Europe–China Digital Commons Observatory 中文

Home › Research › How Europe and China fund open source — chapter with conclusions

How Europe and China fund open source — chapter with conclusions

Source file: docs/FUNDING-CHAPTER.md · updated 2026-11-08

Version 1.0, 3 November 2026. Closes research branches R12–R15 (docs/RESEARCH-REGISTER.md) for this round. Later rounds produce new versions of this chapter; changes are listed at the end.

Open-source research from public sources, not official statistics. Incomplete and correctable. Absence of data means we found nothing, not that nothing exists. Contribute: GitHub issues (label funding-data) or contact@eu-china-commons.eu.

Labels: [F] fact (sourced, evidence A/B) · [C] commitment (announced, not shown as spent) · [I] inference · [P] proposal.

1. What we measured

ComponentRecordsSource file
Public money, open-source-specific onlyEU 24, CN 21 (+ excluded generic schemes logged)data/funding/, docs/FUNDING-COMPARISON.md
Private corporate money via foundationsEU 23 (Eclipse, FSFE, KDE, TDF…), CN OpenAtom 2020–23same, §1a
Modalities (who decides, instruments, strings)qualitative fields on all recordsdocs/FUNDING-MODALITIES.md
Funding of individual projects34 projects (18 CN-origin, 16 EU)docs/PROJECT-FUNDING.md, data/project-funding.json

Amounts in EUR at ECB annual average rates. Depth reached: centres 2–3, peripheries 1–2 (docs/RESEARCH-DEPTH.md).

2. Findings

2.1 Amounts

  1. [F] Documented EU public open-source money that can be summed: €11.5M (2023), €36M (2024), €19M (2025) — Germany and France only. On top, multi-year EU envelopes [C]: DARE up to €120M, NGI Zero ≈€65.5M (2018–27), OpenEuroLLM €20.6M, Chips JU RISC-V €60M.
  2. [F] No Chinese public open-source amount could be summed: China publishes per-project caps and rates, not budgets or disbursements (largest caps: Wuhan East Lake ¥30M, Shanghai ¥20M, Shenzhen ¥10M). [F] No central-government open-source budget line was found.
  3. [F] OpenAtom, China's national open-source foundation, filed ¥270M of income 2020–23 (≈€9–10M/yr in donations), 100% domestic, Huawei 37% (2023), plus ¥20.7M of government subsidy in 2023. [I] The widely cited "¥600M+ since 2020" (Xinhua) is not supported by the filings.
  4. [F] European private corporate money is fragmented and small where it can be isolated (FSFE + KDE ≈€0.1–0.2M/yr); Eclipse Foundation (Belgium) has €9.4–15.4M revenue (2022–25) but does not separate member fees from public grants. [I] With Eclipse as an upper bound, European foundation money is of the same order as OpenAtom.

2.2 Modalities

  1. [F]/[I] China pays for adoption; Europe pays for creation and maintenance. Chinese instruments are local adoption subsidies (Shenzhen 15% of investment + 5% for OpenHarmony-adapted RISC-V), prize-style grants paid after deployment, per-device rates. European instruments are competitive grants for new strategic tech (RISC-V, LLMs), small grants via intermediaries (NLnet), and — uniquely in Germany — a standing maintenance fund (Sovereign Tech Fund: curl €195k, OpenSSL €406k, KDE €1.29M).
  2. [F] Chinese money is tied to named national stacks (OpenHarmony, openEuler, RISC-V, OpenAtom-hosted projects) and local registration; European money is stack-neutral and conditioned on open licences and co-funding. [I] In practice Chinese instruments exclude foreign firms; European ones do not exclude anyone by nationality.
  3. [F] Who decides: in China local industry bureaus and development zones (the centre acts through OpenAtom, under MIIT supervision); in Europe the Commission's programmes and a few national agencies. The EU periphery (member states) is mostly absent: 23 of 27 show no dedicated money in our search [I: light search, not conclusive].

2.3 Projects

  1. [F]/[I] Chinese flagships depend on one company each (Huawei for OpenHarmony/openEuler; Alibaba, Baidu, Tencent, Ant, PingCAP, High-Flyer…), with the top funder close to 100%; Chinese-origin projects at Apache keep the founding company as payer. European projects have more diverse, multinational funders and are paid mostly by subscriptions or dual licensing (Nextcloud, Qt, SUSE, Element), with grants and donations for maintenance (curl, GnuPG, KDE, OpenSSL).
  2. [F] Individual donations are almost absent in China (Ant Design ≈US$18k all-time on Open Collective; KubeSphere 0) versus e.g. curl ≈US$91k/yr.
  3. [F] Governments as paying customers of open-source vendors are a European channel (BSI VS-Desktop for GnuPG, BwMessenger/Tchap for Element, openDesk); the closest Chinese equivalent is 信创 procurement of commercial openEuler distributions, which buys the proprietary product layer, not the commons. [F] Part of OpenAtom's project money flows back to the main donor (¥6.49M of ¥25.8M OpenHarmony spending to Huawei Cloud, 2023).

2.4 Transparency

  1. [F] The transparency gap is the largest asymmetry found: European budgets and grants are published per project; Chinese support is visible as rules and ceilings, and award notices are scattered across local portals. [I] Any headline "China spends X vs Europe Y" is today unfounded in both directions.

3. Limits of this round

4. Implications for EU–China cooperation [P]

  1. Complementarity is real and measurable: China funds adoption at scale, Europe funds maintenance and creation. A shared agenda on maintenance of common upstreams (projects both sides depend on: Linux, OpenSSL, curl, RISC-V toolchains, Apache projects of Chinese origin) is non-competitive and verifiable.
  2. Transparency as the first cooperation deliverable: a common, open template for reporting public open-source funding (amount, instrument, project, disbursement) — the schema used here could be the starting point.
  3. Diversify funding of Chinese-origin global projects: single-company dependence is a resilience risk for everyone using them; European maintenance funds and foundations could co-fund under open governance.
  4. Screen, don't exclude: OpenAtom is a state-supervised, single-country-funded vehicle; engagement should go through neutral foundations (Eclipse, Apache, OpenChain) with economic-security screening, as already practised in Oniro and Catena-X.

5. Next rounds (re-run recipes in RESEARCH-REGISTER R12–R15)

Chinese award notices (Shenzhen, Shanghai, Wuhan, Hangzhou, Guangdong); national budgets of the 10 largest EU economies; Eclipse Belgian statutory accounts; OpenDigger affiliation for openEuler, OpenHarmony, MindSpore, PaddlePaddle; OpenAtom 2024–25 filings when published.

Changelog