China’s Research and Innovation System

Where it stands

• Planned as one system. Science, education, talent and industry are steered together from the top, through five-year plans and since 2023 a Party commission for science and technology.
• Turning towards basic research. After missing its 8% target, the leadership is raising basic research budgets (central spending +16.3% in 2026) and asks for more diverse funding and separate evaluation for basic research.
• Industry leads the spending. Firms carry three quarters of R&D, and policy ties research to strategic and future industries, chips and the diffusion of AI.
• Output has overtaken. China leads on volume, Nature Index and most critical technologies, while the US still leads on elite researchers, cross-country citations and frontier AI investment.
• Integrity and evaluation remain weak points. Reforms against paper counting are years old, but practice lags and a crackdown on misconduct is under way.
• The world around it is narrowing. International collaboration is falling, and technology controls run in both directions.

1. Strategy and planning

Long plans, set at the top and followed through, are the backbone of the system. The current frame is the 15th Five-Year Plan (2026–30) and the goal of a “strong science and technology nation” by 2035.

• 15th Five-Year Plan, adopted 12 March 2026: society-wide R&D to grow more than 7% a year, high-value invention patents to over 22 per 10,000 people, with “high-level S&T self-reliance” leading “new quality productive forces”. It sets no target for the basic research share (SCIO).
• Future industries named in the plan: quantum technology, biomanufacturing, hydrogen and fusion energy, brain-computer interfaces, embodied AI and 6G (Xinhua).
• Basic research priorities for the plan period, set out by MOST on 29 September 2026: mathematics, control of states of matter, synthetic biology, the materials genome, controlled fusion and “intelligent scientist systems” (gov.cn).
• Xi Jinping’s basic research symposium in Shanghai, 30 April 2026, called for a larger share of funding for basic research, more diverse funding sources and a separate evaluation system suited to basic research (CPPCC/Xinhua).
• 2035 goal: a “strong S&T country”, set at the June 2024 national science conference, where Xi acknowledged weaknesses in original innovation (gov.cn); the 2024–2035 Education Powerhouse Plan supplies the education and talent leg (CSET translation).
• The “new whole-nation system” (2022) pools state, market and society to break through in core technologies, with firms in a leading role (China Daily). The revised S&T Progress Law (2022) added a chapter on basic research (NPC Observer); an AI law is planned but not yet drafted.

2. Governance and institutions

Since 2023 the Party steers science directly, while funding and research are spread across ministries, academies, national laboratories, universities and firms.

• Central Science and Technology Commission (2023): a Party body above the ministries that centralises leadership of S&T; MOST was slimmed down to act as its working body and handed applied programmes to sector ministries (East Asia Forum; Naughton et al.).
• NSFC, the main competitive funder of basic research, under MOST since 2018: a 2025 budget of about 39.5bn yuan and about 880,000 projects in 40 years (CGTN). It funds more than 45,000 projects a year, mostly small grants, with its early-career fund now the largest programme (CFRR).
• Laboratory system: national laboratories at the top, then national research centres and about 500 national key laboratories, re-approved after a 2021–22 reform (China Policy).
• “Strategic S&T forces” in the new plan: national laboratories, national research institutions, research universities and leading tech firms (SCIO). The Chinese Academy of Sciences remains the world’s top institution in the Nature Index.
• “New-type R&D institutions”: independent, mixed-funded institutes run on market lines, promoted since 2019 (CSET translation).

3. Money

China now spends about as much on R&D as the US in purchasing-power terms, but a much smaller share goes to basic research, and most money is spent by firms on development.

• R&D 2025 (final): 3,928.8bn yuan, up 8.2%, 2.80% of GDP. Firms spent 77.4%. Basic research 283.5bn yuan, 7.21% of R&D, up from 6.9% in 2024; the preliminary figure was 7.08% (NBS/MOST/MOF).
• Against the US: OECD finds China’s R&D passed the US in purchasing-power terms in 2024 (OECD); at exchange rates it is about half (Nature). Fundamental research spending was about 23% of the US level and 0.17% of GDP in 2023, against 0.50% in the US (CFRR).
• Central government S&T budget 2026: 426.4bn yuan, reported as 10% up, while total central spending grew 5.5%. Central basic research rises 16.3% to 116.9bn yuan; applied research falls about 13% (Yicai).
• The basic research share missed the 14th plan’s 8% target, and the new plan sets none (Nature). Leading OECD countries spend 15–20%.

