\n China may have done it again.\n <\/p>\n
\n The latest Chinese AI model making waves was developed by a lesser-known startup called Moonshot AI. In just two days after launch, Moonshot suspended new subscriptions after demand overwhelmed its computing capacity.\n <\/p>\n
\n The unexpected enthusiasm for Moonshot\u2019s Kimi K3 model stems from its ability to match \u2013 and in some cases outperform \u2013 its most advanced US competitors, such as the latest iterations of OpenAI\u2019s GPT and Anthropic\u2019s Claude, for free.\n <\/p>\n
\n Now its rise is intensifying an already contentious race between the US and China to dominate cutting-edge AI. Washington has ramped up export restrictions against China in recent years to slow the country\u2019s development of AI and chips that could power its military.\n <\/p>\n
\n After Kimi K3 was released, US officials and Anthropic have again accused Moonshot of distilling American models, a technique in which one AI model is trained using outputs from another. Moonshot has denied the allegations.\n <\/p>\n
\n Still, the fears spreading among US companies and government officials underscore the stakes of the AI rivalry, and its far-reaching implications for economy, defense and national security.\n <\/p>\n
\n More broadly, Kimi K3\u2019s capabilities are renewing questions about the effectiveness of Washington\u2019s tech export curbs. The new model also challenges an industry belief that leading-edge AI requires heavy spending on data centers and the advanced chips that the US has restricted access to.\n <\/p>\n
![\"Moonshot](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/gettyimages-2286500505a.JPG\")
<\/p>\n
![\"People](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/gettyimages-2285951625a.JPG\")
<\/p>\n
\n Unlike most American proprietary models, Kimi K3 was released as an open-weight model. That allows developers to download it, modify it, and run it on their own servers for free, positioning it for easier adoption and a wider market globally.\n <\/p>\n
\n Constrained by limited access to AI hardware and investment<\/a> compared with American firms, China has embraced so-called open models, in contrast to US rivals that mostly sit behind paid subscriptions.\n <\/p>\n \n \u201cMany organisations value the ability to self-host open models, allowing them to retain the weight and control over sensitive data instead of relying solely on closed-model providers,\u201d said Chelsey Tam, senior equity analyst at financial services firm Morningstar.\n <\/p>\n \n Moonshot AI was founded three years ago by Yang Zhilin, a promising software engineer in his early 30s from the southeastern Chinese city of Shantou.\n <\/p>\n \n He demonstrated potential early on when studying at Tsinghua University, China\u2019s most elite engineering school, under Tang Jie, a professor who would go on to found another prominent AI startup, Z.ai, in 2019.\n <\/p>\n \n \u201cHe is the most talented and exceptional student I\u2019ve encountered in recent years,\u201d Tang said while recommending Yang for a scholarship in 2014.\n <\/p>\n \n Yang then went to Carnegie Mellon University for his PhD in computer science, studying under Ruslan Salakhutdinov, a former AI executive at Meta and Apple. By the end of his studies, Salakhutdinov recalled, Yang turned down a job offer from a senior Apple executive who reported directly to Tim Cook.\n <\/p>\n \n \u201cI remember him telling me that if he didn\u2019t at least try starting his own company, he would regret it for the rest of his life. I respect that, and he was right,\u201d Salakhutdinov said in a post on X<\/a>.\n <\/p>\n \n Yang later said the decision came down to two things. \u201cFirst was the environment. We saw tremendous opportunities, from government and venture-capital support to the advances China had made in education over the past few decades, all of which created favorable conditions for AI startups,\u201d he told state-run media<\/a> in 2024. \u201cSecond, we believed AI represented a massive opportunity.\u201d\n <\/p>\n \n Yang stood out in China\u2019s AI industry not only for founding Moonshot at the age of 32, making him one of the country\u2019s youngest AI entrepreneurs, but also for choosing to return to China to launch his venture.\n <\/p>\n \n From the start, Yang had grand ambitions. He told a podcast<\/a> in 2024 that the startup\u2019s name was a nod to his favorite album by British rock band Pink Floyd, \u201cDark Side of the Moon,\u201d in the hope that it would \u201caccomplish something that would create enormous value but would also be extremely difficult \u2014 much like landing on the moon.