Anthropic's voice: US government's "cut" on Google (GOOGL.US) dampens AI investment enthusiasm.

AI developer Anthropic recently filed a motion in federal court to dismiss a proposal by the U.S. government aimed at stopping Google (GOOGL.US) from investing in AI startups to rectify its illegal monopoly in the online search market.

The AI startup said in court filings on Friday that the remedy of ending Google's relationship with Anthropic "would not only harm Anthropic, but would have a general impact on competition."

Google has invested about $3 billion in the company, and Amazon (AMZN.US) is also an investor, according to the filings.

Last year, a federal judge ruled that Google had illegally monopolized the online search and search advertising markets, and the Justice Department and several states proposed major reforms to Google's business, including forcing it to sell its Chrome web browser. The remedy proposal also would have barred Google from acquiring, investing in or partnering with companies that control consumer search information, including AI products.

Anthropic said in the filings that such a forced sale would "provide an unreasonable windfall to Anthropic's larger competitors in AI, including OpenAI, Meta (META.US), and, ironically, Google itself (through its subsidiary DeepMind) via the sale of its AI language model Gemini."

Anthropic is known for its Claude series of large language models, which compete with OpenAI's products. Like its peers, the company has been raising large amounts of money to fund investments in expanding its computing capacity and staying ahead in the AI development race.

The U.S. Federal Trade Commission has expressed concerns about these investments, as well as Microsoft's investment in OpenAI. In a January report, the FTC said that the tech giants were requiring some of their investments in AI startups to be used to buy their own products and services, which could lead to the concentration of advantageous data in the hands of the tech giants related to chip development, model training and data center construction.