Copilot class action targets Nadella as 15M paid seats lag 450M users

Microsoft is dealing with two linked headaches in 2025: a class action over Copilot claims and operational pressure from AI-driven usage spikes at GitHub. The lawsuit, filed by the City of St. Clair Shores Police and Fire Retirement System in Seattle US District Court, alleges Microsoft made “materially false and/or misleading” statements about how Copilot is being adopted. It names Microsoft bosses including CEO Satya Nadella, and it argues the reality is worse than what investors were told.

The adoption numbers in the case point to the kind of mismatch that makes regulators and boards nervous. On January 28, Microsoft announced results for its fiscal second quarter, including a slowdown in Azure growth and a disclosure that paid Microsoft 365 seats had reached only 15 million out of 450 million Microsoft 365 users. The complaint ties this to investor-impactful outcomes, saying Microsoft’s flagship proprietary AI model “ranked well below competitors on a number of benchmark tests,” and that Microsoft “had failed to convert a significant percentage of its commercial Microsoft 365 users to paid Copilot subscriptions,” with Copilot “lost market share to rival products,” a trend the complaint says was increasing.

In plain English: the lawsuit is arguing that investors got an overly optimistic story about Copilot performance and customer uptake, while the reported numbers show conversion lagging and competitive pressure building. That matters because Copilot is not just a feature. It is the monetization engine Microsoft is trying to scale on top of a massive installed base. If paid conversion stays stubbornly low while competing AI tools gain traction, the narrative shifts from “AI acceleration” to “AI adoption friction,” and markets can punish the gap.

The market reaction described in the complaint underscores why this is not an abstract legal exercise. After Microsoft’s January 28 fiscal second-quarter announcement, the company’s shares reportedly declined by more than $48 per share, around ten percent of their value at the time, according to the complaint. Microsoft disputes the allegations. A spokesperson told The Register: “We are aware of the complaint and believe the claims are without merit. Microsoft stands by the integrity of its public statements and will vigorously defend itself in court.” That statement does not remove the core issue for boards: even when a company believes it is right, the time, cost, and uncertainty of a class action can drag down decision-making and distract executives who need to focus on delivery, pricing, and scaling.

The lawsuit also specifically claims that Microsoft’s SEC filings did not clearly explain problems “regarding the development and customer adoption of Copilot products and Microsoft’s proprietary AI models.” That phrasing is key. In US markets, the legal risk often hinges on disclosure clarity, not whether a product exists. Investors typically want to understand what is working, what is not, and where performance and adoption metrics are coming from. When AI is involved, expectations can get out ahead of real-world capacity and customer behavior, and that gap becomes a governance problem.

Meanwhile, Microsoft’s AI-related stress is reportedly spilling into infrastructure, and this is where the second act of the story gets uncomfortable for an enterprise giant that sells cloud reliability. Microsoft-owned GitHub, which has sometimes struggled with availability amid a surge in AI-assisted workflows, is reportedly facing reliability and scalability woes. Reports say the site has attempted to shift workloads to Azure, but for many users it has remained unreliable. Azure, for its part, has had “infamously” capacity problems recently, according to The Register.

The most eyebrow-raising operational claim is that GitHub may be “propped up with additional resources from AWS.” The Register notes it is not clear whether this is temporary triage or something more permanent. Either way, it is a classic multicloud tension: even if one vendor owns you, the performance of AI-assisted workflows depends on capacity that can be scarce, throttled, or simply overloaded. The Register adds that “given the choice, few IT managers would entrust all their workloads to a single vendor,” and that a multicloud approach is sensible.

For executives, the second-order implication is straightforward. When AI demand spikes, reliability stops being a technical footnote and becomes a revenue and reputation risk. If GitHub experience degrades, developers look elsewhere. If Copilot adoption lags, customers hesitate to pay for upgrades they do not feel yet. And if both issues happen under the same corporate umbrella, board oversight becomes more complex: legal counsel, cloud engineering, product leadership, and investor communications all end up in the same room.

There is also a governance signaling problem. The Register frames it as “a little embarrassing when your owner operates its own cloud service,” which may be editorial snark, but the underlying issue is real. Customers expect the parent company to lead on reliability, especially when its own roadmap increasingly depends on AI scale. For leaders at peers who are rolling out Copilot-like tools, the lesson is less about blaming any single benchmark or outage and more about building adoption plans that match capacity realities, with disclosures that do not leave investors filling in the blanks.