UK’s AI probe into Microsoft isn’t just about Windows—it’s about control

The UK’s Competition and Markets Authority (CMA) has opened its second formal investigation into Microsoft in as many years, but this time the focus isn’t on cloud computing or Activision Blizzard. Instead, regulators are zeroing in on how Microsoft’s AI integration—particularly in Windows and its business applications—could distort competition through interoperability barriers and ecosystem control.

This probe arrives as AI-driven features like Copilot and Microsoft 365’s AI tools become central to enterprise workflows. The CMA’s concern, according to available information, is whether these tools create de facto lock-in by restricting how third-party services interact with Microsoft’s platforms. Early signals suggest the investigation may examine whether businesses face undue pressure to adopt Microsoft’s AI stack—or risk losing compatibility with core productivity tools.

The timing is critical. Microsoft’s AI push coincides with a broader industry shift toward closed-loop ecosystems, where proprietary AI models and data pipelines replace open standards. For regulators, the question isn’t just about fair competition in software, but about whether AI’s rapid adoption is entrenching monopolistic structures under the guise of innovation.

The CMA’s move reflects a growing consensus among antitrust bodies that AI’s competitive risks differ from traditional software markets. Unlike past probes into Microsoft’s bundling practices (e.g., Internet Explorer in the 1990s), today’s concerns hinge on data portability and algorithmic transparency. If confirmed, the investigation could target how Microsoft’s AI models—trained on vast proprietary datasets—interact with rival services, or whether APIs are designed to favor in-house solutions.

What’s notable is the probe’s scope: it spans both consumer-facing Windows features and enterprise tools like Power Platform and Dynamics 365. This dual focus underscores how AI’s role in business operations is blurring the line between productivity software and infrastructure. The CMA’s preliminary findings (where available) will likely scrutinize whether Microsoft’s dominance in office software translates into unfair advantages in AI adoption cycles.

For now, the investigation remains in its early stages, with no formal allegations. But the signal is clear: as AI becomes the default layer for enterprise and consumer tech, regulators are treating interoperability not as a technical detail, but as a cornerstone of market fairness.

Operationally, this investigation may force Microsoft to preemptively open APIs or document data-sharing policies—moves that could ripple across the industry. The alternative? A future where AI’s promise of efficiency comes at the cost of vendor lock-in, with businesses paying the price in flexibility.