CMA clears Google over Anthropic partnership

The Competition and Markets Authority (CMA) has said Alphabet’s partnership with Anthropic does not qualify for investigation under the merger provisions of the Enterprise Act 2002.

In October 2023, Alphabet invested $2bn in OpenAI rival Anthropic. The artificial intelligence (AI) startup has also received $4bn funding from Amazon.

The CMA is concerned that the foundational model sector is developing in ways that risk negative market outcomes. In particular, the likes of Google, Amazon, Meta, Microsoft and Apple have the market dominance to buy up or shut down competition. It is also worried that partnerships between these major technology providers and developers of AI foundation models may limit choice and be anti-competitive.

In September, the CMA concluded its investigation of Microsoft’s hiring of key staff from Inflection, finding that Inflection AI was not a strong competitor to the consumer chatbots Microsoft has developed directly in partnership with OpenAI.

Discussing the outcome of the latest investigation, Joel Bamford, executive director of the CMA, wrote on LinkedIn: “Our investigation has shown that Google has not acquired the ability to materially influence Anthropic’s commercial policy and therefore the partnership does not meet the jurisdictional threshold for UK merger control to apply.”

He described the conclusion of this latest investigation as “another decision by the CMA which provides greater clarity for businesses and their investors”.

In a summary of its findings from the phase one investigation into the deal, the CMA said it did not believe Google had acquired material influence over Anthropic as a result of the partnership. The CMA said it looked at the risk of Google exercising influence over Anthropic at shareholder and/or board level, along with an assessment of Google’s own Vertex AI product.

“The available evidence did not indicate that Google has the ability to exercise material influence over Anthropic through the partnership,” the CMA concluded.

The CMA said it had considered the fact that Anthropic and Google offer two of the leading foundational AI models globally. However, given Anthropic’s turnover is below the £70m threshold, which is one of the criteria it takes into account when assessing whether to look further into a deal, pursuing this thread of investigation was not necessary.

The CMA is also looking at whether it should investigate Amazon’s partnership with Anthropic, due to the $4bn funding the AI startup received from Amazon. 

Some industry experts believe the CMA should continue looking at the foundation model market. Josh Mesout, chief innovation officer at Civo, said: “While the CMA has decided not to pursue an investigation into the Anthropic/Alphabet partnership, the broader concerns raised in the investigation about potential market concentration in AI remain valid.

“Over-dependence on a handful of major firms could still stifle innovation, limit consumer choice and potentially lead to a monopoly that favours Big Tech. Even without a formal investigation, it is the responsibility of everyone in the industry to ensure the AI market remains fair, competitive and conducive to ongoing technological advancement.”