OPINION
One might think that if a government sacks the chairman of an agency after controversial decisions that were inconducive to the country’s economic growth, the message would be easily understood by those who were lucky to keep their jobs. Today there is evidence that even such a drastic measure can fall on deaf ears:
I wrote in May that the UK Competition & Markets Authority (CMA) would have to choose between Google and its government when issuing a decision at the end of its cloud market investigation. The summary of that decision (PDF) has just been released and — guess what — the CMA opted for Google all the way. It is the most obviously arbitrary decision by an antitrust authority I’ve ever seen in a scenario where several companies (here, Amazon, Microsoft and Google) are similarly situated. No law- and self-respecting regulatory would let one of them (Google) off the hook when all three have a lot of market share, are growing, and have unique offerings of one kind or another. The CMA is a UK government department, but the cloud market report makes it look like an Alphabet subsidiary.
This is the same CMA that tried to block a clearly procompetitive merger (Microsoft & Activision Blizzard) that has since brought many advantages to consumers (October 13, 2024 games fray article). Google opposed that deal, too, though Sony was more vocal.
As if its former activist chair Marcus Bokkerink had not been fired earlier this year, the CMA’s cloud market decision makes mockery of the UK government’s strategic steer. “Using CMA tools proportionately, with growth and investment in mind” means something other than wasting resources on a market that has absolutely no UK-specific characteristics, in which there are many players, and in which prices have been falling for a long time. And it’s not just that they have already wasted resources by picking up an investigation that was first conducted by telecommunications regulator Ofcom: they now want to keep throwing good money after bad, referring the matter to a recently-created CMA division, the Digital Markets Unit (DMU).
Whoever came up with the idea that they should look at the same non-issue (that even if there were issues would not be a logical priority for the UK) for a third time has an agenda that is not aligned with the priorities the UK government identified.
Maybe some people already have a premonition. The next milestone is a January 2026 decision by the CMA Board on Strategic Market Status (SMS) designations, which means targets of potential investigation by the DMU. After voicing in the penultimate paragraph of the summary the “expect[ation] that [their] findings and recommendations will be taken into account as part of [the SMS] decision,” they “recognise that there is some uncertainty around the implementation of {their] recommendations.”
The headline of the document says “decisions.” But the final paragraph downgrades it all to “recommendations.” It’s a decision not to dispose of this file for good, in light of overwhelming evidence that the cloud market does not intervention, and the intervention suggested by Google would, of course, benefit Google, but not customers.
There still is hope that the DMU will agree that the cloud market is not a priority for competition enforcement, much less for UK competition enforcement. Even though the DMU is a CMA division as opposed to agency of its own, it is neither legally nor politically-psychologically bound by a report that makes no sense. The DMU can’t declare too many digital products and services a strategic priority: it has resource constraints, and there are gatekeeper issues especially in mobile ecosystems that require a lot of resources because resistance by the deep-pocketed incumbents is as stiff as it is massive.
The DMU can’t turn a blind eye to the AI dimension of this. Google has two perspectives on AI and the cloud business, and the CMA should pick the right one. When Google talks to investors, the cloud business is now all about AI, but when it plays a regulatory capture game with the CMA, it says the opposite (April 25, 2025 ai fray article). The summary of the CMA decision adopts the denial version (“AI capabilities currently play a limited role in driving customers’ choice of cloud provider”), but goes on to nuance it: “However, these capabilities are likely to become more important to customers over time and so AI could impact competition in cloud services to a greater degree in the future.”
Come January 2026, the evidence should be even stronger that AI is not just a vision for the future, but a driver of demand for cloud services in the present. The market investigation that closed today started in October 2023, and picked up the thread from an earlier Ofcom investigation. The extent to which AI drives demand for cloud services changed fundamentally during that period, while legacy business software licenses play ever less of a role in cloud infrastructure decisions.
For now, only the summary of the final decision is available. I will read the detailed decision with interest (aAbove all, the part on AI), and publish a summary and some observations, but in the form of an article while this was just an initial reaction as an opinion piece.