured Into AI, But Returns Uncertain

To support its AI ambitions, Microsoft has poured tens of billions of dollars into data centers, custom silicon, and partnerships with OpenAI. Capital expenditures grew from $12.3 billion in fiscal 2023 to over $28 billion in fiscal 2025, with no signs of abating. The lawsuit points to this spending spree as a red flag that was not adequately communicated to investors.

CFO Amy Hood has repeatedly assured investors that this spending would generate strong returns, tied to "concrete customer contracts" and "unprecedented demand for AI infrastructure." However, the lawsuit argues that Microsoft's internal capacity utilization rates were far below what was publicly implied. By mid-2025, a significant portion of new AI compute capacity remained unutilized because enterprise customers were slow to adopt GPU-intensive workloads. The depreciation timeline for these massive capital assets—often just 4-5 years—means that every quarter of underutilization chips away at potential returns.

While CEO Satya Nadella frequently noted that "every dollar of capex is tied to customer demand," the pension fund claims Microsoft was building capacity far in excess of near-term demand, gambling on a future AI boom that might not materialize on the expected timeline. The glut of AI infrastructure, the suit argues, now threatens margins and raises the risk of impairment charges.

GPU Scarcity: The Hidden Bottleneck

A central allegation is that Microsoft misled investors about the availability of GPUs, the essential hardware for training and running AI models. Despite public statements that the company had "strong supply chain relationships" and was "managing inventory effectively," the complaint cites internal emails from early 2025 showing that Microsoft's GPU allocations were falling 30-40% short of requirements.

This scarcity forced Microsoft to make difficult trade-offs: delaying expansion of Azure OpenAI Service for new regions, postponing the general availability of Copilot features that required real-time inference, and even reallocating GPUs from profitable cloud services to unproven AI projects—effectively cannibalizing existing revenue streams. One internal email cited in the complaint from a cloud sales executive reads, "We're turning away [Fortune 500] companies because we can't guarantee GPU capacity for their pilots. This is hurting our credibility in the market." If proven, this directly contradicts Satya Nadella's assertion in a January 2025 earnings call that "we have the infrastructure to meet customer AI demand across every sector."

Copilot Economics: High Costs, Slow Adoption

When Microsoft launched Copilot for Microsoft 365 in 2023, the company set a price of $30 per user per month, leading industry analysts to project that Copilot could generate over $40 billion in new annual revenue within five years—a figure that Microsoft's leadership never publicly corrected. The lawsuit alleges that Microsoft knew such projections were unrealistic because the true cost of delivering Copilot services was much higher than anticipated.

Each Copilot query requires significant computational power, and with millions of potential users, the inference costs quickly cascade. The complaint asserts that Microsoft's internal financial models, shared with select senior leaders, showed that even at scale, Copilot's gross margins would be negative or barely breakeven unless model efficiency improved by orders of magnitude. Instead of disclosing these headwinds, Microsoft touted "strong enterprise engagement" and "record seat additions," while failing to mention that a large percentage of those seats came from heavily discounted bundle deals or free trial periods that were unlikely to convert.

The pension fund points to a May 2026 report from Morgan Stanley that downgraded Microsoft, noting that "Copilot adoption has decelerated, and competitive offerings from Google and Salesforce are eroding the premium Microsoft can charge." That report triggered the sharpest single-day drop in Microsoft's stock in two years.

The lawsuit is grounded in Section 10(b) of the Securities Exchange Act of 1934 and SEC Rule 10b-5, which prohibit fraudulent misstatements or omissions in connection with the purchase or sale of securities. To prevail, the plaintiff must show that Microsoft made materially false statements, that those statements were made with scienter (knowledge or reckless disregard of the truth), and that the stock price declined when the truth was revealed.

The pension fund points to several "corrective disclosures"—events where the market learned of the alleged troubles. These include Microsoft's April 2026 earnings report, where Azure growth came in below expectations and CFO Amy Hood warned of "moderating AI capex returns," causing the stock to drop 8% in one day. Another was a May 2026 analyst report detailing unsold data center capacity, leading to a further 12% decline.

Plaintiffs estimate damages could run into the billions of dollars, covering the losses incurred by all affected shareholders during the class period. Microsoft has not yet filed a response, but legal experts expect a vigorous defense, arguing that forward-looking statements were protected under the Private Securities Litigation Reform Act's safe harbor provision and that the risks were adequately disclosed in public filings.

Broader Implications for the AI Industry

This lawsuit comes at a time when Big Tech's AI spending is under intense scrutiny. Google, Amazon, and Meta have all faced similar questions about the return on their massive investments. If the case proceeds, it could set a precedent for how companies must disclose the risks of AI capex and the operational constraints of emerging technologies.

For Microsoft, the suit is a reputational blow as it battles to maintain its early lead in the AI race. Analysts are already adjusting their models to reflect the possibility that AI infrastructure spending could weigh on earnings for years, even if long-term demand materializes. The case may also embolden other institutional investors to file their own suits, given the widespread ownership of Microsoft stock.

What Investors Should Watch

As the lawsuit unfolds, several key developments will be critical: whether the judge certifies the class, discovery processes that could reveal internal communications about GPU supply and Copilot adoption, Microsoft's quarterly filings where further changes in capex guidance or write-downs could corroborate the allegations, and any regulatory inquiries from the SEC or other bodies into Microsoft's disclosures.

For now, Microsoft's stock has partially recovered from the post-disclosure dips, but the cloud of litigation is unlikely to lift quickly. The case highlights a fundamental tension in the AI era: the mandate to build boldly for the future versus the obligation to provide truth-in-the-moment to those funding the buildout.

In the end, the Michigan pension fund's lawsuit may do more than seek compensation—it could force a reckoning over how the tech industry communicates the colossal risks hidden inside its most transformative bets.