Britain’s competition regulator, the Competition and Markets Authority (CMA), said on Wednesday that Microsoft’s partnership with OpenAI doesn’t qualify for an investigation under the merger provisions of the U.K.’s Enterprise Act 2002, the country’s anticompetitive practices law.
“Overall, taking into account all of the available evidence […] the CMA does not believe that Microsoft currently controls OpenAI’s commercial policy, and instead exerts a high level of material influence over that policy,” the CMA wrote in its decision. “In other words there is no change of control giving rise to a relevant merger situation.”
The CMA began investigating Microsoft’s partnership with OpenAI in December 2023. The tech giant is one of the AI company’s top investors, with nearly $14 billion invested since 2019. Microsoft also packages many of OpenAI’s technologies in a managed offering called the Azure OpenAI Service, and it works closely with the outfit to develop products like its Copilot chatbot and GitHub Copilot AI coding assistant.
The CMA was initially concerned that Microsoft had acquired control over OpenAI’s commercial policy in 2019, and that this control increased following Microsoft’s role in securing OpenAI CEO Sam Altman’s re-appointment in November 2023. The CMA believed such control could result in a “substantial lessening” of AI industry competition in the U.K.
“[A]n increase in Microsoft’s control over OpenAI could give rise to potential competition concerns if Microsoft was able to restrict rivals’ access to OpenAI’s leading models in markets where access to [AI] is likely to be important and where Microsoft already holds strong market positions,” the agency wrote in a filing. “The CMA was also concerned that the partnership could potentially impact competition in the emerging market for the supply of accelerated compute, given OpenAI’s potential to act as an important customer in this market.”
But as the CMA noted in its decision on Wednesday, recent developments have potentially weakened — not strengthened — Microsoft’s influence over OpenAI.
In January, Microsoft said it had renegotiated elements of its cloud computing agreement with OpenAI, moving to a model where the tech giant has the “first right of refusal” for certain OpenAI workloads. Microsoft has also granted waivers to let OpenAI build additional computing capacity, including a $500 billion U.S. data center deal with investor SoftBank. Previously, Microsoft was OpenAI’s exclusive cloud provider.
Last year, Microsoft also dropped plans to take a board seat at OpenAI, which might have invited additional regulatory scrutiny.
“Material aspects of the [Microsoft] partnership have been changing over the course of the investigation,” the CMA wrote in its decision. “Furthermore, there is no ‘bright line’ between factors which might give rise to material influence and those giving rise to de facto control.”
The CMA has aggressively investigated the tech industry’s investments in AI and AI startups in an effort to prevent a consolidation of power in the nascent sector. However, while the agency has previously characterized big tech partnerships with AI startups as an “interconnected web,” it hasn’t found evidence of wrongdoing. In November, the CMA cleared Google parent company Alphabet’s dealings with OpenAI rival Anthropic.