UK competition watchdog steps up scrutiny of big tech’s role in AI startups
The UK competition watchdog has stepped up its scrutiny of big tech involvement in artificial intelligence startups, asking for comment on three deals by Microsoft and Amazon.
The Competition and Markets Authority (CMA) announced that it was examining Microsoft’s investment in the French firm Mistral and the hiring of the DeepMind co-founder Mustafa Suleyman as head of the US company’s new AI division. The watchdog is also scrutinising Amazon’s $4bn (£3.2bn) investment in the US AI firm Anthropic.
The CMA has issued “invitations to comment” on the tie-ups, a procedural move that paves the way for a formal investigation, amid concerns that these partnerships are effectively giving the big tech companies backdoor control over potential rivals and stifling competition. The watchdog has already asked for comments on Microsoft’s relationship with OpenAI, the developer of ChatGPT.
Joel Bamford, the executive director of mergers at the CMA, said: “We will assess, objectively and impartially, whether each of these three deals fall within UK merger rules and, if they do, whether they have any impact on competition in the UK.”
Meanwhile, Sarah Cardell, the chief executive of the CMA, has warned that the watchdog had “real concerns” about the AI market. The CMA has singled out six tech companies at the heart of an “interconnected web” of AI partnerships: Google; Microsoft; Meta; Amazon; Apple; and Nvidia, the leading supplier of chips for training and operating AI systems.
The CMA’s next step with Microsoft and Amazon will be a “phase one” investigation, during which the watchdog will examine whether the partnerships fall under the UK merger regime and whether they raise competition concerns. If the CMA finds there are concerns and decides to proceed with a phase two investigation, it could seek remedies from the companies involved.
In February, Microsoft announced it was investing €15m (£13m) in Mistral, a Paris-based startup specialising in open-source AI models, which can be freely downloaded and adapted by users. Microsoft said the deal represented an opportunity for Mistral to “unlock new commercial opportunities” and “expand to global markets”.
The Mistral deal was followed weeks later by the announcement of Suleyman’s appointment. Several employees at his Inflection AI startup were also recruited to join the new Microsoft division. The tie-up included a $650m cash payment to Inflection and making the company’s AI models available on Microsoft’s Azure cloud service.
Amazon’s deal with Anthropic, the developer of the Claude chatbot, involves the US startup using the big tech company as its main cloud provider and using Amazon’s custom chips to build, train and deploy AI models.
Alex Haffner, a partner at the UK law firm Fladgate, said the move underlined the CMA’s interest in the AI market. “Noises coming out of the CMA
in recent times have indicated that they are very keen to get involved early in the evolution of the AI market and today’s announcement certainly reflects that degree of seriousness,” he said.
Microsoft said it would provide the CMA with the information the watchdog needed to carry out its inquiries. “We remain confident that common business practices such as the hiring of talent or making a fractional investment in an AI startup promote competition and are not the same as a merger,” the company said.
Amazon said the CMA move was “unprecedented” and that the Anthropic deal was different from other tieups announced by the big tech firms. In a pointed reference to Microsoft’s partnership with OpenAI, Amazon said the deal did not give it a “board director or observer role” and continued to have “Anthropic running its models on multiple cloud providers”.