CMA launches full blown probe of Amazon's Anthropic tie-up

Britain's competition regulator is embarking on a full blown deep dive into Amazon's multi-billion dollar investments in Anthropic to ascertain if the exchange equates to a stealthy "merger situation."

The Competition and Markets Authority invited interested third parties in April to provide input on the potential impact to their business of Amazon injecting $1.25 billion into Anthropic in September 2023, which was followed by $4 billion funding in March.

The watchdog says [PDF] it now has "sufficient information in relation to the partnership between Amazon.com and Anthropic PBC" to launch an probe "for the purposes of deciding whether to make a reference for a phase 2 investigation" – a stage that would itself necessitate a even deeper inspection of the relationship.

The decision deadline for that stage is scheduled for October 4, 2024.

The ultimate aim is to discover if the investment "resulted in the creation of a relevant merger situation" under the provisions of the Enterprise Act 2002, and if so, could this result in a "substantial lessening of competition within any market or markets" within the UK for goods or services.

In the burgeoning market for LLMs, regulators are wary of cloud giants Amazon Web Services, Microsoft, and Google using their financial muscle to control fledgling organizations and tighten their grip on the tech.

Amazon's $5.25 billion investment in the AI upstart includes Anthropic agreeing to purchase cloud resources on AWS and a non-exclusive accord to host its models of Amazon Bedrock, its platform for building GenAI apps.

The Jeff Bezos empire previously told The Register in the spring that the CMA's review of a "collaboration of this type" was "unprecedented."

Today, a spokesperson at the megacorp said in a statement emailed to us: "We're disappointed that the UK's Competition and Markets Authority (CMA) has not ended its probe yet.

"Amazon's collaboration with Anthropic does not raise any competition concerns or meet the CMA's own threshold for review. The early days of generative AI have largely seen one successful option available for customers. Anthropic has worked hard to become an emerging viable alternative.

"But, building models is expensive, and companies like Anthropic need access to a substantial amount of capital to train these models. By investing in Anthropic, Amazon, along with other companies, is helping Anthropic expand choice and competition in this important technology. Amazon holds no board seat nor decision-making power at Anthropic, and Anthropic is free to work with any other provider (and indeed has multiple partners)."

The bit about the board seat, who could Amazon be referring to? Likely Microsoft's cosy relationship with OpenAI, although Microsoft quit its board seat in July when regulators began taking a closer look.

The spokesperson added: "Amazon will also continue to make these Anthropic models available to customers via Amazon Bedrock, a service that makes it easier for developers and companies to leverage large language models (LLMs) and build generative AI applications."

Anthropic sent a statement to The Register. "We are an independent company. Our strategic partnerships and investor relationships do not diminish our corporate governance independence or our freedom to partner with others... We intend to cooperate with the CMA and provide them with a comprehensive understanding of Amazon's investment and our commercial collaboration."

The alliance between the pair has already caught the attention of the Federal Trade Commission in the US, and the American regulator is also inquiring about Google's investment in Anthropic and Microsoft's investment in OpenAI in the same probe.

4 stories

• OpenAI meltdown: How could Microsoft have let this happen after betting so many billions?
• Microsoft exits OpenAI's boardroom to sidestep regulatory scrutiny
• OpenAI meltdown: How could Microsoft have let this happen after betting so many billions?
• Microsoft hiring Inflection team triggers interest from EU's antitrust chief

In addition to OpenAI, Microsoft did a deal with Inflection AI, after co-founders Mustafa Suleyman and Karén Simonyan, and a bunch of staff, joined the Redmond HQ'd org in March. Microsoft paid Inflection AI $650 million in a licensing deal, while poaching the staff. This move too is under the watchful gaze of the CMA and the EC's authorities.

Regulators want to know if these moves are stealthy strategies to lock in fledgling developers of LLMs to the already powerful cloud titans, and what impact this will have on the wider user landscape.

Reg readers should feel free to give the CMA and other authorities across the globe some insight in the comments section below. ®