Nate Russo

The CMA Probe Won't Fix Microsoft's Grip - But It'll Make Them Nervous

April 1, 2026

On March 31st, 2026 - literally yesterday - the UK's Competition and Markets Authority announced it is launching a Strategic Market Status investigation into Microsoft's business software ecosystem. We're talking Windows, Word, Excel, Teams, and Copilot. The investigation kicks off in May and could take up to nine months. The CMA is going to examine whether Microsoft's licensing practices are strangling competition in the cloud and locking UK businesses into a tech stack they can't escape without bleeding money.

I want to be clear upfront: I think this matters. I think it's probably the right call. And I also think it won't change a single thing about how deeply embedded Microsoft is in the daily operations of most UK businesses. Both of those things are true at the same time.

That's not defeatism. That's just reading the scoreboard.

What's Actually Happening Here

The CMA is preparing to launch a new antitrust investigation into Microsoft over its business software and cloud services. The probe will focus on whether Microsoft is using its strong position in products like office productivity tools and Azure cloud to make it harder for companies to switch to rival providers. Regulators are examining licensing rules, pricing structures, and bundled services that could be limiting competition.

The concerns center on products such as Windows Server and SQL Server being more expensive to run on certain rival clouds, notably Google Cloud and AWS, as Microsoft permits customers to reallocate on-prem licenses to cover cloud-based instances running on Azure. Google in particular claimed that it costs up to five times as much to move legacy workloads to a non-Microsoft cloud, serving as a major incentive for businesses to choose Azure instead.

Five times. Let that land. You're not choosing Azure because it's technically superior in your situation. You're choosing it because your existing Microsoft licenses make everything else prohibitively expensive. That's not competition. That's a toll booth.

The products under review - Windows, Word, Excel, Teams and Copilot - are not random inclusions. They form the front door for many UK organisations into Microsoft's wider platform, and that makes the investigation about more than one market definition. It is about whether Microsoft can bundle, steer, or condition access in ways that make rivals less viable, even where customers think they are making independent purchasing choices.

And that last clause is the real thing. The part about customers thinking they're making independent choices. This is the architecture of lock-in. It doesn't look like coercion. It looks like convenience. Then one day you try to leave and you find out the exit costs five times what staying does.

This Isn't the First Swing

Here's the part that should make you understand why I'm skeptical that this investigation changes the real-world outcome, even while I think launching it is the right call.

In the eyes of an inexperienced optimist, 2025 was supposed to be a year of reckoning for Microsoft's licensing practices. Regulators across three continents had Microsoft in their sights. The European Commission was investigating Teams bundling. The UK Competition and Markets Authority was probing the entire cloud infrastructure market. European cloud providers had lodged formal complaints about anti-competitive licensing. The US Federal Trade Commission had launched its most comprehensive Microsoft investigation since the landmark 1990s case. And a coalition of nearly 60,000 UK businesses was preparing to sue. By the end of the year, Microsoft had emerged largely unscathed.

The Teams investigation ended in a settlement with no fine. The CISPE complaint was withdrawn after a cash payment. The CMA recommended further investigations rather than taking action itself. The FTC probe continued quietly with no public developments since March.

That's a remarkable track record for a company supposedly under siege from regulators. They paid some money to make the CISPE crowd go away. They agreed to "address concerns" on cloud egress. The Teams thing evaporated. And now we're starting the next investigation.

Cal asked me yesterday why I care so much about this. He was eating a sad desk salad and I could tell the question was more about him than me. But I thought about it. I care because I've watched this exact pattern play out in the software industry for years - a slow accumulation of power, a cascade of regulatory concern, and then... nothing structurally changes. We wrote about this dynamic with the software industry's general anxiety response to competitive pressure, and Microsoft is basically the platonic ideal of that piece.

The Numbers Tell The Story The Regulators Are Slowly Realizing

Productivity software leadership remains unmatched, with 87.5% global market share exceeding Google Workspace by a wide margin. Cloud infrastructure ranks second globally near 20% share, trailing AWS at approximately 30% while maintaining a lead over Google Cloud near 13%.

FY2025 revenue reached $281.7 billion, reflecting 15% growth year over year. Operating income totaled $128.5 billion, supported by scale efficiencies despite elevated infrastructure spending. Azure AI services expanded 40% year over year in Q1 FY2026, driven by enterprise deployment of generative AI workloads. Microsoft Cloud revenue reached $49.1 billion in Q1 FY2026, growing 26% annually.

This is a company that controls 87.5% of the global productivity software market and is still growing Azure at 40% year-over-year on AI workloads. The market cap cleared $4 trillion. And they just announced price hikes that will hit hundreds of thousands of businesses in July.