4. Industrial strategy

Research is tied explicitly to industrial and security goals: self-reliance in key technologies, strategic and “future” industries, and the diffusion of AI across the economy.

• Made in China 2025 (2015–25): an SCMP analysis found more than 86% of some 260 targets met, with electric vehicles and renewables far ahead and advanced lithography, wide-body aircraft and some materials behind (Northern Trust).
• “AI+” action plan (August 2025): over 70% adoption of intelligent terminals and agents by 2027 and over 90% by 2030, with AI for science first among six areas (Caixin).
• Semiconductors: the third state “Big Fund” (2024) has 344bn yuan of capital (SCMP); SMIC makes 7nm chips without EUV, and US approval of Nvidia H200 sales in January 2026 stalled in reviews on both sides (CNBC).
• Humanoid robots: MIIT reports 2025 as the first year of mass production, with more than 140 manufacturers (China Daily). Biotech: China-origin assets were 32% of global biopharma out-licensing deal value in the first half of 2025, up from 5% in 2020 (PharmaSource).
• The research–industry link: firms carry three quarters of R&D but only about 9% of basic research (CFRR); MOST’s line for the new plan is to plan “science and industry together”.

5. People and talent

The workforce is large and young, density is still low, and talent policy works on both training at home and recruiting from abroad.

• 3.00 million researchers (full-time equivalents, 2023), but 4.1 per 1,000 employed against 16–17 in Denmark, Sweden, Finland and Korea; 639,280 doctoral students enrolled in 2024 (CFRR).
• Early-career support: about 80% of researchers starting careers in 2019 had early work acknowledging NSFC, against about 50% of the 2000 cohort (CFRR).
• Recruiting from abroad: the Qiming programme, successor to Thousand Talents, targets chip and other sensitive fields (Reuters via Context); a K visa for young foreign STEM talent since October 2025 (China Briefing).
• Flows are shifting: high-profile moves from US universities to China are rising (Inside Higher Ed), while most top Chinese-origin AI researchers remain in the US; 69% of China-trained NeurIPS authors now stay in China (Nature).
• Gender: NSFC has raised age limits for women in some schemes, but no current national figure for women’s share of researchers was found.

6. Assessment, integrity and openness

Reforms aim to move away from paper counts, while integrity problems and a push for Chinese journals run alongside.

• Evaluation reform: the 2018 policy against judging researchers only on papers, titles, diplomas and awards, and the 2020 rules against “SCI supremacy” (HKUST Library). Paper counts still drive hiring and promotion at many universities (Nature).
• Integrity crackdown: NSFC named 31 misconduct cases in August 2026, and MOST announced a national database of serious violations in September (Nature; SCIO). China-affiliated authors appear on about 60% of retracted papers in 2014–24, partly a sign of scrutiny (Chemistry World).
• Journals: a national plan since 2019 builds Chinese STM journals; in December 2025, 70 new English-language titles were named in fields such as AI and quantum (HIT).
• Open data and NSFC review: 20 national scientific data centres hold over 270 PB; NSFC sorts proposals as curiosity-driven or goal-oriented and pilots double-blind and “non-consensus” review, none of it yet evaluated (CFRR).

7. Outputs and global position

On volume and many quality measures China now leads; on field-weighted impact, elite researchers and frontier models the picture is closer.

• Nature Index 2026: China’s share is about twice the US’s, and nine of the top ten institutions are Chinese (Xinhua).
• Critical technologies: ASPI finds China leading research in 69 of 74 technologies, 41 with “high monopoly risk” (ASPI).
• Patents: 73,718 international (PCT) applications in 2025, ahead of the US with 52,617 (WIPO).
• The other side: China’s share of Highly Cited Researchers fell slightly to 19.7% in 2025, against 37.4% for the US (Clarivate); excluding domestic citations, China receives 9.5% of cross-country citations against 24.9% for the US (CFRR).

8. International and security

Collaboration is narrowing, technology controls run in both directions, and China builds its own networks and governance initiatives.

• Collaboration: China’s international co-authorship rate is about 20% (2024), down from a peak around 2018 (CFRR); only about a quarter of its critical-technology collaborations now involve the US (Bloomberg Law).
• US–China S&T Agreement: renewed in December 2024 for basic research only, excluding critical and emerging technologies (AIP).
• China’s own controls: rare-earth and magnet export controls, broadened in October 2025 and suspended after the Busan meeting until 10 November 2026 (FDD).
• Networks and governance: Belt and Road science conferences (2023, 2025) and joint labs (CAS); a Global AI Governance Action Plan (2025) (Shanghai government); the ISC’s Beijing Science Dialogue and the open China Fundamental Research Report (October 2026) (China.org.cn).

