The Handshake That Became a Fist

In 2006, guanxi was a word American business readers were supposed to learn the way they once learned kaizen or glasnost — as a key to unlocking a foreign system that was, at bottom, eager to be unlocked. Buderi and Huang's book operates entirely within this framework. Microsoft Research Asia is presented as a triumph of cultural fluency: hire the right bicultural leader, cultivate relationships with the Ministry of Education, sponsor the right programming competitions, and China opens like a flower. The thesis is that guanxi — those layered, reciprocal, trust-based relationships — is the operating system of Chinese business life, and that a sufficiently patient American corporation can learn to run on it. The book is detailed, well-reported, and genuinely useful as a chronicle of how MSRA became one of the most productive industrial research labs in the world. What it cannot see, because no one standing in 2006 could easily see it, is that the relationships it celebrates were load-bearing walls in a structure that would be deliberately demolished.

The prescience is real but narrow. The book correctly identifies China's deep bench of computer science talent as a strategic resource of global consequence. It sees that search is the coming battlefield — Bill Gates's 2003 email about Google reads now like a dispatch from a war that Microsoft would half-win with Bing and half-lose everywhere else. The emphasis on speech recognition, natural language processing, and handwriting interfaces as research priorities at MSRA was genuinely forward-looking; these were the precursors to the language model revolution that would reshape the industry two decades later. Kai-Fu Lee's trajectory from Microsoft to Google to his own AI ventures in China tracks almost perfectly with the book's implicit argument that talent flows follow opportunity, not flags. And the portrait of Harry Shum, quietly accumulating influence across research, education, and product groups, is the portrait of a man who would eventually run all of Microsoft AI before departing in 2020. The book saw the people clearly.

What it did not see — could not see — was decoupling. The entire narrative assumes a future of deepening integration between American technology companies and the Chinese state, Chinese universities, and Chinese talent pipelines. The word "espionage" does not appear. The possibility that the U.S. government might restrict chip exports, blacklist Chinese AI firms, or treat the flow of researchers between Redmond and Beijing as a national security threat exists nowhere in these pages. Nor does the idea that China itself might decide it no longer needs or wants American platforms — that WeChat, Baidu, Alibaba, and later DeepSeek and dozens of domestic LLM efforts would render Microsoft's patient guanxi-building in China strategically moot for its consumer products. The book treats piracy as the main obstacle to Microsoft's China business. Twenty years later, the obstacles are sanctions, the Great Firewall's maturation, and a Chinese tech ecosystem that has achieved escape velocity. The Venus set-top box failure, narrated as a cautionary tale about insufficient guanxi, now reads as an early signal of something more structural: China was never going to let a foreign company own its living room.

The most striking absence is any serious engagement with the Chinese state as something other than a relationship to be managed. The Ministry of Education is a partner. Local officials are doors to be opened. The Communist Party is background furniture. This was the standard Western business book posture of the era — the same posture that informed a generation of "China strategy" consulting — and it aged the way milk ages. Kai-Fu Lee's "Making It in China" report, described here as a blueprint for foreign companies, reads now less like strategic wisdom and more like a time capsule from the last years of mutual optimism. Lee himself would go on to write "AI Superpowers" in 2018, a book that takes the competition between the U.S. and China as its foundational premise rather than its unthinkable conclusion. Guanxi is the before picture. The relationship it documents — between American capital, Chinese talent, and a shared belief in technology as a solvent for political difference — was real. It was also temporary. The book sits in the corpus alongside works like James McGregor's "One Billion Customers" and Clyde Prestowitz's "Three Billion New Capitalists" as artifacts of a specific moment when the West believed engagement was a one-way ratchet.

If guanxi is a system of reciprocal obligation built on trust, and if the trust between American technology firms and the Chinese state has been not merely eroded but actively dismantled by both sides, then the question this book now raises is one its authors would have found absurd in 2006: Was the guanxi ever real, or was it always a transaction dressed in the language of relationship — and if so, who was deceiving whom?