What happened
Moonshot AI, a Beijing-based startup, closed a $3.5 billion funding round at a $35 billion valuation—far exceeding its initial $1 billion to $2 billion target. The company is now fundraising at a $50 billion pre-money valuation ahead of a potential Hong Kong IPO as soon as this year.
Why it matters
Moonshot's K3 model has become a landmark moment in AI competition. The model approached the performance of frontier models by Anthropic and OpenAI, triggering a tech stock selloff and earning comparisons to the "DeepSeek moment" as evidence that Chinese AI labs can keep pace with the US despite export restrictions. The company's K3 became the world's largest open-weight model with 2.8 trillion parameters and a 1 million-token context window—capabilities that allow it to process massive documents or codebases in a single prompt.
What to watch
Moonshot achieved $300 million of annual recurring revenue in June, up from $200 million in April, and daily sales have surged at least six times since the K3's launch. A White House official has accused the company of training K3 on restricted Nvidia chips and using distillation (extracting outputs from competitors to boost performance)—allegations Moonshot has not publicly addressed.
Summaries like this, in your inbox every morning.
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—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.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Vinci4D provides chip-design simulation software that trains on physics equations instead of customer data, an…
SEMI President and CEO Ajit Manocha told DIGITIMES that the semiconductor industry reached the US$1 trillion r…
CoreWeave fell more than 25% this week, its debt yield briefly rose above 16% on Thursday, and SK Hynix droppe…
Nvidia's AI Enterprise lists at $4,500 per GPU annually, and paid subscriptions add supported software, securi…
Anthropic told investors this week that it is making 85% gross margins, a figure the columnist calls a cash co…
Modal Labs, an AI infrastructure provider, raised $750 million in a Series D led by Accel at a $15.75 billion…
