NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC

China's US$14B State Venture Fund Opens, With 70% Bound for Seed Stage

China formally launched the National Venture Capital Guidance Fund on 26 December 2025 with RMB 100B (≈US$14.3B) in central money, a 20-year life and three regional funds for hard tech.

2026-11-17 · 1019 words · NeuroAI
China's US$14B State Venture Fund Opens, With 70% Bound for Seed Stage

A chip designer in Shenzhen has a prototype and no lead investor. A robotics founder in Suzhou needs five years before revenue arrives. On 26 December 2025, Beijing switched on a fund built for exactly that gap: twenty years of patience, with most of the money required to reach companies before they look fundable.

The National Venture Capital Guidance Fund (国家创业投资引导基金) was formally launched that day at a press conference held by the National Development and Reform Commission (NDRC). It is not an AI fund in the narrow sense. It is a state-backed machine for early-stage "hard technology" (硬科技) — and chips, AI and robotics are squarely inside its mandate.

Key takeaways

  • The fund was formally launched on 26 December 2025, jointly initiated by the NDRC and the Ministry of Finance (MOF), according to the Chinese government's official website.
  • Central government money totals RMB 100 billion (≈US$14.3B / HK$111B), sourced from ultra-long special treasury bonds, with a stated target of RMB 1 trillion (≈US$143B / HK$1.11T) in overall scale.
  • Three regional funds — Beijing-Tianjin-Hebei, the Yangtze River Delta, and the Guangdong-Hong Kong-Macao Greater Bay Area — were established and running at launch.
  • The structure has three layers: a fund company, regional funds and sub-funds. The top-level company has a 20-year lifespan, split into 10 years of investment and 10 of exit.
  • At least 70% of capital must reach seed-stage and early-stage enterprises; average sub-fund size is capped at about RMB 1 billion (≈US$143M / HK$1.11B).
  • Reuters reported that each regional fund is expected to exceed RMB 50 billion (≈US$7.1B / HK$55.6B) — a figure from reporting, not from the official announcement.

What was actually launched

The official account is precise. The fund uses proceeds from ultra-long special treasury bonds, with RMB 100 billion (≈US$14.3B / HK$111B) contributed at the national level, and encourages social capital to participate at the regional and sub-fund levels. officials described a leverage effect: central money is meant to pull in local governments, central state-owned enterprises, financial institutions and private investors until the whole system reaches roughly RMB 1 trillion (≈US$143B / HK$1.11T).

That is a target, not a deposited sum. The distinction matters, because China has a long history of guidance funds whose announced scale exceeded what they ultimately deployed.

At launch, the three regional funds signed preliminary investment intentions covering integrated circuits, quantum technology, biomedicine, brain-computer interfaces (脑机接口) and aerospace. The NDRC's English summary also lists artificial intelligence and future energy. This is a broad hard-tech mandate, not a single-sector bet.

The three-tier machine

The design is worth understanding because it determines where a startup actually encounters the money.

  • Fund company (引导基金公司). A corporate vehicle at the top, with a 20-year lifespan and the job of setting policy direction.
  • Regional funds (区域基金). Three of them, set up as limited partnerships. They invest through "sub-funds plus direct projects" (子基金+直投项目), with at least 80% of their capital going into sub-funds and the direct portion coordinated with national key projects.
  • Sub-funds (子基金). Market-selected managers run these. A regional fund will not be the largest investor or largest shareholder in a sub-fund — a deliberate choice to keep the state in a guiding role rather than a controlling one.

The rules then narrow the target further. Sub-funds must put at least 70% of their capital into seed and early-stage companies. Official guidance reported by the Financial News put the "invest small" (投小) test concretely: target companies valued below RMB 500 million (≈US$71.5M / HK$557M), with any single investment capped at RMB 50 million (≈US$7.15M / HK$55.6M).

What "patient" means in practice

A 20-year life — 10 years investing, 10 exiting — breaks the standard seven-to-ten-year venture clock. For a semiconductor or robotics company, that changes which milestones are fundable. A foundry process or a dexterous hand may need years of iteration before a customer will sign a volume order; a fund that must exit in year eight cannot comfortably underwrite that.

The government also says it will not set geographic reinvestment requirements and will not manage funds day to day. Managers are chosen through competitive selection. Those are meaningful design choices, because a "return to the local region" clause has historically distorted how Chinese guidance funds deploy capital.

According to Li An, chief scientist at BrainNet (脑机网), the practical value of very long-dated state capital is not the headline size but the freedom it gives founders to pursue hardware milestones that a conventional fund cycle would force them to abandon.

Not the same fund you have read about

This vehicle is distinct from the National AI Industry Investment Fund (国家人工智能产业投资基金), the separate MIIT-and-MOF vehicle registered in Shanghai in January 2025 and backed by the third phase of the integrated-circuit "Big Fund." The two are often conflated in English coverage. One is a broad, early-stage, multi-region guidance structure funded by special treasury bonds; the other is a single AI-themed fund with an equity base. They have different parents, different mandates and different clocks.

Honest limitations

This article relies on the official NDRC and Chinese government announcements of 26 December 2025, plus Reuters reporting on regional fund size. The RMB 1 trillion figure is an aspirational target, not committed or deployed capital; the RMB 100 billion is national fiscal contribution, not the whole pool. The "70%" and sub-fund size rules are stated policy requirements, not audited deployment figures. Reuters' "more than RMB 50 billion per regional fund" is a reported expectation rather than an official number in the launch materials. Early-stage funds historically under-deploy against announced targets, and no portfolio data is available here. Sector priorities can shift. This is not investment advice.

What readers can do now

  • If you are a hardware founder in China, check the sub-fund managers actually selected in your region — the rules matter less than who is writing the cheques.
  • For anyone tracking AI or robotics capital, distinguish this fund from the separate National AI Industry Investment Fund before citing either figure.
  • Watch the first disclosed sub-fund close: it will show whether the 70% seed-stage rule translates into real early cheques or is absorbed by larger deals.

Related coverage

Cite this article

NeuroAI (脑机网), "China's US$14B State Venture Fund Opens, With 70% Bound for Seed Stage," 17 November 2026, https://neuroai.site/a/na-capital-national-vc-guidance-fund

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Capital & Markets” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories