Chinese AI research is no longer being watched only through company launches, translated blog posts, and outside analysis. A growing number of researchers and executives from Chinese AI labs are using X to speak directly with the wider AI community.
That matters because models such as Kimi K3 and DeepSeek R1 have often been treated in Silicon Valley as if they appeared suddenly. The activity now visible on X suggests a different picture: researchers, founders, former employees and collaborators are participating in public discussions about AI development, commercialization and technical work.
A More Direct Window Into Chinese AI Labs
Moonshot AI, the company behind Kimi K3, stands out as one of the clearest examples. On Thursday, around 30 accounts on X could be found that were run by people who claimed to be currently affiliated with Moonshot, including two of the startup’s cofounders. There were also half a dozen former employees and collaborators.
These accounts are not simply placeholders. They discuss major Moonshot releases, smaller research papers, relationships with Western researchers, and parts of their personal lives and hobbies. That mix gives outside readers a broader view than a formal company announcement can provide.
Moonshot appears to be the Chinese AI lab most active on X, but it is not alone. Executives and technical staff from Minimax and Z.ai are also present. Several DeepSeek employees, who reportedly are unable to leave China because the government confiscated their passports, regularly post about AI news and share job listings.
The result is a public conversation that is more immediate and more personal than the older pattern of waiting for official releases or relying on intermediaries to interpret Chinese AI work for overseas audiences.
Why X Has Become Attractive
Meng Fanqing, cofounder of Evolvent AI and a former intern at Moonshot who worked on the Kimi K2.5 model, described X as a particularly useful venue for AI discussion. “X is the best community” for AI discussions right now, he says.
His explanation is straightforward: X includes both Chinese and overseas audiences, and its recommendation algorithm helps surface material within the same technical circle. Compared with Chinese platforms, where serious technical discussion can be harder to sustain, some Chinese researchers would rather spend their time on X.
Meng says he began actively posting on X in mid-2025. Earlier that year, DeepSeek’s R1 model went viral around the world and brought intense attention to China’s AI industry. According to Meng, many Chinese researchers used that moment to begin sharing their own perspective internationally.
“Around that time, all the [Chinese] researchers are starting to think about branding on an international level,” Meng says.
That branding is not only corporate. In AI, individual researchers can become visible voices in their own right, especially when they explain why a model or paper matters, what tradeoffs they see, and how they interpret fast-moving developments.
Chinese Platforms Are Not Filling the Same Role
One reason X has gained importance is the absence of an equivalent Chinese platform for high-quality technical AI conversation. Zhihu, a Quora-like platform, once positioned itself as a destination for educated professionals. Since 2020, however, it has increasingly moved toward fiction content, and expert contributors have gradually disappeared.
Deng Yu, a Chinese mathematician who was awarded a Fields Medal this year, publicly wrote on Zhihu in 2018 that he had decided to stop posting because of low-quality content. That example reflects a broader concern among some technical users about whether the platform still supports the discussions they want.
Xiaohongshu, also known as RedNote, has become increasingly popular among young, urban and educated people in China. Some junior AI researchers have recently opened accounts there and used them to discuss their work. But the audience is described as less technical than what researchers can find on X.
The contrast became visible when Moonshot published a new research paper in March. The paper drew attention on both Chinese and Western social media and even attracted praise from Elon Musk. A current Moonshot researcher, speaking anonymously because he was not authorized to talk to media, said Chinese platforms focused heavily on the authors’ backgrounds, ages and parents, while foreign readers discussed technical details and praised the work’s breakthroughs.
The same researcher said Zhihu readers tend to be more mean-spirited and that the platform has more AI slop than X. Moonshot did not immediately respond to a request for comment.
Western Audiences Are Paying Attention
Chinese researchers arriving on X have generally been welcomed by academics and researchers at Western AI companies. Many of those users were already looking for more direct information about Chinese AI labs.
Tiezhen Wang, an independent AI analyst and former researcher at Hugging Face, says this interest existed before the recent wave of Chinese researchers joined X. “Before they all came to X, we already saw a lot of influential researchers translating Chinese [AI] blogs into English,” Wang says.
Wang points to Su Jianlin as an example. Su, a respected AI researcher, has long maintained an active Chinese-language blog documenting his ideas. He now works for Moonshot and also has a public X account.
Another factor is that researchers at OpenAI and Anthropic have become more cautious about sharing their thinking online as their employers have grown into larger companies. Wang says they may see model architecture and training methods as secrets, which means they are posting less. That creates more room for Chinese researchers to attract attention from AI audiences trying to understand the field.
Marketing, Feedback and Less Guesswork
For Chinese AI companies, X is not only a discussion forum. It is also a way to market products, attract attention, and recruit. Meng says the platform offers major traffic and exposure, and that the AI industry now resembles finance in one important way: individuals and companies both need brands.
Wang says he advised Moonshot during a one-off consultation session in 2025 to bring more researchers forward on social media. In his view, people often find a researcher’s personal explanation more compelling than an official announcement. That kind of posting can also generate audience feedback, connect researchers with peers, and potentially lead to collaboration.
Moonshot may be especially suited to this kind of public presence. Chinese media has described it as one of the AI labs with a distinctive internal culture. Its CEO has decorated his office with band posters and souvenirs tied to his love of rock music, and the company’s Chinese name, 月之暗面, comes from Pink Floyd’s album The Dark Side of the Moon.
The larger significance is informational. Wang argues that when Chinese AI researchers explain their work directly on X, it helps reduce blind speculation about what Chinese labs are doing. In his words, “People have a lot of wrong hallucinations about what China is doing. And now, X is the door for them to better understand China.”