Google/Anthropic: the latest chapter in the CMA’s scrutiny of generative AI

Generative AI has recently become a policy and enforcement priority for competition authorities across the globe¹. In our previous post we looked at the European Commission’s September 2024 Policy Brief on generative AI and virtual worlds. The CMA has been examining similar issues since it launched a review of AI foundation models (FMs) in May 2023 (see here).

The CMA published an update paper and a technical update report on FMs in April this year. In these papers, the UK regulator identifies various risks to competition in AI-related markets, including restriction of access to critical inputs for developing FMs and the proliferation of partnerships between key industry players that might reinforce or extend existing positions of market power through the value chain.

While the CMA acknowledges that strategic partnerships can give rise to pro-competitive benefits such as increased innovation, it is also “vigilant against the possibility that incumbent firms may try to use partnerships and investments to quash competitive threats”². This echoes comments made by EU Commission Executive Vice-President Margrethe Vestager in a June 2024 speech:

“[…] We are seeing a trend of big companies setting up partnerships with small AI developers. It is becoming a feature of the industry. Actually, these investments are important: they give access to the necessary components and allow AI systems to be developed. So generally, we consider these deals to be pro-competitive. But they can sometimes create entrenched market positions, especially through exclusivity rights. So we need to keep an eye on them to ensure fair play.”

As part of its review of AI foundation models, the CMA has been tracking partnerships across the FM value chain, identifying over 90 partnerships between Google, Amazon, Microsoft, Meta, Apple and Nvidia on the one hand and FM developers on the other³. The UK regulator has subjected certain partnerships to particular scrutiny in the last year. It invited comments on Microsoft’s partnership with OpenAI last December, and this case remains open⁴. In May 2024 it decided that Microsoft’s partnership with Mistral AI did not qualify for investigation under the UK merger rules⁵;  and it cleared Microsoft’s hiring of former employees of Inflection AI in September 2024⁶. That same month, the CMA decided that Amazon’s partnership with Anthropic did not qualify for investigation under the UK merger rules⁷.

Anthropic is once again in the CMA’s focus, this time in relation to its partnership with Alphabet Inc (Google). On 24 October the CMA notified the parties of its merger inquiry into the partnership, following reports that Google had invested around $2 billion in Anthropic. Anthropic is a US-based company specialising in AI safety and research and the developer of the Claude family of FMs. It uses Google Cloud as its cloud provider⁸.

According to a Reuters report, an Anthropic spokesperson has confirmed the company’s intention to cooperate fully with the CMA and provide a “complete picture” of its partnership with Google. The spokesperson added: “We are an independent company and none of our strategic partnerships or investor relationships diminish the independence of our corporate governance or our freedom to partner with others.”

The deadline for the CMA’s Phase 1 decision on the Google/Anthropic partnership is 19 December. At that point the CMA will have to decide whether to refer the partnership to a full Phase 2 investigation. In this context, the following extract of the CMA’s April 2024 update paper on AI foundation models is worth highlighting:

“We are […] stepping up our use of merger control to examine whether [strategic partnerships and investments] fall within the current rules and, if so, whether they give rise to competition concerns. It may be that some arrangements falling outside the rules are problematic even if not ultimately remediable through merger control. Equally some arrangements may not give rise to competition concerns. We consider it is appropriate to step up our review more generally so that we can start to identify more clearly and coherently which types of partnerships fall within the merger rules and the circumstances in which they may give rise to competition concerns. That will also be of benefit to the businesses involved.”

It remains to be seen whether Google/Anthropic will proceed to a Phase 2 investigation. One thing seems clear though: further CMA investigations into strategic partnerships in the generative AI space are yet to come.

Follow our Spotlight on – competition and generative AI series for further updates.

==============================

1. See e.g. the July 2024 Joint Statement of the US FTC and DoJ, the CMA and the European Commission on Generative AI Foundation Models and AI Products; the JFTC’s October 2024 Discussion Paper on Generative AI and Competition; and the Australian Digital Platform Regulators Forum August 2024 Working Paper on Multimodal Foundation Models.  
2. CMA technical update report on AI foundation models (16 April 2024), para 2.60. See also paras 2.54 - 2.59 and 5.16 - 5.19. 
3. CMA technical update report on AI foundation models, para 2.58. 
4. See https://www.gov.uk/cma-cases/microsoft-slash-openai-partnership-merger-inquiry.
5. See https://www.gov.uk/cma-cases/microsoft-slash-mistral-ai-partnership-merger-inquiry. 
6. See https://www.gov.uk/cma-cases/microsoft-slash-inflection-ai-inquiry. 
7. See https://www.gov.uk/cma-cases/amazon-slash-anthropic-partnership-merger-inquiry.
8. Anthropic press release, “Anthropic Partners with Google Cloud”, 3 February 2023, https://www.anthropic.com/news/anthropic-partners-with-google-cloud.