\u201d\n <\/p>\n \n Moonshot quickly reached national fame after releasing its first Kimi model \u2013 named after Yang\u2019s English name \u2013 in 2023. That same year, Moonshot secured nearly $300 million in its first two funding rounds, according to a 2024 interview with Chinese media<\/a>.\n <\/p>\n \n The international attention that Kimi K3 has garnered echoes the rise of DeepSeek in early 2025, when its R1 reasoning model sent shock waves through the global AI community.\n <\/p>\n \n This time, K3 delivered performance approaching frontier systems such as OpenAI\u2019s GPT 5.6 Sol and Anthropic\u2019s Claude Fable 5 on various benchmarks, further narrowing the gap between Chinese and American AI models.\n <\/p>\n \n But analysts say Moonshot\u2019s achievement should be viewed through a different lens.\n <\/p>\n \n DeepSeek challenged the conventional wisdom that developing an advanced AI model requires massive financial investment and vast computing power, and offered competitive pricing.\n <\/p>\n \n But K3\u2019s discount to frontier models is much smaller than R1\u2019s was at launch, suggesting its cost advantage may be less significant than DeepSeek\u2019s, according to Ivan Su, director at Morningstar.\n <\/p>\n \n What K3 demonstrates is that Chinese companies are continuing to build on the success of DeepSeek, and close the gap between Chinese and US AI models, said Poe Zhao, analyst and founder of the newsletter Hello China Tech.\n <\/p>\n \n \u201cIts broader significance is that DeepSeek increasingly looks like the beginning of a pattern,\u201d he said. \u201cDeepSeek was a surprise. K3 is evidence that China\u2019s progress is becoming more repeatable.\u201d\n <\/p>\n \n Since DeepSeek, Chinese AI firms like Moonshot, Alibaba and Z.ai are getting more attention for each new model, with the pace of the rollouts speeding up.\n <\/p>\n \n With K3, Wei Sun, principal AI analyst at market analysis firm Counterpoint Research, said the general gap between Chinese and American models has narrowed to three to six months, though it varies significantly by task.\n <\/p>\n \n Still, other analysts have expressed reservations toward K3\u2019s true capabilities.\n <\/p>\n \n \u201cWhile K3 constitutes progress, we\u2019d hesitate to ascribe it near-parity with American frontier models, such as Fable 5, in real-world tasks,\u201d said Malik Ahmed Khan, senior equity analyst at Morningstar in a research note last Friday. Many open model developers have optimized software performance around benchmarks for better results in the past, he explained.\n <\/p>\n Is China overtaking the US in AI?\n <\/p>\n Is China overtaking the US in AI?<\/p>\n 1:25<\/p>\n \n For American tech companies, free and comparable Chinese models like K3 threaten to undermine their own products.\n <\/p>\n \n Leading US AI firms Anthropic and OpenAI have long accused<\/a> Chinese labs, including Moonshot, DeepSeek and Alibaba, of extracting capabilities from their models. Those claims have intensified with the arrival of K3.\n <\/p>\n \n Huang Zhenxin, Moonshot AI\u2019s head of enterprise business, denied the accusations of distillation in an interview with state-run media<\/a> on Monday.\n <\/p>\n \n But that has not stopped new allegations from Washington.\n <\/p>\n \n \u201cWe have information that Moonshot AI distilled Anthropic\u2019s Fable for the development of its K3 model,\u201d said Michael Kratsios, director of the White House Office of Science and Technology Policy, in a Wednesday post on X<\/a>. \u201cLarge-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.\u201d\n <\/p>\n \n Sarah Heck, Anthropic\u2019s policy head, echoed Kratsios in an X post<\/a> that accused Moonshot of engaging in \u201cIP theft and industrial espionage,\u201d calling it a \u201cnational challenge that creates serious national security risks.\u201d\n <\/p>\n \n CNN has reached out to Moonshot for comment.\n <\/p>\n \n Treasury Secretary Scott Bessent also reiterated Wednesday that the US could sanction Chinese firms over such IP theft, without mentioning Moonshot or others by name.\n <\/p>\n![\"Yang](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/ap26084551982585a.JPG\")
<\/p>\n![\"DeepSeek's](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/ap25028376456937.jpg\")
<\/p>\n![\"china](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/china-us-ai-16x9.jpg\")
<\/p>\n![\"china](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/china-us-ai-16x9.jpg&q=w_860,c_fill\")
<\/p>\n![\"The](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/1784869453_905_2026-06-04t232737z-1911130883-rc2ewjaj90q5-rtrmadp-3-anthropic-ai.jpg\")
<\/p>\n