Sintra AdvisorySeven Chinese universities share a nickname: the Seven Sons of National Defence. They report to the Ministry of Industry and Information Technology, the same agency that oversees China’s defence industry through its subordinate body SASTIND. The Australian Strategic Policy Institute rates all seven as very high risk on its China Defence Universities Tracker. Four sit on the US Bureau of Industry and Security Entity List. More than half of their research budgets go to defence work, and all seven hold accreditation for classified weapons research.
Until December 2025, they were eligible for European Union research funding.
A 2023 investigation exposed the funding pipeline
The Horizon Europe programme, the EU’s €93.5 billion research and innovation framework, had been sending money to these institutions for years. A 2023 investigation by Science|Business identified five active Marie Skłodowska-Curie Actions involving four of the Seven Sons, worth a combined €4.4 million in EU contributions. The projects covered swarm robotics, high-performance motor drives for propulsion systems, vehicular communications security, and heat transfer technologies with published applications in fighter jet thermal management.
One project captures the contradiction more precisely than any policy paper could. Micro-FloTec, a €437,000 Horizon Europe grant involving Beihang University (CORDIS 101082394), started on 28 March 2023. Two days later, Commission President Ursula von der Leyen delivered her landmark speech on de-risking Europe’s relationship with China. The EU was funding a Seven Sons university at the same moment its leadership was publicly warning about strategic dependencies on Beijing. When Science|Business asked the Commission about these projects, the Commission declined to comment. Under the rules at the time, nothing was technically wrong.
The scope widened from five projects to €26 million and 20 new agreements
The five MSCA projects were the visible edge of a larger pattern. In March 2025, Swedish MEP Karin Karlsbro (Renew Europe) filed Parliamentary Question E-000926/2025, putting harder numbers on the problem. During the past decade, she wrote, the EU had funded at least 14 projects worth approximately €26 million involving Chinese universities that independent experts considered high-risk. And between December 2024 and January 2025, while the Commission was already drafting new restrictions, over 20 new agreements, partnerships, and joint research publications were announced between Seven Sons universities and European institutions.
The timing of those agreements deserves scrutiny. In December 2024, Beihang University President Wang Yunpeng spent a week touring France and Austria under the banner of the Beihang-Europe Cooperation Committee, a body the university had quietly founded the year prior. The delegation visited ENAC (France’s national civil aviation school, where Beihang runs a joint aviation institute in Hangzhou), IIASA in Austria, Dassault Systèmes, ESPCI Paris, and INSA Toulouse. The trip produced new agreements on AI, integrated circuits, metaverse technologies, chemistry, and a scholarship pipeline funnelling Beihang undergraduates into European institutions. Beihang’s School of Integrated Circuits separately visited the Catalonia Institute of Nanoscience in Barcelona to discuss two-dimensional materials research. All of this in a single month.
Meanwhile, Northwestern Polytechnical University, the only Chinese university with a dedicated defence drone laboratory according to ASPI, signed a new collaboration with the University of Salamanca on aeronautical engineering. Nanjing University of Aeronautics and Astronautics inked a joint laboratory agreement with Samara University in Russia. Beijing Institute of Technology deepened its partnership with Bauman Moscow State Technical University on underwater robotics and lunar rovers. These were not academic courtesies. They were institutional commitments, signed and announced, creating obligations that would outlast any policy shift in Brussels.
Karlsbro’s parliamentary question cited these Russia-facing partnerships directly, raising the concern that EU-funded research collaborations with the Seven Sons could be indirectly supporting Russian military technology development. The agreements documented by the Jamestown Foundation showed Beihang, NPU, NUAA, and BIT all strengthening ties with Russian defence-linked institutions in the same months that European restrictions were being drafted. European institutions were building new ties with the same universities that European policymakers were preparing to ban.
Institutions drew their own lines before Brussels acted
Some institutions did not wait for the Commission. On 20 January 2024, the Flemish government announced a complete ban on new collaborations with all Seven Sons, calling them deeply rooted in China’s defence industry and under the military control of the Chinese army. Flanders simultaneously stood up a knowledge security desk, modelled on the Dutch National Contact Point for Knowledge Security, and published new guidelines for its universities.
