Anthropic in talks to acquire Decart AI for $6 billion

Anthropic PBC is in advanced talks to acquire the artificial intelligence startup Decart AI for about $6 billion, according to people familiar with the negotiations who spoke on condition of anonymity. The deal has not been finalized, and discussions could still fall apart, they cautioned. If the acquisition goes through, it would represent Anthropic's largest known acquisition and would come ahead of the company's planned initial public offering, which has been hotly anticipated by investors.

Decart's core strength lies in software that reduces the cost of training AI models by enabling chips to work more efficiently. This technology would allow Anthropic's existing computing infrastructure to handle more demand—a critical advantage as the company scales its operations and serves a growing customer base. Beyond training optimization, Decart also specializes in generative video work, leveraging what are called world models that can instantly modify live video feeds. This second area of focus is regarded as evidence of the startup's high-quality infrastructure engineering talent, making it an attractive acquisition target for Anthropic, which has been spending heavily on computing power to develop new products.

Decart, founded in 2023 by three Israeli engineers—brothers Dean and Orian Leitersdorf and Moshe Shalev—has grown quickly in valuation. In May 2026, the startup closed a $300 million funding round led by Radical Ventures, with participation from Nvidia Corp., Atreides Management, Valor Equity Partners, and Adobe Ventures, alongside returning investors Sequoia Capital, Benchmark, and Zeev Ventures. That round valued Decart at almost $4 billion, up from $3.1 billion just nine months earlier in August 2025. Both Anthropic and Decart declined to comment on the acquisition discussions.

The acquisition talk comes as OpenAI and Anthropic each commit to spending tens, if not hundreds, of billions of dollars on data centers filled with expensive chips, and both are increasingly sourcing hardware from multiple vendors to meet their expanding computing needs. These heavy infrastructure outlays—which both companies say are necessary to build and maintain cutting-edge AI models—are expected to exert pressure on their finances as they prepare for their respective Wall Street debuts in the coming months.