Nvidia is reportedly weighing an investment of up to $10 billion in Anthropic’s planned initial public offering, a move that could secure a major anchor investor for what may become the largest IPO in history. Sources told Reuters the AI developer is aiming to raise as much as $100 billion at a valuation near $2 trillion, with the listing expected before November’s U.S. midterm elections.
This potential deal would deepen Nvidia’s ties with Anthropic, one of its key customers. Anthropic already relies heavily on Nvidia GPUs, and in November 2025 Nvidia said it planned to invest up to $10 billion through a broader partnership. Under that arrangement, Anthropic agreed to spend $30 billion on Microsoft Azure compute capacity powered by Nvidia hardware. Anthropic has separately committed to spending more than $100 billion with Amazon Web Services over ten years and plans to use over 1 million Amazon Trainium2 chips, while also expanding TPU capacity with Google and Broadcom.
At the same time, Moonshot AI is projecting its annualized revenue will reach $2 billion by the end of 2026, up from roughly $100 million a year earlier. The Beijing-based startup told investors that annual recurring revenue exceeded $1 billion in August, compared with about $300 million in June. The growth follows the July release of Kimi K3, a model with 2.8 trillion parameters that performed near the top of benchmarks while costing far less to run than large US models. Moonshot is also raising fresh capital at a valuation of about $50 billion ahead of a possible Hong Kong IPO this year.
However, Anthropic has accused Moonshot, DeepSeek, and Alibaba of unauthorized use of Claude outputs in a practice it called illicit distillation. Anthropic said it blocked activity tied to the companies from December 2025 through August 2026. For Moonshot, Anthropic alleged that some Kimi user prompts were secretly routed to Claude and that responses were returned to users who believed they were interacting with Kimi. In one 10-day period, close to 300,000 customer prompts were transmitted to Anthropic, primarily through accounts originating from Singapore and Japan. Anthropic also said Moonshot retained some exchanges and gathered Claude transcripts for training, with more than 23 million discussions credited to Moonshot between May and July. Some prompts contained sensitive information from individuals, multinational companies, and state-affiliated actors.
Anthropic reported more than 12 million suspected distillation attacks linked to DeepSeek over 14 days in July 2026, and more than 151 million exchanges connected to Alibaba between May and July. The cases were part of a broader report covering cyber operations, influence campaigns, surveillance, scams, biological misuse, conventional weapons development and AI model distillation.
