Dominance Is a Dashboard, Not a Crown

The title of the Fresh Air interview is almost irresistible: Has the United States lost its dominance to China? It promises one event, one possession, one answer. Evan Osnos’s reporting supplies vivid material for the concern: an electric-vehicle factory that feels almost unpeopled, a research system gaining mass, and a country that can coordinate investment at a scale the United States often struggles to match. The strongest version of his warning is not that China has won everything. It is that American confidence was built on advantages that no longer look permanent.

I wanted to know what happens when the noun dominance is forced to declare its unit. Is it a share of factory value added, a supply-chain bottleneck, the ability to finance discovery, publication volume, highly cited science, currency use, military spending, standards, alliances, or the freedom to absorb criticism? Each has a source and a denominator. They do not share a conversion rate. A factory cannot be divided by a reserve currency; a publication count cannot be added to a carrier group.

This Notebook therefore treats NPR and The New Yorker as initiating reported interpretation, not as the final numerical authority. Every dashboard value comes from an official statistical body, an intergovernmental or specialist institution, or peer-reviewed research with a disclosed method. The purpose is not to drain the reporting of urgency. It is to find exactly where the urgency is earned.

The shortest defensible answer is no: the reviewed evidence does not establish a singular transfer of global dominance. It establishes a much more consequential change. China has become the central manufacturing system for the energy transition and a peer-scale research power. In some production chains, the question is no longer whether China leads but whether the rest of the world can function through a disruption in Chinese processing or manufacturing.

The United States, meanwhile, is not a former power waiting for a verdict. Its economy remains much larger at current exchange rates. The dollar accounted for 57.13 percent of disclosed allocated reserves in the first quarter of 2026, compared with 1.99 percent for the renminbi. SIPRI estimates US military spending at $954 billion in 2025, against $336 billion for China. Stanford records $285.9 billion in US private AI investment in 2025, although its China figure misses important public-guidance capital. Those are not decorative legacies. They finance capabilities, alliances, research organizations, and policy options.

But the old American model (lead the invention, outsource much of the production, and assume high-value control will remain at home) looks less secure when learning, supplier density, engineering, standards, and iteration accumulate around the factory. The dashboard does not crown China. It removes the excuse to treat production as a lower-order activity that follows innovation rather than shaping it.

UNIDO’s 2025 yearbook supplies the cleanest wide-angle measure. China’s share of global manufacturing value added rose from 2.8 percent in 1990 to an estimated 32.0 percent in 2024. The United States stood at 15.0 percent. Japan, Germany, and South Korea added another 14.2 percent; even those four economies together remained below China. That is an extraordinary structural fact, but value added still does not reveal who owns equipment, controls software, earns profits, or can replace a missing component under pressure.

The IEA makes the concentration concrete. China produced nearly three quarters of electric cars in 2025, more than four fifths of battery cells, and about 85 percent of solar-PV supply-chain capacity. Its share reaches roughly 95 percent at the wafer stage. The same pattern appears upstream: China is the top refiner with more than 90 percent shares for gallium, graphite, manganese, and rare earth elements in the IEA’s review. These are different denominators, so the page refuses the tempting but meaningless act of averaging them into one clean-China percentage.

Concentration becomes geopolitical power when a buyer cannot switch quickly. The IEA estimates that a full implementation of expanded rare-earth export measures could put $6.5 trillion of downstream production outside China at risk. That is a modeled exposure scenario, not a realized loss, and the expanded measures were suspended for a year through November 2026. The distinction matters: a scenario measures vulnerability; it does not prove that Beijing can impose costs without also accelerating substitution or damaging its own suppliers.

Automation complicates the demographic story. The International Federation of Robotics counted 295,000 industrial-robot installations in China in 2024, or 54 percent of the world total, and an operational stock above two million. The United States installed 34,200 and operated roughly 393,700. Robot counts do not equal productivity, but the scale shows why a smaller future labor force does not translate mechanically into proportionally lower output.

The frontier is not one race either. NSF’s 2026 state-of-science report estimates China spent $1.028 trillion in purchasing-power-comparable R&D in 2024 and the United States $1.009 trillion, or 30 and 29 percent of the world total. The values are preliminary. China produced 31 percent of global science and engineering articles against 12 percent for the United States. Yet the United States retained disproportionate representation among highly cited work and accounted for 31 percent of internationally coauthored articles. Scale and network position are both real advantages.

Biopharma offers a particularly sharp transition. A March 2026 JAMA study using Clarivate’s Cortellis database found China-originated early-stage programs grew from 829 in 2015 to 6,145 in 2024, while US programs rose from 5,024 to 7,107. China’s share moved from 8.0 to 32.3 percent; the US share fell from 48.2 to 37.4 percent. In biologics, the 2024 shares were 35.1 and 35.7 percent. In cancer, China was slightly ahead. But the unit is an early-stage program attributed by originator headquarters, not a successful medicine. The study did not assess quality or approvals.

AI again splits the verdict. Stanford reports that the aggregate performance gap between the top US and Chinese models narrowed to 2.7 percent by March 2026. China led publication, citation, patent-output, and industrial-robot measures, while the United States led top-tier model production, higher-impact patents, and private AI investment. The private-capital gap is spectacular in the recorded series, but Stanford explicitly warns that Chinese government guidance funds are not captured. A good dashboard carries that caveat next to the number, not in a footnote designed to be forgotten.

