Anthropic is reportedly in advanced discussions to acquire Nvidia-backed artificial intelligence startup Decart for approximately $6 billion. The potential deal, which remains unfinalized, marks a major strategic move by the Claude developer to expand its computing infrastructure and secure optimization technology ahead of a highly anticipated public offering.
Artificial intelligence developer Anthropic PBC has entered into negotiations to acquire Israeli startup Decart AI in what would be the largest acquisition in Anthropic’s history. The transaction, first reported by Bloomberg News and confirmed by Reuters, is valued at roughly $6 billion. People familiar with the matter cautioned that the deal has not been finalised and talks could fall through.
Representatives for both Anthropic and Decart declined to comment on the negotiations, while sources indicate that the discussions remain at an early stage. If successfully concluded, the acquisition would bring Decart’s engineering team directly into Anthropic’s inference and performance organisation.
Infrastructure Optimization and Rising Industry Pressures
The pursuit of Decart highlights a central economic bottleneck confronting the artificial intelligence industry: the soaring cost and scarcity of computing infrastructure. Leading model developers like OpenAI and Anthropic have committed to spending tens, if not hundreds, of billions of dollars on data centers stocked with expensive chips. As these startups prepare for Wall Street debuts, optimizing the hardware stack has become an urgent priority.
Founded in 2023 by brothers Dean and Orian Leitersdorf alongside Moshe Shalev, Decart specializes in infrastructure and optimization technology designed to reduce the cost of training artificial intelligence models by helping chips work more efficiently. For Anthropic, integrating this technology could provide the operational bandwidth needed to absorb growing user demand for its Claude assistant as the company races to overcome capacity constraints.
Real-Time Generative Video and Simulation Models
Beyond core chip optimization, Decart has gained industry attention for its advancements in generative video and simulation environments. The startup developed Oasis, a model that generates simulated environments to train and test robotics and autonomous-driving systems, as well as the Lucy
model, which can edit live video in real time. According to company disclosures, Oasis rapidly captured public attention, surpassing ChatGPT in downloads and reaching 1 million users within three days of its release as a playable video game.
Industry observers note that Decart’s work with world models capable of modifying live video feeds instantly demonstrates the kind of high-quality infrastructure talent that major AI labs are aggressively pursuing to slash generation costs.
Rapid Valuation Growth and Competing Suitors
A $6 billion transaction would represent a striking upward trajectory for a company that has existed for less than two years and employs roughly 100 people. In May, Decart announced a $300 million funding round led by Radical Ventures, with Nvidia joining as a new investor alongside Atreides Management, Valor Equity Partners, and Adobe Ventures, alongside returning backers Sequoia Capital, Benchmark, and Zeev Ventures. That financing valued the startup at nearly $4 billion, up from $3.1 billion in August 2025.
The path to Anthropic’s doorstep involved a shifting field of heavyweights. Prior to talks with Anthropic, Nvidia had held advanced negotiations to acquire Decart directly, though those discussions collapsed after higher offers emerged from other technology players. Amazon, Nebius, and SpaceX were also identified by market sources as prospective suitors before Elon Musk confirmed that SpaceX would not pursue a deal.
Preparation for a Major Public Offering
The acquisition talks coincide with a broader push by Anthropic to scale its technical operations ahead of a mega initial public offering. The company has moved aggressively to secure computing power, hiring engineers with experience across the hardware and software stack to co-design custom chips. By bringing Decart into its fold, Anthropic aims to lock down the optimization tools necessary to make its AI models faster and more efficient as infrastructure expenditures mount across the sector.
Worth a look