Close reading: the China Fundamental Research Report 2026

On 20 October 2026 China’s research council NSFC published, with Wuhan University, the first open, system-level account of how China funds, organises and evaluates fundamental research (cfrr.net). It compares 25 systems, Denmark among them, refuses composite rankings and frames research systems in terms of diversity and contribution.

Timeline

1. Mar 2023
   Party-state reform creates the Central Science and Technology Commission and slims down MOST.
2. Sep 2023
   Xi introduces “new quality productive forces”; Huawei launches a phone with a 7nm SMIC chip.
3. May 2024
   Big Fund III registered with 344bn yuan for semiconductors.
4. Jun 2024
   National science conference sets the goal of a strong S&T nation by 2035.
5. Jul 2024
   Third Plenum resolution calls S&T the strategic support of modernisation and promises research-system reform.
6. Dec 2024
   US–China S&T Agreement renewed, narrowed to basic research; China restricts gallium, germanium and antimony exports.
7. Jan 2025
   Education Powerhouse Plan 2024–2035; DeepSeek-R1 released as open source.
8. Aug 2025
   State Council “AI+” action plan.
9. Oct 2025
   K visa for young foreign STEM talent; sweeping rare-earth controls, suspended a month later after the Busan meeting; Fourth Plenum adopts the 15th plan recommendations.
10. Jan 2026
    US approves Nvidia H200 sales to China with conditions; shipments stall.
11. Mar 2026
    NPC adopts the 15th Five-Year Plan; central basic research budget up 16.3%; OECD finds China’s R&D passed the US in purchasing-power terms.
12. 30 Apr 2026
    Xi’s basic research symposium in Shanghai.
13. Jun 2026
    Nature Index 2026: China about twice the US; nine of the top ten institutions.
14. Aug 2026
    NSFC publishes 31 misconduct cases.
15. 29 Sep 2026
    MOST sets out basic research priorities for the plan period and a national integrity database.
16. 16 Oct 2026
    Final 2025 R&D figures: 2.80% of GDP, basic research 7.21%.
17. 19–20 Oct 2026
    ISC Beijing Science Dialogue; launch of the China Fundamental Research Report.
18. 10 Nov 2026 · upcoming
    Suspension of China’s broadened rare-earth export controls due to lapse; an extension is reported but unconfirmed.

How to read it: the main debates

• Scale or originality? China publishes more top-cited papers than any country on simple counts (Wagner, Ohio State), yet its own leadership names original innovation as a weakness and basic research remains about 7% of R&D.
• Direction or discovery? The new whole-nation system centralises priorities, while curiosity-driven categories, non-consensus pilots and calls for diverse funding show that steering alone is not expected to produce breakthroughs (Naughton).
• How to compare money? Purchasing-power figures put China ahead of the US; exchange rates put it at half; the composition gap in basic research matters more than the totals (NSF).
• Innovation or diffusion? Jeffrey Ding argues leadership comes from spreading general-purpose technologies through the economy, and that China has a diffusion deficit (Georgetown); the AI+ targets are in effect a diffusion policy.
• Decoupling or a shared frontier? ASPI frames China’s lead as monopoly risk; others argue for keeping basic research cooperation (CSIS) and the report’s authors pitch it against a race narrative.
• Do evaluation reforms work? Rules against paper counting coexist with paper-count hiring and high retraction shares (Zhang & Sivertsen).

What to watch

• Does the basic research share keep rising without a target, and do the new evaluation rules for basic research reach universities?
• Do NSFC’s review reforms (curiosity-driven categories, double-blind and non-consensus review) get evaluated and published?
• Does the integrity crackdown and national database change incentives, or mainly punish cases?
• How fast do chips, fusion, quantum and embodied AI move from plan targets to production?
• What happens to the rare-earth control suspension after 10 November 2026, and to US–China research ties?
• Do talent flows keep shifting towards China, and does the K visa attract foreign researchers?
• Does the China Fundamental Research Report become an open, annual resource that others can check and build on?

My reading

Sources and method

Figures come from official Chinese releases (National Bureau of Statistics, MOST, Ministry of Finance, Xinhua, SCIO), the China Fundamental Research Report 2026, OECD, WIPO, Nature, ASPI and other linked sources. Several official pages were read through secondary coverage, so check key numbers against the original before citing. Spending comparisons depend on whether purchasing power or exchange rates are used.