Microsoft 365 Business Basic increases from $6 to $7 per user per month, representing a 16.7 percent increase. Microsoft 365 Business Standard moves from $12.50 to $14 per user per month, a 12 percent increase.

That timing - announcing price increases in December 2025 for July 2026, right as regulators are gearing up a new probe - is not an accident. Microsoft said it planned to raise commercial pricing for its flagship subscriptions, upping costs by an average of 16% starting in July 2026 as it includes access to more AI tools. The justification is Copilot. You're getting AI, so you pay more. Whether you want the AI or not.

This is the core of what the CMA is correctly pointing at. The CMA's concern is not simply that Microsoft is big. It is that market power in one layer can spill into adjacent layers, especially when licensing terms or bundle structures make it harder to mix products from different suppliers. The cloud investigation already flagged concerns that software licensing could distort cloud competition, and now the regulator appears ready to ask whether the same practices also affect the business software market more broadly.

I think about this like the scene in The Last Jedi where the Resistance fleet is just slowly bleeding out in space - outgunned, outmaneuvered, and being forced to burn fuel they can't replace. Everyone watching is screaming about strategy, but the structural situation was baked in long before the battle started. The CMA probe is, at best, a hyperspace jump - dramatic, feels decisive, might buy some time. It doesn't fix the underlying power imbalance.

The SMS Designation: The Tool That Could Actually Do Something

Here's what makes this probe different from the ones that fizzled in 2025. The CMA now has a real weapon it didn't have before.

The Digital Markets, Competition and Consumers Act 2024 came into force on 1 January 2025 and allows the watchdog to designate firms with strategic market status, or SMS, in relation to a particular digital activity. Once designated, the CMA can impose conduct requirements or pro-competition interventions designed to open markets and address entrenched dominance. The regulator used that framework earlier this year to launch SMS investigations into Apple and Google's mobile ecosystems, showing that it intends to use the regime aggressively and in phases rather than waiting for every market to finish a traditional inquiry.

An SMS designation is not a fine. It's not a settlement. It's legally binding conduct requirements. It's the CMA being able to say: you will change how you license this software, you will lower these switching costs, you will make it possible for customers to choose AWS or Google Cloud without paying a 5x penalty. That's structurally different from anything that's landed on Microsoft before in the UK.

The CMA seeks to determine whether the company's use of software licensing hurts competition in cloud as concerns rise about AI adoption entrenching customers with the technology provider. If the investigation leads to a strategic market status designation, it would let the CMA address Microsoft's licensing practices and level the playing field as AI is embedded in business software tools.

The AI piece is crucial. This isn't just about cloud egress fees anymore. The watchdog said the rapid embedding of artificial intelligence, including agentic technologies, into workplace tools, with products such as Microsoft Copilot, Enterprise GPT and Claude Enterprise already becoming commonplace, made this a pivotal moment, with implications for UK productivity and competitiveness.

The CMA is doing something smart here: getting ahead of the next wave of lock-in before it fully sets. Once Copilot is the glue holding your entire workflow together - once it knows your inbox, your calendar, your Excel models, your Teams history - leaving Microsoft won't just be expensive. It'll feel cognitively impossible. The AI becomes institutional knowledge storage. That's a moat that makes egress fees look quaint.

We talked about this exact dynamic in the piece about AI embedding itself in software you already use. It applies here with ten times the force, because Microsoft's distribution is ten times larger than anyone else's.

But Let's Not Kid Ourselves About the Timeline

The strategic market status investigation into Microsoft's business software ecosystem is due to start in May, and could take up to nine months to complete. And that's just the investigation phase. Getting to an actual SMS designation, then imposing conduct requirements, then watching Microsoft's legal team contest every single intervention - we're talking years before anything changes operationally for a UK business running Microsoft 365.

Meanwhile, AI integration across productivity, cloud, and developer platforms has strengthened customer lock-in and expanded average revenue per user. Every month the investigation drags, Microsoft's position compounds. The lock-in deepens. More workflows wrap around Copilot. More data lives in Azure. More procurement decisions get made by IT teams who've standardized on the Microsoft stack and can't remember why they ever considered alternatives.

A key driver of the SMS decision was the CMA's finding that Microsoft had made no material progress on licensing concerns since its cloud market investigation concluded in July 2025. No material progress. After a full investigation, commitments were made, and the needle didn't move. That tells you something about the gap between regulatory intentions and Microsoft's actual behavior change calculus.

There's also a weird asymmetry in the new probe. The decision to exclude AWS raises practical concerns with both providers being structured in the same way from a lock-in perspective, which could create a regulatory imbalance between the two parties that would keep one side unchallenged. AWS gets a pass while Microsoft gets the full SMS treatment. That's not inherently wrong - Microsoft's productivity software dominance is a different animal than AWS's cloud infrastructure play - but it creates an odd dynamic where the investigation framework itself might have blind spots.

