Moonshot AI hits $35B valuation after $3.5B funding round

Moonshot AI, a Beijing-based startup, secured a $35 billion（約5.6兆円） valuation after closing a $3.5 billion（約5600億円） funding round—substantially larger than the $1 billion（約1600億円） to $2 billion（約3200億円） it initially targeted. The company is already reaching out to new backers at a $50 billion（約8兆円） pre-money valuation, with plans for a Hong Kong initial public offering as soon as this year.

The funding frenzy reflects investor enthusiasm for Moonshot's K3 model, which the company released this month. The K3 features 2.8 trillion parameters—a measure of AI complexity—and a 1 million-token context window that enables it to process entire documents or codebases in a single prompt. These specifications made it the world's largest open-weight model. Critically, the K3's performance approached that of frontier models developed by Anthropic and OpenAI, triggering a selloff in tech stocks and prompting comparisons to the "DeepSeek moment"—shorthand for evidence that Chinese AI labs can match US capabilities despite export curbs and computing constraints. On Monday, Moonshot released the model's weights, allowing developers to download, modify, and host the technology freely.

Commercially, Moonshot achieved $300 million（約480億円） of annual recurring revenue in June, up from $200 million（約320億円） in April. Since the K3's launch, daily sales have surged at least six times, according to one person familiar with the matter. However, a White House official has accused the company of training K3 on restricted Nvidia chips and of using distillation—a technique involving extraction of outputs from competitors to enhance performance. Moonshot has not publicly responded to these allegations or to repeated requests for comment.

Moonshot AI's funding round underscores the intensity of investor appetite for Chinese AI startups that can match or challenge US-developed models. The company's ability to raise $3.5 billion（約5600億円）—more than triple its initial target—reflects market confidence in both its technical capabilities and commercial traction. The K3 model's release of openly available weights marks a strategic pivot toward an open-source approach, contrasting with the proprietary stance of US competitors like OpenAI and Anthropic.

The White House's public accusation that Moonshot trained K3 on restricted Nvidia chips signals a new front in US-China AI competition. Export controls and computing constraints have been expected to slow Chinese AI development, yet the K3's performance has contradicted that assumption. The model's 2.8 trillion parameters and 1 million-token context window represent technical achievements that narrow the perceived gap between Chinese and Western systems—a shift with policy implications beyond the private market.

No comments yet. Be the first to share your thoughts!