The scientist-migration claim needs similar care. The PNAS study found nearly 20,000 scientists of Chinese descent left the United States for other countries between 2010 and 2021, with annual departures rising from about 900 to 2,621. China became a more common destination, but the nearly-20,000 total includes moves elsewhere. The paper connects departures to fear about US research conditions as well as family and opportunity. Talent security is therefore not only a visa or counterintelligence problem. It is also the daily institutional question of whether scientists believe they can work, collaborate, and belong.

Industrial concentration, military capacity, and monetary reach impose different kinds of consequences. A mineral-processing bottleneck can interrupt a factory. A navy can deter or coerce in a particular geography. A reserve currency can lower financing costs, extend sanctions reach, and make global institutions responsive to domestic financial conditions. None reliably substitutes for the others.

SIPRI’s estimates keep the military scale visible: $954 billion for the United States in 2025 and $336 billion for China. The US total was 33 percent of global expenditure; China’s was 12 percent after thirty-one consecutive increases. Spending is still an input. Geography means China can concentrate forces near its coast while the United States funds a global posture; alliances and operating experience further complicate direct comparisons. The right claim is that the United States retains enormous military resource superiority, not that every contingency therefore has a predetermined result.

The dollar-renminbi comparison is more lopsided. IMF COFER data put the dollar at 57.13 percent of allocated official reserves in 2026’s first quarter and the renminbi at 1.99 percent. Valuation changes moved the shares, and reserve disclosure is incomplete, but there is no evidence here of a near-term monetary handoff. China’s trade and production network can grow faster than use of its currency because convertibility, institutions, market depth, trust, and policy control are part of monetary power.

The practical lesson is to ask not only who has more, but who can turn a stock into collective action without destroying the conditions that made it valuable. A chokepoint loses value as others redesign around it. Financial sanctions invite alternatives. Military threats change alliance behavior. Power is exercised inside feedback loops.

China’s demographic constraint is concrete. The National Bureau of Statistics reported a 2025 population of 1.40489 billion, down 3.39 million in a year, with 7.92 million births and 11.31 million deaths. People aged sixty or older made up 23.0 percent; ages sixteen through fifty-nine, 60.6 percent. That means fewer new workers, heavier pension and health burdens, and pressure on a growth model accustomed to abundant labor and housing demand.

The comparison guards against a common visual error. The United States is smaller and still grew by about 1.8 million from July 2024 to July 2025, with net international migration contributing about 1.3 million. Yet when Census single-year estimates are summed into the same display bands, 25.1 percent of the US population is sixty or older, higher than China’s official 23.0 percent. The systems are not identical: Census and NBS definitions, reference dates, citizenship and resident coverage, and migration regimes differ. The point is not that the United States has the worse problem. It is that age share, total scale, movement, fertility, and policy have to be read together.

China still has roughly 851 million people in the official sixteen-to-fifty-nine band. The challenge is the direction and the dependency ratio, not sudden labor disappearance. Automation can raise output per worker but may concentrate returns, demand expensive capital, and leave care work unresolved. Migration can extend US growth but depends on a political and institutional capacity to attract and integrate people. Demography sets constraints; governance determines much of the response.

The timeline begins before the factories because today’s dashboard was built through institutional choices. The PRC’s 1971 UN representation, reform and opening from 1978, US normalization in 1979, and WTO accession in 2001 progressively changed China’s access to markets, capital, technology, and diplomatic arenas. None was a straight line toward convergence. Political control persisted while economic methods changed.

Made in China 2025 formalized the next ambition: move from high-volume production toward advanced equipment, quality, integration, and domestic innovation. The 2018 Section 301 tariffs marked a US turn toward explicit technology and supply-chain rivalry. The 2022 CHIPS and Science Act marked another turn, from diagnosing dependence to subsidizing capacity and research at home. Plans and statutes are only beginnings; delivered factories, workers, research, costs, and spillovers determine whether they change the balance.

By 2024, the industrial lead and science near-parity could be seen in the same frame. By 2025, export controls showed how concentration could be deployed as leverage. By August 2026, a mainstream American interview could reasonably ask whether dominance had been lost. The question’s arrival is historically meaningful. Its binary form is still analytically inadequate.

I began with an image of competition between two national stacks: factories on one side, finance and frontier invention on the other. The sources replaced that image with overlapping systems. Chinese industry contains foreign equipment, overseas demand, and knowledge accumulated through global integration. American labs, firms, and markets depend on Chinese-origin talent, processing, components, and manufacturing feedback. The competition is real, but interdependence is not a footnote to it; interdependence is the terrain on which it is fought.

I also understand the manufacturing argument more strongly after auditing it. The hype does not come from inventing the direction. It comes from compressing several already-extraordinary facts into one totalizing percentage. Disaggregated, the record is more useful: 32 percent of manufacturing value added, nearly 75 percent of electric-car production, more than 80 percent of battery cells, about 85 percent of solar capacity, and more than 90 percent in specified refining chains. Each can be monitored and challenged.

What I reject is the crown. A crown makes power singular, stable, and transferable. The dashboard shows a plural, conditional balance: China’s production ecosystem can set costs and pace; American capital, currency, military resources, research institutions, and alliances can still shape choices; both face demographic, political, and feedback constraints. The strategic task is not to declare a winner. It is to identify which capacity matters for which outcome, then build resilience without pretending dependence can disappear by slogan.