Helen came by while I was reading through all the CMA documentation and mentioned that Harold spent three hours on the phone with their Microsoft account rep last year trying to figure out why their renewal was so much higher. He never got a straight answer. He gave up. That's not anecdote - that's the lived experience of this market. The complexity is a feature, not a bug.

What This Actually Means If You Run a Business

The CMA probe won't save you. Not by July. Probably not by 2027. But it matters for a few real reasons that go beyond regulatory theater.

First, it signals that the SMS framework is going to be used aggressively. Microsoft is now the fourth SMS investigation alongside Apple and Google's mobile ecosystems. The probe places Britain alongside other jurisdictions scrutinising Microsoft. Brazil's CADE has opened a parallel investigation into Microsoft's corporate software and cloud conduct, while Japan's JFTC is examining whether Microsoft Azure restricts customers and rivals from combining services across providers. This isn't a single regulator being frisky. This is a coordinated global recognition that the current structure of the Microsoft ecosystem is not compatible with open markets.

Second, it changes Microsoft's negotiating posture with large customers. When you're under a formal SMS investigation, the calculus on "how aggressive can we be with licensing terms" shifts. Not dramatically. But at the margin. Brad Smith already responded publicly saying Microsoft is "committed to working quickly and constructively" with the CMA to address its concerns. That's boilerplate, but it's also a tell. They're paying attention.

Third - and this is the one I think gets underreported - it gives competitors oxygen. "The UK will benefit most where a broad range of competitors can integrate with Microsoft's business software, so that businesses and public sector organizations can mix-and-match AI software across suppliers to best suit their needs." The CMA isn't just trying to constrain Microsoft. It's explicitly trying to create space for the alternatives to breathe. That matters for anyone building on top of non-Microsoft stacks, or for anyone advising clients on software strategy. The regulatory wind is behind diversification now in a way it wasn't two years ago.

I've been using tools that slot alongside the Microsoft stack rather than competing with it directly, and you can see the difference when there's genuine interoperability versus when a vendor is grudgingly tolerating you. The distinction matters enormously if you're trying to run automations that touch multiple platforms. If you want to understand how this plays out in vendor positioning more broadly, the piece we did on vendors and their AI threat responses gives some useful context - because the companies that answered honestly were the ones already thinking about interoperability as their survival mechanism.

My Actual Take

The CMA probe is the right call, launched too late, with tools that are better than anything the regulator has had before, against a company that has become structurally too embedded to move quickly no matter what the finding is.

That's not a condemnation of the CMA. That's just an accurate description of the problem. For many businesses, Windows remains the endpoint operating system, Microsoft 365 the productivity layer, Teams the collaboration tool, and Copilot the new AI-inflected interface tying the bundle together. That combination gives the company a structural advantage because the user experience, identity layer, admin tools and procurement channels can all be integrated in ways that feel seamless to customers but difficult for rivals to break into.

Microsoft will cooperate, selectively. They'll make some commitments on egress fees - they already started. They'll talk a lot about how competitive the market is. Brad Smith will write thoughtful blog posts. And unless regulators require quick changes from Microsoft, the company "will continue to further entrench its dominance in AI just as it has in software and cloud." That quote is from the Coalition for Fair Software Licensing and it's blunt, but it's right.

The investigation will make Microsoft nervous. Nervous enough to make some visible concessions. Nervous enough to be more careful about the most egregious licensing games. Not nervous enough to actually loosen the grip.

Rey didn't fix everything in the sequel trilogy either. She moved the thing forward. Set up the next thing. That's what the CMA is doing here. I keep telling people The Last Jedi is the most ambitious Star Wars film ever made precisely because it understood that systemic problems don't resolve cleanly - they escalate, they transform, and the heroes rarely get the clean win they wanted. I get the look. I don't care.

For what it's worth: key issues such as interoperability and egress fees - which place high costs on customers transferring data to other services - were highlighted in the probe. Last year, the CMA concluded the duo's dominance actively harmed competition in the country's cloud computing market. The conclusion was reached. The harm was named. Now the question is whether the new tools in the DMCC Act have enough teeth to actually do something about it.

I think they do. I think the SMS designation, if achieved, is genuinely significant. I think it'll take long enough that most businesses won't notice a material change for years. And I think Microsoft's July 2026 price increases will land before any of this changes anything on the ground - which is its own kind of message about where the power actually sits right now.

Watch the SMS investigation. Watch whether Microsoft actually moves on licensing or just makes noises. And if you're running a business that's entirely on the Microsoft stack and you've never seriously stress-tested what leaving would cost you - the answer to that question is worth knowing, regardless of what the CMA finds.