The whirlwind management drama at OpenAI last month ended with co-founder Sam Altman reinstated within a week of his surprise dismissal and a much bigger role for Microsoft, which ended up with a seat on the board for the first time since investing billions into the startup earlier this year. That new, cosier relationship is the focus of a new inquiry that was launched today by the Competition and Markets Authority (CMA) in the U.K. over whether the two companies are now effectively in a “relevant merger situation”.
The first part of the process is today’s announcement of the the CMA interest, and an “Invitation to Comment”, which is open to both companies and any interested third parties to provide feedback for the CMA to consider as it eyes up what steps it might take next, if any.
“The invitation to comment is the first part of the CMA’s information gathering process and comes in advance of launching any phase 1 investigation, which would only happen once the CMA has received the information it needs from the partnership parties,” said Sorcha O’Carroll, Senior Director for Mergers at the CMA, in a statement.
“Relevant merger situation” is a specific regulatory term and it is meant to account for situations where a company is not being bought or merged with outright, but the relationship between the two effectively impacts competition for the rest of the market.
The CMA notes that “a range of different kinds of transactions and arrangements may constitute a relevant merger situation,” which can include minority shareholding and commercial arrangements. Both of these, of course, exist in the relationship between Microsoft and OpenAI. Microsoft not only made a huge investment worth billions in OpenAI last year that gives it just under 50% of the business. But the pair work very closely in developing a range of AI services, including a number incorporating Microsoft’s Azure cloud platform.
If AI has been a moving target, it is one that has very prominently featured both Microsoft and OpenAI as two of the strongest and most accurate darts. OpenAI has been setting the pace for the building of LLMs, and of services build using those LLMs. Microsoft has both supported that financially, but also operationally. And Microsoft played a very pivotal role in the last month of turmoil, which appears to have been the latest move that triggered the CMA’s attention.
“There have recently been a number of developments in the governance of OpenAI, some of which involved Microsoft,” it noted today.
Although OpenAI has never fully disclosed what led to the ouster of Altman and his co-founder Greg Brockman, Microsoft wasted no time in offering the pair prominent jobs at their company, along with jobs to any other OpenAI employees who wanted to leave the startup in protest. When Altman and Brockman were reinstated, it was a victory lap for all of them: Microsoft ended up with a board seat for the first time, yes, as a non-voting observer, but still at the table.
“The speed at which artificial intelligence (AI) is scaling across use cases and markets is unrivalled in economic history, while advances in powerful foundation models (FMs) mean that this is a pivotal moment in the development of this transformative technology,” the CMA writes. Essentially, the regulator is worried that in these early days, a handful a companies are making it hard to compete in the building and operation of these foundation models. “The partnership between Microsoft and OpenAI (including a multi-year, multi-billion dollar investment, collaboration in technology development and exclusive provision of cloud services by Microsoft to OpenAI) represents a close, multi-faceted relationship between two firms with significant activities in FMs and related markets,” it added.
There are a number of other criteria to meet to be considered part of a relevant merger situation as defined under the country’s Enterprise Act. If the CMA pursues this as a full-blown investigation, these points will inevitably come up. They will include questions of whether the two businesses — in this case, in the area of AI — are distinct enough; how much revenue is generated through their relationship (there is a £70 million revenue target); and whether they can be argued to account for more than 25% of the market for the product in question. These are all points I imagine both sides would argue exist or do not exist.