China's University Patent Boom Signals Deeper Tech Threat Than DeepSeek

The story of American technological dominance has long centered on a handful of corporate giants—IBM, then Microsoft, then Google and Nvidia. China's rise, by that same logic, should follow the same script: Huawei, Tencent, and other state-backed champions racing ahead. But a sweeping examination of nearly 14 million patents filed across critical technology fields tells a strikingly different tale, one that suggests the real challenge to American leadership runs far deeper than any single breakthrough like DeepSeek.

Harvard Business School researcher Josh Lerner and his co-authors set out to map who was actually generating patents in the ten technology areas the Pentagon considers vital to national security—advanced computing, space, artificial intelligence, hypersonics, biotech, and others. What they found inverted the expected pattern. Chinese universities account for more than a quarter of their country's critical technology patents, a rate eight times higher than the 3.3% share held by American universities. State-owned enterprises and government bodies combined contribute just 4% of China's total. Fewer than one in ten Chinese patents in these fields involved an inventor with prior work experience in the United States, undermining the long-held assumption that Beijing's innovation engine runs on returning American talent.

Lerner described his team's initial expectation as a kind of intellectual muscle memory. They anticipated finding innovation concentrated in a few recognizable names, the way Samsung and IBM once dominated American patent filings. Instead, the data revealed something far more distributed. "Much more diverse—spread out across the parties," Lerner told Fortune, with universities "more represented" in critical technologies than anywhere else in China's patent ecosystem. In the United States, by contrast, university patents remain peripheral to the main story of technological innovation. The structural difference matters enormously: China has built a system where academic institutions are central to breakthrough research in fields the military cares about most. America has not.

The quality gap that many assumed would protect American technological leadership appears to be closing. Using four separate measures of patent novelty and scientific relevance, Lerner's team found no evidence that China's explosive growth in patenting came at the expense of rigor. Chinese patents have converged with American ones and, in some critical technologies, begun to surpass them. The share of Chinese patents containing what researchers call "disruptive" technical language—language indicating genuinely novel approaches—jumped from under 2% in the early 1990s to 16% by 2022, approaching the American share of roughly 20%. DeepSeek, Moonshot, and other Chinese AI models are no longer trailing American frontier labs; they are competing directly and winning on cost.

The structural difference between how China and the United States organize research funding reveals why this shift is happening. Beijing operates what researchers describe as a "coordinative" system: the state steers research through funding priorities and mandates but does not directly own the universities producing the work. The arrangement allows for alignment without the friction that characterizes American university-government relations. Lerner pointed to Harvard's own protracted legal battles with the Trump administration over research funding as a contrast. "One doesn't suspect the same kind of interactions would characterize those between a major Chinese university and the state," he said.

Meanwhile, the American research funding pipeline has contracted sharply. The Trump administration cut $4 billion from the National Institutes of Health's reimbursement of university research overhead costs, capping rates that historically ran 50 to 60 percent at just 15%. By April, the administration had formally terminated 1,392 NIH grants worth $539 million in unspent obligations. The National Science Foundation cancelled or suspended nearly 2,000 awards of its own. The Brennan Center for Justice estimates total cuts and freezes across both agencies at more than $3 billion since 2025, with roughly $1.4 billion still frozen or unresolved. Some terminations have been reversed by courts, but the disputes continue case by case. The administration has proposed further cuts and new rules allowing agencies to cancel research awards mid-stream—a practice critics warn would destabilize long-term research programs.

The consequences of these cuts may not be immediately visible. Lerner cautioned that the lag between research funding and actual innovation can stretch across years or decades. But the rankings are already shifting. In June, Harvard lost the top spot in the Nature Index Research Leaders—a closely watched measure of high-quality research output—for the first time since the index began tracking universities in 2015. Zhejiang University, the alma mater of DeepSeek founder Liang Wenfeng, moved into first place. Harvard fell to second, or third if you include the Chinese Academy of Sciences. Nine of the top ten academic institutions globally are now Chinese; five years ago, eight of the top ten were American.

Dan Wang, a China analyst at Eurasia Group, urges caution about reading too much into patent counts alone. Chinese university patents have low commercialization rates and are rarely filed overseas, he notes, while American companies like Nvidia and Google continue to drive exceptionally strong research and development. But Wang also warns that Washington's strategy of sanctioning individual "national champions" like Huawei and Tencent misses the real foundation of China's technological rise: a decentralized network of universities, suppliers, and state-owned enterprises. Wang estimates there are 18,000 state-designated "little giant" firms working on components and materials alone. "Sanctioning a few major firms alone is unlikely to halt the broader ecosystem," he said.