The myth about China’s tech rise remains stubborn and simple: “It’s all government subsidies and copycat tech.” The reality, measured in hard numbers, is structural and cumulative. The Science and Technology Indicators 2026 report — issued by the National Institute of Science and Technology Policy in Tokyo — shows that China’s combined corporate and academic research and development spending reached 97.1 trillion Japanese yen ($614 billion) in 2024, surpassing the United States (95.3 trillion yen) for the first time.
Chinese companies alone accounted for 75.4 trillion yen of that total, a 13 percent year-on-year increase. China’s official figures for 2025 confirm the trend: National R&D expenditure hit 3.9 trillion yuan ($584.4 billion), up 8.2 percent, with intensity rising to 2.8 percent of GDP — the first time it has exceeded the Organization for Economic Cooperation and Development average of 2.72 percent. Enterprises contributed 77.4 percent of the total, crossing the three-trillion-yuan threshold for the first time and remaining above 77 percent for four consecutive years.
These are not abstract aggregates. They reflect a sustained decision, repeated across thousands of firms and multiple cities, to treat R&D expenditure as a core operating cost rather than an optional expense. Industrial policy may set priorities, but the decisive variable has been the willingness of companies — especially private and mixed-ownership ones — to keep writing large checks for talent, equipment, and iterative engineering over long periods. There are no shortcuts in deep technology. Every sustained percentage point of GDP devoted to R&D, every firm that chooses to reinvest 15 to 25 percent of revenue into laboratories rather than short-term profits, adds another layer of real capability.
The enterprise engine
The shift toward enterprise-led R&D is now structural. In 2025, Chinese companies spent more than 3 trillion yuan on research, contributing 73.7 percent of the overall increase. This share approaches the levels seen in the US and Japan and exceeded those of several major European economies. Growth has been especially strong in computers, electronics, optical products, and high-tech manufacturing — precisely the sectors most affected by export controls since 2022. Rather than retreating, firms accelerated domestic investment in restricted domains. Basic research has also accelerated, rising 13.3 percent in 2025 and lifting its share of total R&D to 7.21 percent. Universities and government institutes remain important, yet the commercial sector now supplies the bulk of both funding and applied output. The Japanese data and Chinese official statistics align on this point: The overtaking of the US was driven primarily by corporate spending, not by a sudden surge in pure academic budgets.
Shenzhen: intensity at city scale
No place illustrates the pattern more clearly than Shenzhen. In 2024 the city’s total R&D expenditure reached 245.3 billion yuan. Intensity hit 6.67 percent of local GDP — the highest among China’s major cities, ahead of both Beijing and Shanghai. Enterprise spending accounted for 93.5 percent of the total. Over the preceding decade, intensity rose steadily from under 4 percent through successive thresholds of 5 percent and then 6 percent. The city now hosts more than 25,000 national high-tech enterprises and leads the nation in the number of specialized “little giant” firms, with the total reaching approximately 1,333 by late 2025. It has ranked first among Chinese cities in new “little giant” designations for three consecutive years. Specialized small and medium-sized enterprises number well over 12,000.
These figures are not bureaucratic tallies; they reflect a dense ecosystem in which private firms treat R&D as nonnegotiable. Local policy reinforces the pattern by directing the majority of science and technology fiscal support toward enterprises, yet the underlying driver remains commercial decision-making. Shenzhen’s ability to move rapidly from prototype to volume production is well known. That speed rests on years of prior capital allocation to engineering talent, testing infrastructure, process knowledge, and supplier networks. The city’s innovation intensity is measurable in the density of firms that survive and scale precisely because they keep investing through cycles.
Company cases: capital allocation as strategy
Individual firms make the aggregate numbers concrete. Technology conglomerate Huawei continues to set a high bar. In 2025 it spent 192.3 billion yuan on R&D — 21.8 percent of revenue and nearly three times its net profit that year. In the first half of 2026 alone the figure reached 121.4 billion yuan, equivalent to more than 670 million yuan per day and a research intensity of 25.9 percent. Cumulative investment over nearly two decades exceeds 1.8 trillion yuan. More than half the workforce is engaged in R&D. The company has maintained research intensity above 20 percent for several consecutive years even under sustained external pressure.
