Moonshot AI Halts New Kimi K3 Sign-Ups as China Cuts U.S. AI Lead to 4-7 Months
Updated
Updated · The Washington Post · Jul 22
Moonshot AI Halts New Kimi K3 Sign-Ups as China Cuts U.S. AI Lead to 4-7 Months
3 articles · Updated · The Washington Post · Jul 22
Summary
Moonshot AI stopped accepting new Kimi K3 subscribers on Sunday, saying demand had exceeded expectations and it needed to preserve computing resources for existing users.
Kimi K3 and Alibaba’s Qwen3.8 are part of a new wave of Chinese models that experts say are rapidly narrowing the gap with U.S. leaders, with Chinese hacking-capable systems now trailing American peers by 4 to 7 months versus 6 to 10 months last year.
That faster progress is testing the Trump administration, which has tried to boost U.S. AI while also delaying some frontier model releases over security fears that hackers could exploit them.
Chinese developers are also gaining traction with open models that can be copied and modified, while U.S. rivals such as OpenAI and Anthropic largely keep their systems closed behind paid access.
Analysts say the U.S. still holds a decisive edge in advanced chips and data centers, but if Chinese model gains continue, the contest may shift from building frontier AI to deploying it fastest.
With China offering powerful free AI, can America's pay-to-use models still compete for global dominance?
If US AI safety features hinder cyber defense, are they becoming a national security liability?
Kimi K3’s 2.8 Trillion Parameter Model: China’s Open-Weight AI Revolution and Its Global Impact
Overview
Moonshot AI's Kimi K3 made a dramatic debut, drawing immediate and overwhelming demand that forced the company to halt new sign-ups. This rapid adoption highlighted both the impressive capabilities of Kimi K3—such as being the largest open-weight AI model with 2.8 trillion parameters and topping key benchmarks—and the challenges Chinese AI models face in scaling for a growing user base. As an open-weight model, Kimi K3 allows developers to download and run it on their own hardware, giving them more control over costs, data privacy, and customization, which has captured global attention and disrupted the AI market.