Matthew Lesh is a Public Policy Fellow at the Institute of Economic Affairs and a Director at Freshwater Strategy.
In a blast from the past, Britain’s competition regulator, the Competition and Markets Authority (CMA), has proposed designating Microsoft’s software, including Windows, Office, Teams and Copilot, as having “strategic market status”. This is the most substantial use of the CMA’s digital market powers, which have repeatedly prompted concerns about excessive intervention in fast-moving markets and undermining investment into the British economy. The status will also apply to security software, its server operating system and database software.
The focus on Microsoft is intriguing. It almost feels like 1998 all over again, rehashing concerns about market dominance that most thought had long since passed in the internet age, given the fierce competition between the hyperscalers. This time the CMA is proposing to regulate a single company precisely in the part of the economy where competition has never been stronger: AI. The new status lets the CMA impose bespoke conduct requirements on how Microsoft designs and sells its products. This gives the regulator broad discretion to direct how software used by more than a million UK organisations is allowed to operate. How it will use these powers is not yet clear.
The CMA’s complaint is that the products are integrated and therefore have ‘lock-in’ effects among businesses. But CMA-commissioned research found businesses see a “key benefit” in the “relatively good value for money” provided by the bundling and consider it “more cost-effective than buying and managing separate licences”. Put bluntly, people keep using Microsoft software because the products work well together and they’re perceived as good value. This is very far from a market failure that requires intervention.
What’s particularly odd is that Copilot is in scope too, going further than Brussels’ similar regime, which has opted not to include Office or Copilot, and further than the CMA itself went just a year ago, when it opted to keep Gemini out of Google’s designation because AI was moving too fast to call.
It’s hard to overstate how competitive this space already is, and Microsoft is far from leading. Gemini has more than one billion monthly users, and ChatGPT has more than one billion weekly users. In UK workplaces, only about one in ten Office users use Copilot. Gemini, Claude and ChatGPT together already have roughly as many workplace users as Copilot, in just three years or so. Those rivals are also likely making significantly more revenue on these products than Microsoft.
The CMA’s legal test is meant to ask whether Microsoft’s position will be entrenched over the next five years. The productivity software sector is already facing immense disruption in the era of AI-led software development, not to mention the AI tools themselves, which are still developing rapidly. Remember that five years ago, ChatGPT didn’t exist. Anyone confident about who will lead workplace AI in 2031 should be running a hedge fund, not a regulator.
The press release justifying the designation also cites the need for “strategic autonomy”, particularly in the public sector. This is the latest excuse for protectionism, based on the premise that using foreign software providers somehow undermines British sovereignty. While this principle may have some limited justifications in today’s unstable geopolitical climate, it’s hard to see the strategic risk in using Word or Teams. To the extent that this is a legitimate concern, the competition regulator is not meant to settle it through conduct requirements; it’s far beyond its mandate.
There’s also the question of who is pushing for all this. Microsoft’s biggest rivals have spent years lobbying regulators for these sorts of broad interventions. When the DMCC Bill was going through Parliament, I warned in Digital Overload, published by the IEA, that the regime, with its broad, discretionary powers, risks becoming a place for companies to win through regulators what they can’t win from customers. When competitors find it cheaper to lobby the CMA than to build better products and attract customers, the regulator stops refereeing competition and starts picking winners.
Ministers also promised at the time that the regime would be targeted and proportionate. The latest proposal is the clearest test yet of that promise, and it is failing. Instead of targeting an issue like the cloud market, where concerns have focused on software licensing lock-in, the CMA has opted for the widest possible net: combining five separate markets into one ecosystem. As a result, nothing will happen until February 2027 at the earliest, and even then, the conduct requirements aren’t yet specified, which will require further consultations and delay action well into 2027. This is hardly swift and targeted action; it is broad and slow.
Britain wants to be the best place in the world to build and adopt AI. Pre-emptively regulating the AI tools businesses are only just starting to use is a strange way to get there.
I would not say that users 'perceive Microsoft products as good value. It is more a case of it is the market leader, and streets ahead of any competition. Microsoft are 'charging what the market will bear' and recent years their annual charges have increased with little concern that users will 'go elsewhere'. It is far too late for any market authority to intervene now. I do agree though that their newer offerings, like Copilot, are meeting customer resistance. We have all learned, to our cost, what happens when we get suckered in, and today's savvy customer will deliberately look to alternatives to prevent the grasp of the mighty Microsoft growing even tighter.
Ps. Teams is nowhere near as intuatively friendly as Zoom.