Highlights:

  • In February of last year, Google LLC invested USD 300 million in Anthropic, becoming a key investor.
  • The CMA will investigate whether Google’s partnership with Anthropic effectively merges the two companies, which could reduce competition.

The U.K.’s antitrust watchdog has recently initiated preliminary steps towards a potential Google-Anthropic Partnership Probe.

San Francisco-based Anthropic, one of OpenAI’s best-funded competitors with over USD 7 billion raised to date, develops the Claude series of large language models. The company’s latest model, launched earlier this month, can outperform OpenAI’s flagship GPT-4o across several types of tasks.

Google LLC became an investor in Anthropic last February with a USD 300 million investment, acquiring a 10% stake in the company. Following this, Google supported a subsequent USD 450 million funding round for Anthropic alongside other investors. More recently, Google has committed to investing up to USD 2 billion in the startup.

Today, the Competition and Markets Authority (CMA) in the U.K. invited feedback on the Google-Anthropic partnership. The regulator is evaluating whether the deal could potentially harm market competition in the U.K. Interested parties have until August 13 to submit their views.

As part of the investigation, the CMA will assess whether Google’s partnership with Anthropic qualifies as a “relevant merger situation.” This regulatory term refers to scenarios where two companies either cease to be distinct entities or are approaching that state.

Requests for feedback from the CMA can initiate a Phase 1 antitrust investigation. This preliminary probe can either conclude with a decision in favor of the scrutinized company or lead to a more comprehensive Phase 2 investigation, which typically spans several months. Upon completing the Phase 2 investigation, the CMA may require the company to alter any business practices deemed to violate antitrust regulations.

The watchdog is initiating its inquiry just days after Anthropic unveiled its latest and most advanced large language model, Claude 3.5 Sonnet. In internal testing, the model outperformed GPT-4o in six out of eight task categories. Additionally, Anthropic claims that Claude 3.5 Sonnet can generate responses approximately twice as fast as its previous flagship model.

Prior to the model’s release, the company submitted it to the U.K.’s Artificial Intelligence Safety Institute, a newly established government organization responsible for examining the potential risks associated with new AI systems.

Anthropic told a leading media house that “we intend to cooperate with the CMA and provide them with the complete picture about Google’s investment and our commercial collaboration. 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.”

Google stated, “Anthropic is free to use multiple cloud providers and does, and we don’t demand exclusive tech rights.”

Anthropic has also secured USD 4 billion from Google competitor Amazon.com Inc., with the investment closing in March. When the deal was announced, Anthropic mentioned that it would use Amazon Web Services Inc. infrastructure to train its new models. This partnership also attracted the CMA’s attention, leading the regulator to request input from interested parties about the deal in April.

The watchdog had previously sought feedback on Microsoft Corp.’s investment in OpenAI. Meanwhile, the European Commission has indicated that it might initiate an antitrust investigation into the companies’ partnership. Recently, the commission sent Microsoft a series of questions concerning “certain exclusivity clauses” in the partnership agreement, which officials suspect may have violated antitrust regulations.