Political Capital, covered in Letter 170, can be spent and rebuilt within a single news cycle. Knowledge Capital moves on a completely different clock — it accumulates over decades of research funding and erodes just as slowly, which is exactly what makes the 2024–2026 data on global R&D spending and patent output worth taking seriously: a shift this large, this quietly, rarely reverses within an investing horizon.
For most of the postwar era, American R&D leadership was close to uncontested — the assumption underneath a huge share of global technology valuations. That assumption did not survive 2024 intact, and the 2026 data confirms the shift was not a blip.
China's R&D spending, measured in purchasing-power-parity terms, reached an estimated $785.9 billion in 2024 against $781.8 billion for the United States — a gap so narrow that, depending on methodology, China may already be ahead. What makes the number significant isn't the crossover itself so much as the trajectory: China's share of global R&D spending rose from 4.0% in 2000 to 27.4% in 2024. Over the past decade, Chinese firms increased R&D spending by 537%, against 150% growth for US firms and just 32% for the rest of the world combined. On a size-adjusted basis, US firms out-invested Chinese counterparts in eight of nine advanced industries as recently as 2014; by 2024, the US had lost its lead in four of those nine.
The patent data tells the same story with a wider margin. Of 3.7 million patent applications filed worldwide in 2024 — a record — China's IP office received roughly 1.8 million, compared to about 603,000 at the US Patent and Trademark Office, a gap of roughly three to one. In 2024, China also overtook the United States in total scientific publications for the first time since the US itself overtook the United Kingdom in 1948, and posted a 17% citation advantage over the US in the Nature Index, which tracks publication in the most selective journals in the world.
The mechanism behind the shift is less a Chinese acceleration than an American deceleration in one specific category. US federal R&D spending peaked at roughly $160 billion in 2010 and fell more than 15% over the following five years, even as private-sector R&D — which is overwhelmingly commercially near-term — kept growing. China's model runs the opposite way: a large share of its research flows through centrally funded government labs directed at national priorities including energy, biotech, and frontier computing, insulating basic research from the funding cycles that have squeezed US federal science agencies in recent years.
Patents and papers are lagging indicators of research funded five to ten years earlier. The R&D spending crossover of 2024 is the leading indicator for a patent and citation gap that is still widening today — which means the full economic consequences of this shift haven't been priced yet, in either direction.
The honest complication is that raw patent and paper counts overstate the case in China's favor — filing volume is influenced by domestic incentive structures as much as by underlying research quality, and the US retains real structural strengths: businesses perform 77% and fund 75% of US R&D, giving it a commercialization pipeline from discovery to market that remains faster than China's in most sectors, pharmaceuticals and biotech included, where nearly $140 billion in US business R&D flowed in 2023 alone. The risk for investors is not that China has "won" the innovation race — it's that the compounding advantage the US market has priced into decades of American technology valuations is eroding in specific, identifiable sectors (electric propulsion patents, where China's global share rose from 2.4% to 26.9% between 2010 and 2020, is the sharpest single example), while the broader market narrative has not caught up to the sector-level data.
Knowledge Capital is the slowest-moving asset in this series and, for that reason, the one most likely to be mispriced by markets that reward quarterly news over decade-long funding trends. The next letter turns to the capital this deficit feeds most directly: AI Capital, where compute access is now determined in large part by exactly the research and patent base this letter measures.
The 2024 R&D spending crossover between China and the US is real, verified across multiple independent methodologies, and not a temporary artifact of exchange rates or accounting convention — it reflects a decade of diverging federal science funding decisions on one side and centrally directed research investment on the other. It does not mean the US has lost its innovation edge broadly; commercialization speed and business-sector R&D intensity remain genuine American advantages. It does mean that sector-by-sector, the compounding research advantage priced into a generation of technology and biotech valuations needs re-underwriting on a case-by-case basis, not assumed forward from history.