In Germany, FAU Erlangen-Nuremberg became the first university to suspend cooperation with scholars funded by the Chinese Scholarship Council, effective June 2023, citing risks of espionage and conflicts between CSC contracts and Germany’s Basic Law. In the Netherlands, TU Delft began refusing candidates linked to the Seven Sons outright, and Maastricht University withdrew three PhD positions connected to the group.
Denmark’s Aarhus University rejected 24 foreign researchers in the first seven months of 2025 on security grounds, one in every twelve applicants from China, Russia, and Iran. The university hired five specialists fluent in Russian, Chinese, and Persian specifically to evaluate incoming applications. Vice-Dean Brian Vinter was direct about the reasoning: the applicants were rejected based on the risk that they would gain access to information that could put them under pressure to share it with a foreign power. Aarhus had already stopped admitting PhD students funded by the Chinese Scholarship Council in March 2022, after concluding that CSC contracts imposed obligations incompatible with academic freedom. The University of Copenhagen announced its own screening procedures in August 2025, employing dedicated staff and an external consultancy to assess applicants in the natural and health sciences.
The Copenhagen model points to a broader pattern: institutions that moved early did not rely solely on internal capacity. Several turned to third-party expertise to conduct background screening, evaluate institutional affiliations, and assess the risk profiles of prospective partners and personnel. The logic is straightforward. Few research offices have in-house fluency in Chinese military-industrial structures, Russian defence procurement networks, or the overlapping entity lists maintained by multiple jurisdictions. Outsourcing that assessment to specialists who track these affiliations full-time produces faster, more defensible determinations than asking a department head to evaluate whether a visiting scholar’s home institution appears on an export control list they have never read.
The most carefully reasoned action came from the Université Libre de Bruxelles. On 5 May 2025, ULB’s Academic Council voted to suspend its institutional partnership with Beihang University and to rule out new agreements with any of the Seven Sons. The trigger was a review by ULB’s Committee on Compliance with International Law and Dual Use, which had been asked to evaluate both a proposed new agreement with NUAA and the renewal of an existing memorandum with Beihang. The Council’s reasoning went beyond the usual security language. It cited the total opacity with which the Seven Sons operate, the extent of their classified military research, and the risk of sensitive technologies being transferred to universities that work directly with the Chinese armed forces. The conclusion should give every European research administrator pause: risk management with these institutions is not difficult; it is impossible. No European university can reliably assess what its research contributions to these partners will ultimately be used for. Individual academic relationships with colleagues at Seven Sons universities were explicitly left unaffected. The restriction targets the institutional pipeline, not the personal one.
The December 2025 work programme reversed the default on Chinese participation
The October 2025 Flagship Conference on Research Security, held in Brussels with roughly 500 policymakers from 15 countries, set the final stage. Commissioner Ekaterina Zaharieva announced six new measures, including a European Centre of Expertise on Research Security and, for the first time, a commitment to embed research security into law through the forthcoming European Research Area Act.
Six weeks later, on 11 December 2025, the Commission adopted the Horizon Europe Work Programme 2026-2027 (Decision C(2025) 8493) with a budget of €14 billion. The General Annexes, applying Article 22(6) of the Horizon Europe regulation, reversed the default on Chinese participation entirely. Previous work programmes had listed specific calls where Chinese entities were restricted. The new programme states that unless otherwise specified, Chinese entities cannot participate in any Research and Innovation Action. The shift from opt-out to opt-in may sound procedural. In practice, it rewrites the terms of consortium-building across the continent.
Three clusters are now entirely closed to Chinese entities: Cluster 1 (Health), Cluster 3 (Civil Security for Society), and Cluster 4 (Digital, Industry and Space). That covers AI, semiconductors, quantum technologies, biotechnology, telecommunications, and health research. Chinese entities can still participate in parts of three remaining clusters covering culture, climate, and food systems, along with a handful of bilateral topics co-funded by China’s Ministry of Science and Technology. The Seven Sons are banned from all of it, even the clusters that remain open to other Chinese entities.
