Chen Tianqiao's MiroMind and Dai Jifeng's Open-Source Venture: A Dispute Shaped by Geopolitics

Forum topic · ✨步子哥 · 2026-04-29

Summary

This forum post chronicles the dispute between Chen Tianqiao's MiroMind and AI researcher Dai Jifeng. Chen Tianqiao, former Shanda Group chairman and once China's richest man, moved into brain science after leaving gaming, founding the Tianqiao Brain Research Institute. With the rise of ChatGPT, he launched MiroMind in March 2025 with Tsinghua associate professor Dai Jifeng to develop open-source large models, releasing the MiroThinker 1.5 agent model. Dai left in January, offering 15% equity for IP rights; after investors sought 5% plus unrestricted global IP licensing, MiroMind refused. Dai later told The Washington Post he was pressured to work overseas; MiroMind's internal notice accused him of bad faith and trade-secret violations. The dispute traces to Meta's blocked $2 billion acquisition of Manus, after which China's NDRC prohibited the deal, prompting Chen to impose a US-China "separation wall": no cross-border code or data sharing. Dai founded Naiveai, raising about $300 million at roughly $800 million valuation. The author urges both sides to focus on real engineering over publicity wars.

Chen Tianqiao's Overseas Island and Dai Jifeng's Open-Source Awakening

Chen Tianqiao: From NASDAQ Billionaire to Brain-Science Recluse

Chen Tianqiao was once China's richest man. Riding the online gaming boom, he led Shanda Group to NASDAQ in 2004, raising $152 million and becoming a billionaire at thirty. After losing ground to Tencent and Alibaba, Shanda was privatized in 2012 and converted into an investment group. Suffering from panic disorder, Chen flew only once from China to California; for the following sixteen years he lived overseas — first Singapore, then California — and never returned to China.

Five years ago, the author interviewed Chen, who by then had turned to brain science, founding the Tianqiao Brain Research Institute and TCCI, aiming to explore brain-computer interfaces without opening the skull. He described the brain-computer revolution as: "Musk goes left, I go right" — one side implants chips, the other pursues non-invasive approaches.

ChatGPT Sparks a New AI Venture

The ChatGPT wave awakened Chen's dormant brain-science ambitions. In March 2025, he personally incubated MiroMind together with Dai Jifeng, an associate professor at Tsinghua University's Department of Electronic Engineering. Dai is a leading scholar in computer vision — his representative work includes Deformable Convolutional Networks and the multimodal foundation model InternVL, with over 50,000 citations. He previously served as principal researcher at Microsoft Research Asia and executive research director at SenseTime, returning to academia in 2022. MiroMind targets open-source large models, benchmarked against DeepSeek, while supporting brain science in parallel with Musk's Neuralink. The company released its flagship search-agent model, MiroThinker 1.5.

The Short-Lived Partnership: Equity and IP Dispute

March 2025: Dai joins MiroMind and requests US work; the company sponsors his O-1 visa (approved in November).
January 15: Dai resigns, offering 15% equity of his new company as consideration for legally using MiroMind's IP, and is permitted to take key domestic team members. MiroMind even helps connect him with investors such as IDG and Sequoia.
March 2: Dai's backing investors demand the equity drop to 5%, plus an "unconditional, worldwide, royalty-free, unlimited-term, irrevocable, sublicensable grant with no restrictions or objections." MiroMind refuses.
April 22: Dai speaks via The Washington Post, claiming that after the Manus incident MiroMind pressured him to develop overseas, which he refused, leading to his departure.

MiroMind's Internal Notice

On April 24, MiroMind issued an internal notice titled "Regarding Dai Jifeng's False Statements," accusing him of "bad-faith breach," detailing the visa sponsorship and its respect for his departure, claiming reputational damage, and stating it holds leads on "trade-secret infringement" while reserving rights to criminal complaint and civil litigation.

The Manus Deal and the Geopolitical Trigger

The root cause lies in Meta's $2 billion acquisition of Manus. Though registered in Singapore, Manus's founding team originally developed its technology in China. Chen invested $100 million in MiroMind, part of Shanda's $2 billion "discoverable AI" investment. Manus had hired over sixty scientists across Singapore, Tokyo, and Seattle, relocating Chinese staff to Singapore and cutting dozens of positions.

April 27: China's NDRC lawfully prohibited the foreign acquisition and ordered the transaction revoked, ending Meta's $2+ billion deal (~14 billion RMB). CCTV commented that banning "relocation-bath-style" overseas restructuring targets security risks within openness.
April 28: Chen told Bloomberg that MiroMind had implemented protocols prohibiting cross-border sharing of information or code, minimizing movement of personnel, data, and assets. He said: "Before, I believed we could gather Chinese and foreign AI talent for humanity's future. After the Manus incident, we had to implement comprehensive risk controls. This is like 'cutting off our own retreat,' but in the current environment it is a necessary compromise." He stressed: "The international environment is extremely complex. Under geopolitical conditions, companies have no choice but to pick sides."

The Separation Wall

In March, Chinese officials contacted Chen's team, warning against unilateral technology transfer. Chen's internal isolation plan — regional businesses managed locally — resolved the issue. MiroMind plans its first external fundraising in the second half of the year and has gained revenue through asset-management and energy-infrastructure transactions. Dai subsequently founded Naiveai, raising roughly $300 million at about $800 million valuation, with a core team including former MiroMind members. Chen argues these events force companies to pick sides and that global AI OS technology rules have been rewritten.

Commentary: Substance Over Noise

The author, declaring no shareholder ties or interests with Chen, argues that one-sided online pile-ons and deliberate publicity stunts are pointless. Only two things matter for a company: whether core technology can genuinely reach mass production, and whether the team can run a real business closed loop. A Jiaguang Year report said MiroMind claimed Dai and Chinese AI researchers remotely served a Singapore company using US AI chips and compute, facing cloud control and legal risk. The author's view: Chen's approach was reasonable, Dai's statements weren't wrong either — both sides have a point, and the key is turning conflict into progress.

> "Picking sides" under geopolitical pressure is not a simple political choice but a survival calculation — like Silk Road caravans forced to choose a route at checkpoints. AI startups must consider markets and exit channels from day one; selling to foreign tech giants now carries high policy risk.

Conclusion

AI is the great tide of technology. Chen Tianqiao's brain-computer dream and Dai Jifeng's open-source path, though temporarily diverging, both serve as lighthouses of human intelligence. The author hopes both sides will turn conflict into cooperation and let technology truly benefit everyone.

---

References 1. Chen Tianqiao's statement to Bloomberg on MiroMind's risk controls and the Manus incident, April 28, 2026. 2. MiroMind internal notice regarding Dai Jifeng's statements, April 24, 2026. 3. NDRC decision prohibiting Meta's acquisition of Manus, April 27, 2026. 4. Yicai and Caixin reports on MiroMind/Naiveai equity and departure details, 2026. 5. Bloomberg in-depth interview on Chen Tianqiao's MiroMind investment and the AI separation-wall plan, 2026.