A key "enhancement" on the eve of the IPO? It is rumored that Anthropic is negotiating a $6 billion acquisition of Decart AI, focusing on optimizing AI computing efficiency.

As the highly anticipated initial public offering (IPO) approaches in just a few weeks, AI giant Anthropic PBC is accelerating its efforts to strengthen its technological moat through acquisitions. According to informed sources, this AI star company is in deep negotiations to acquire the Israeli AI infrastructure startup Decart AI, with a deal valuation of around $6 billion. If this transaction goes through, it will mark Anthropic's largest acquisition to date and represents a critical "technology enhancement" move before its IPOit is noteworthy that Nvidia was originally a potential buyer for this deal but was outmaneuvered by Anthropic as negotiations neared completion. Decart AI: The Israeli upstart that makes chips "run faster" Founded in 2023 in Israel by brothers Dean and Orian Leitersdorf alongside Moshe Shalev, Decart AI's core competency lies in maximizing the operational efficiency of various chipsits software can optimize the training and inference processes of AI models, helping developers extract more computing power from existing hardware. Technical layout: Chip performance optimization layer: Helps AI developers efficiently run models across various chips from Nvidia to AMD, providing strategic value in the current context of scarce AI computing power; Generative video and "world models": Can instantly modify real-time video streams, showcasing its high-quality talent pool in infrastructure. Financing trajectory and valuation leap: In May of this year, Decart completed a $300 million financing round led by Radical Ventures, with participation from Nvidia, Atreides Management, Valor Equity Partners, and Adobe Ventures, resulting in a post-investment valuation close to $4 billion. This means that Anthropic's offer has pushed Decarts valuation up by 50% in just three months. Even more dramatically, Decart was poised to finalize a sale agreement with Nvidia when another "larger" buyer intervened, leading the founders to pivot. The industry widely speculates that this "interloper" is Anthropic. Anthropics strategic logic: The "last-mile" enhancement before the IPO This $6 billion acquisition, if completed, holds triple strategic significance for Anthropic: Firstly, the "leverage effect" of computing efficiency. Anthropic has committed to investing tens of billions of dollars in building data centers equipped with expensive chips. Decart's optimization technology allows existing infrastructure to handle greater demand, effectively adding "leverage" to Anthropic's computing investment amid a continuing chip supply crunch. The Decart team is expected to be integrated into Anthropic's inference and performance organization. Secondly, "safeguarding" valuation on the eve of the IPO. Anthropic plans to go public in September or early October, with a valuation target of $965 billion to $1 trillion. Completing a high-profile strategic acquisition just before the IPO will help tell a more comprehensive "full-stack AI" story to public market investors. Thirdly, differentiation in competing with OpenAI. Against the backdrop of OpenAI's scale triumphing, Anthropic has chosen to establish barriers on the efficiency dimension. Decart's ability to achieve "higher performance on the same chips" complements Anthropic's philosophy of "safe and controllable" technology. Countdown to IPO: "Stress testing" the trillion-dollar valuation Anthropics IPO has entered its final sprint phase. The company had previously submitted a confidential S-1 prospectus draft to the U.S. Securities and Exchange Commission (SEC). Market predictions from platform Kalshi indicate that traders believe the probability of Anthropic announcing its IPO by 2026 has risen to 85%. However, the "stress test" for the trillion-dollar valuation has already shown cracks during pre-IPO investor meetings. Investors are concentrating on three major questions: the competitive threat from low-cost Chinese AI systems; the tense relationship between Anthropic and the Trump administration; and the precedent set by SpaceX's stock price plummeting from $225 to $108 post-IPO. Industry perspective: Fundamental model companies acquiring "efficiency layer" assets Anthropics acquisition of Decart reflects a deeper trend in the AI industry: fundamental model companies are competing to acquire reasoning and efficiency capabilities rather than relying solely on internal development. As AI models continue to scale and computing costs remain high, "efficiency" is becoming as critical a competitive dimension as "performance." At the same time, both OpenAI and Anthropic have committed to investing tens of billions or even hundreds of billions of dollars in building data centers. Under the pressure of high capital expenditures, enhancing the utilization efficiency of existing infrastructure through acquisitions is becoming a more cost-effective pathway. Nvidia, initially an investor and potential buyer of Decart, was outmaneuvered by Anthropic in the final stages of negotiations. This reveals a profound trend: AI model developers are actively building their own tech stack rather than merely relying on chip suppliers. As the competition for computing scale becomes increasingly homogenized, "how to use computing power more efficiently" is emerging as the core competitive advantage for AI companies. The Decart acquisition marks the official launch of the AI efficiency war. Of course, this transaction still carries uncertaintiesinsiders stress that negotiations are not yet finalized and could still fall through. However, for Anthropic, which aims for a trillion-dollar IPO, acquiring Decart AI is not just a business decision but a demonstration of its technological ambition in the marketa true "IPO declaration."