China’s New AI Model Tops US Rivals, Sparks Alarm

A Chinese startup released a powerful new artificial intelligence model on Thursday, immediately topping global rankings and igniting debate about America’s lead in the AI race. Moonshot AI launched Kimi K3, which it calls the largest open-weight AI system in the world. Open-weight means the model’s core code is publicly available for anyone to use or modify. The model trails top-tier US systems like Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6 Sol on overall performance. However, it beats their second-tier models in key areas like coding and completing complex tasks. Within one day, Kimi K3 ranked first on Arena’s frontend coding leaderboard, ahead of every leading US model. It also placed third on Artificial Analysis’s Intelligence Index. The release is the latest sign that Chinese labs are closing the gap with their US counterparts. It also rattled markets: Moonshot AI is valued at roughly $31.5 billion, a fraction of the trillion-dollar valuations of Anthropic and OpenAI. Tech and academic leaders quickly reacted online. Here is what they said. **David Sacks, former White House AI czar:** Called the release “concerning.” He said the US is hobbling itself with regulations on data centers and new models. “This is how you lose the AI race,” he wrote, urging “permissionless innovation.” **Vinod Khosla, billionaire investor:** Agreed with Sacks and highlighted an “even bigger issue”: US immigration policies that scare away talented foreigners. He pointed to new restrictions on student visas and H-1B applications. **Aaron Levie, CEO of Box:** Called the release a “huge win” for companies using AI. He said cheaper, high-performing AI will let businesses automate more tasks, which was previously limited by high costs. **Ethan Mollick, Wharton professor:** Offered “a note of caution.” He said Kimi K3 “messed up” when asked to audit his academic work, misapplying statistical methods. **Jason Calacanis, venture capitalist:** Said AI progress is accelerating. He predicted “open-source models are compounding” and that 2026 will be the year of AGI (artificial general intelligence), with superintelligence following in 2027 or 2028. **Russ Salakhutdinov, CMU professor:** Congratulated his former student, Moonshot’s CEO, calling the release “a huge win for the open-source community.” **Gary Marcus, NYU professor emeritus:** Reacted bluntly, saying “Congress should investigate.” He pointed to a Goldman Sachs chart showing US cloud spending could reach $1 trillion by 2027, eight times China’s projected spending. **Gavin Baker, chief investment officer:** Called the release an “inflection point.” He said it is bad for closed AI startups like OpenAI and Anthropic but good for almost every other company, as it lowers costs and increases competition. **Dean Ball, OpenAI’s head of strategic futures:** Said he was surprised China allowed such a good model to be open source. He warned this aligns with China’s political philosophy of making AI a “public good” provided by the state.