Good news for Microsoft: UK antitrust regulators have announced that they won't launch a full investigation into the tech giant's acquisition of the team behind AI startup Inflexion because it doesn't raise competition concerns.
However, the Competition and Markets Authority (CMA) said the deal fell within its regulatory jurisdiction as a “relevant merger situation”, meaning similar transactions in the future could be investigated on competition grounds even if an outright acquisition does not take place.
“Quasi-merger”
Microsoft launched a new consumer AI division in March, spearheaded by Inflection founders including Google DeepMind co-founder Mustafa Suleiman. CEO Satya Nadella confirmed that many other Inflection AI members have joined Microsoft's new AI division, including AI scientist Jordan Hoffman, who now leads Microsoft's UK AI hub in London.
In July, the CMA announced it was opening a “phase one” merger investigation into the deal, beginning its investigation to gather evidence and determine whether to proceed to a full investigation. The first phase centred on whether the deal could in fact be investigated as a “merger”, given that Microsoft had not actually acquired Inflection AI, and, if so, whether further steps were needed to address competition concerns.
At the heart of this is a growing effort by big tech companies to circumvent regulatory scrutiny of AI, fueled by a new form of M&A some call “quasi-mergers,” which can involve anything from strategic investments to hiring startup founders and technical talent, as in the case of Microsoft and Inflexion.
In its findings, the CMA noted that Microsoft employed “substantially the entire Inflection team”, including the two founders, and at the same time entered into a series of commercial agreements, including non-exclusive licence agreements to use Inflection's intellectual property (IP).
Additionally, the CMA concluded that the Inflection team's self-described mission prior to their recruitment was “creating AI for everyone,” but that with most of the team leaving Inflection, Microsoft “gained the team's collective know-how” and began developing its own AI technology. The CMA stated in its evaluation:
“Given that technologies in this field can quickly become outdated without continued development, CMA
Point out the importance of expertise in the development and supply of FMs [foundation models] and chatbots. Based on the evidence seen by the CMA, a team of staff responsible for development would be central to any company seeking to develop FMs or chatbots. In this context, the CMA considers that the acquisition of a team with relevant know-how may fall within the CMA's merger control jurisdiction even in the absence of any additional assets.”
One of the key tests used in the judgment comes from the UK Enterprise Act 2002 and helps define what a “relevant merger situation” means in practice, including whether “two or more enterprises become indistinguishable” as a result of the transaction.
While Inflection still exists as an independent company, the core of the company is now at Microsoft.
“Following staff transfers and other tactical arrangements, the two firms are no longer separate,” CMA executive director Joel Bamford said in a LinkedIn post today.
Despite all this, the second aspect of the CMA's investigation concluded that the deal did not raise significant competition concerns because Inflection was not a sufficient competitor to Microsoft's existing AI tools.
“Inflection AI is not a strong competitor to the consumer chatbots developed directly by Microsoft (Copilot) and in partnership with OpenAI (ChatGPT),” Bamford wrote. “For this reason, we approved the transaction.”
Regulatory Oversight
The Microsoft-Inflexion investigation was one of several similar investigations announced by the CMA in April, which also included Microsoft's investment in French start-up Mistral AI, though the regulator quickly concluded that the deal was not subject to investigation under current merger regulations due to the size of the investment. At the same time, the CMA revealed it was investigating Amazon's relationship with Anthropik, following its $4 billion investment in AI start-up Anthropik, which is now the subject of a formal investigation.
Meanwhile, the CMA is seeking comment from parties and interested parties about Google's relationship with Anthropic after the internet giant invested $300 million in it last year and a further $2 billion since then.
So the main takeaway from this case is that while Microsoft may have escaped regulatory intervention this time, future deals involving large tech companies and smaller start-ups are subject to the CMA's scrutiny, whether or not an outright acquisition occurs.