Automotive manufacturer BYD provides another clear illustration. In 2024 its R&D expenditure reached approximately 54.2 billion yuan; in 2025 the figure rose to roughly 63.4 billion yuan, placing it at the top of A-share listed companies for research spending. The company employs more than 120,000 research personnel. For most of the past 14 years BYD has spent more on R&D in a given year than it reported in net profit. The investments have underpinned successive generations of battery technology, powertrains, intelligent driving systems, and manufacturing processes that have reshaped global electric-vehicle supply chains.
In semiconductors, Semiconductor Manufacturing International Corp recorded R&D spending of approximately $774 million in 2025, equivalent to about 8.3 percent of revenue. The company has steadily expanded its patent portfolio while scaling capacity under difficult external conditions. Across the broader Chinese semiconductor sector, listed firms collectively spent tens of billions of yuan on research in recent years, with intensity often exceeding 10 percent of revenue.
DJI, the global leader in consumer and professional drones, has long devoted a substantial share of its workforce — roughly one-quarter — to product development. Its continued expansion of research facilities in Shenzhen, including large new campuses dedicated to intelligent aviation systems, reflects the same underlying logic: Sustained technical investment is the foundation of market position.
Similar patterns appear at Contemporary Amperex Technology Co Ltd (CATL), the global leader in EV batteries, where patent leadership in power and solid-state technologies has accompanied large cumulative R&D outlays, and at numerous specialized equipment, materials, and industrial artificial intelligence firms concentrated in the Guangdong-Hong Kong-Macao Greater Bay Area and other high-tech clusters.
These examples are not isolated. They represent a widespread corporate practice of treating research intensity as a strategic necessity rather than a discretionary cost.
Output metrics: from papers to patents
Quantity of spending is only one dimension. China already led the world in total scientific papers by 2018, in the top 10 percent of highly cited papers by 2018, and in the top 1 percent of elite papers by 2019. More recent patent data reinforce the shift. China accounts for a large share of global AI patent grants and leads in numerous critical technology domains tracked by independent assessments. In batteries, CATL has overtaken previous leaders in patent families. High-value invention patents and next-generation technology patents continue to rise as a share of the national total.
Citation and quality metrics remain imperfect, and not every patent or paper translates into commercial advantage. Yet the direction of travel is consistent with the spending data: Earlier gains in research output are now being matched and reinforced by capital intensity, particularly in domains under external constraint.
Pressure as accelerator
Export controls imposed since 2022 did not reduce Chinese R&D in restricted fields. They increased it. Spending in hardware, optics, semiconductors, and related areas rose as firms sought to reduce vulnerability. The response has been capital-intensive and multiyear. Building parallel capabilities under constraint is expensive and slow — the opposite of a “copycat” strategy, which is cheap and fast when open access to foreign technology exists.
Core mechanism
Industrial policy in China is real. It identifies priority sectors and mobilizes resources. The decisive long-term variable, however, has been the repeated choice by thousands of companies to allocate large and growing shares of revenue to R&D. In Shenzhen, that choice has been institutionalized at city scale, with enterprises supplying more than 90 percent of the R&D funds. Nationally, the enterprise share of total R&D has risen to levels comparable with the most advanced economies.
There are no shortcuts in deep technology. Process knowledge, materials science, systems engineering, and the feedback loops between laboratory and factory all require sustained funding over long horizons.
The Japanese report, Chinese official statistics, corporate financial disclosures, and patent and publication metrics converge on the same conclusion: China’s technological position is the cumulative result of continuous, large-scale investment in the people and facilities that generate new knowledge and convert it into products. The myth of subsidies and imitation no longer fits the evidence.
The author is chief economic adviser at the Shenzhen Qianhai Authority.
The views do not necessarily reflect those of China Daily.
