Matt Stoller delivers a jarring correction to the prevailing narrative on the U.S.-China tech race: the gap isn't just about who has the most powerful chips, but who has built a system that forces efficiency over excess. While the White House focuses on export controls and subsidies, Stoller argues that China's real advantage lies in a fragmented, hyper-competitive ecosystem that extracts far more intelligence per unit of compute than America's concentrated oligopoly. For busy leaders tracking global supply chains, this piece reframes the AI race not as a hardware contest, but as a clash between two incompatible economic philosophies.
The Architecture of Monopoly
Stoller begins by dismantling the idea that China's rise is solely a result of state aggression or military posturing. Instead, he points to a fundamental structural divergence in how capital is deployed. "Chinese consumption rates are at roughly 40% of GDP, with about that same amount going to investment. Those are insane levels," Stoller writes, highlighting a system where the fruits of production are funneled back into industrial capacity rather than consumer wallets. This isn't just about cheap goods; it is a deliberate strategy to create massive, redundant industrial capacity that the rest of the world cannot match.
The author draws a sharp parallel to the rare-earth element crisis, noting how China's ability to cut off supplies revealed a deep vulnerability in the global supply chain. "We saw this most obviously when China cut off rare earth magnets to the U.S. and the rest of the world last year, but the vulnerability exists across many sectors," Stoller notes. This historical context is crucial; it suggests that the current friction isn't a sudden spike in tension but the culmination of a long-term strategy to monopolize the industrial base. Critics might argue that this view overlooks the genuine demand for Chinese goods in a global market hungry for affordable solar panels and electric vehicles. However, Stoller counters that this short-term benefit masks a long-term predatory reality: "As with Amazon offering cheap prices until it monopolized the e-commerce market, this short-term strategy will not work well for anyone."
The Chinese system is organized around having massive levels of unnecessary industrial capacity, which it turns towards waste and exports.
The Efficiency Moat
The most compelling section of the piece shifts from macroeconomics to the specific mechanics of artificial intelligence. Stoller introduces the concept of the "efficiency moat," a term borrowed from analyst Azeem Azhar, to explain why Chinese models are closing the performance gap despite having less computing power. "Chinese labs are extracting 4-7x more intelligence per unit of compute than naive scaling predictions would suggest," Stoller writes, citing data that shows Chinese firms achieving comparable results at a fraction of the cost.
This argument challenges the assumption that the U.S. lead in hardware automatically translates to a lead in software and application. Stoller explains that the U.S. market is dominated by five key frontier labs—OpenAI, Anthropic, Google DeepMind, Meta, and xAI—which operate in a closed, vertically integrated manner. In contrast, "In China, a thousand flowers are blooming," he observes. This fragmentation forces Chinese companies to innovate on efficiency because they cannot rely on brute-force compute. The result is a pricing structure that is nearly impossible for American firms to compete with: "DeepSeek charges $0.43 per million input tokens... Opus 4.6 is 11 times more costly in input and 28 times more expensive in output."
Stoller connects this to a broader historical trend, noting that the U.S. has lost the very ecosystem that once made it a manufacturing powerhouse. He references the 19th-century machine tooling shops in Boston that allowed innovators like Alexander Graham Bell to build the telephone, contrasting that with today's reality where "America has virtually no machine tooling anymore, whereas China leads the world." This loss of physical infrastructure has created a feedback loop where American intellectual property is locked away by dominant firms, stifling the very innovation the country claims to protect. "Today, American intellectual property is locked into dominant firms, who spend large amounts of money making sure no one else can use it," Stoller argues, pointing out that Apple spends $1 billion a year on its legal division to maintain these barriers.
The Policy Paradox
The commentary concludes by exposing the contradiction in current U.S. policy. The administration is simultaneously trying to protect a lead in hardware while ignoring the fact that the software and application layer is becoming more efficient elsewhere. "The first is that Western AI chips... are so much better at mass compute... and the second is that China is nipping at our heels with its own artificial intelligence models," Stoller writes, highlighting the cognitive dissonance in Washington. The solution, he implies, isn't just more subsidies for chip factories, but a fundamental restructuring of competition policy to break up the vertical integration that stifles new entrants.
Stoller's framing of the "Little Giants" program—where the Chinese government supports tens of thousands of small businesses with preferential loans and patent processing—offers a stark contrast to the U.S. approach. "Once they obtain the coveted 'Little Giants' designation, they receive access to an impressive package: preferential loans, direct subsidies, state equity investment, quick patent processing," he details. This creates an environment where innovation is decentralized and rapid, rather than concentrated and slow.
Critics might note that relying on state-directed industrial policy carries its own risks, including inefficiency and the potential for corruption. However, Stoller's evidence suggests that the current American model of financialization has already produced a different kind of inefficiency: a system where companies become more valuable on paper but worse at building things. "In the U.S., it's the opposite. As our companies have become more financially valuable, they have become worse at building things," he asserts.
Our venture capitalists and entrepreneurs shy away from competing with giants or accidentally stepping on patents.
Bottom Line
Stoller's strongest contribution is his identification of the "efficiency moat" as the critical vulnerability in the U.S. tech strategy, shifting the debate from hardware supremacy to systemic adaptability. The piece's greatest risk is its reliance on a binary view of economic systems that may oversimplify the complexities of global trade, yet the data on pricing and compute efficiency is undeniable. Leaders should watch for whether U.S. antitrust enforcement can evolve quickly enough to dismantle the vertical integration that is currently costing the nation its competitive edge in the next generation of technology.