The programme also introduced a provision that will occupy grants officers for the foreseeable future: EU-based entities directly or indirectly controlled by China are excluded from Innovation Actions in AI, semiconductors, quantum, and biotechnology. The General Annexes never define what “controlled” means. No ownership threshold. No board composition test. No guidance on joint ventures or spin-offs with partial Chinese investment. Grants offices are expected to make these determinations on a call-by-call basis, using criteria borrowed from the 5G security provisions that amount to a list of factors without a decision rule.
The operational aftermath is where things get complicated
Every institution preparing a Horizon Europe application now needs to verify that no consortium member is established in China for restricted clusters, that no EU-based partner is controlled by a Chinese entity for critical technology Innovation Actions, and that no MIIT-affiliated university appears in any capacity anywhere in the application. For institutions with long-standing Chinese partnerships, cotutelle agreements, visiting scholar arrangements, and joint PhD supervision, the review reaches into institutional relationship infrastructure that many European universities spent decades building. Some of those relationships will survive scrutiny. Others will not. The problem is that most institutions have never conducted the kind of systematic assessment required to tell the difference.
The Commission’s work programme text justifies the restrictions by citing a persistent lack of progress in the EU-China Innovation Cooperation Dialogue and substantive concerns about undesired transfer of IP to China, naming both the Made in China 2025 strategy and China’s Military-Civil Fusion strategy. The restrictions do not apply retroactively to existing agreements, but the direction of travel is clear.
The proposed next Framework Programme, FP10, would make dual-use research eligible by default across a €175 billion budget, breaking with four decades of civilian exclusivity. The ERA Act, expected in the second half of 2026, would make research security legally binding for the first time. Today’s work programme restrictions will likely become tomorrow’s legal obligations.
University associations have pushed back carefully. The Guild questioned whether expanding restrictions to full research clusters was justified. EUA warned against risk aversion replacing risk management. The Coimbra Group cautioned against ignoring the cost of not collaborating. Joy Zhang, sociologist at the University of Kent, told Nature in February 2026 that IP theft is more common among commercial companies than cutting-edge researchers, arguing the restrictions may represent a right intention applied to the wrong target. Commission business analyst Beatrice Plazzotta countered at a January 2026 webinar that the current geopolitical and economic situation requires paying attention to what institutions do and who they collaborate with. These are fair arguments, and the debate will continue long after the current work programme expires. The compliance obligations will not wait for it to finish.
The trajectory points in one direction
ULB’s Academic Council arrived at the clearest framing of the underlying problem: the opacity of these institutions makes risk management impossible. That observation applies well beyond the Seven Sons.
Thirty months ago, the Commission was funding joint projects with military-linked Chinese universities and declining to comment on it. Today, those universities are excluded from the entire programme, and the definition of who else might be excluded is expanding faster than the guidance on how to determine it. The restrictions have moved from named universities to undefined “controlled” entities, from work programme conditions to binding legislation on the horizon, from opt-out exceptions to opt-in defaults. Each step has been wider than the last. Each has arrived faster than institutions expected.
The question for research offices is no longer whether these restrictions are justified. It is whether the institutional infrastructure exists to answer the screening question before a funder or auditor asks it. Institutions that have already built that capacity will absorb whatever comes next. Those that treat December 2025 as a one-time adjustment will find themselves rebuilding their compliance approach every time the rules tighten further.
Sintra Advisory works with research institutions on partnership vetting and research security compliance.
Sources
EU documents and proceedings
Horizon Europe Work Programme 2026-2027, Decision C(2025) 8493 (European Commission, December 2025)
New measures to strengthen research security (European Commission, October 2025)
Institutional decisions
Decision of the Academic Council on collaborations with the Seven Sons universities (ULB, May 2025)
Flanders bans collaborations with Chinese Seven Sons universities (Belga News Agency, January 2024)
Danish universities reject Chinese researchers over espionage concerns (Scandasia, July 2025)
Investigations and analysis
EU continues to fund research with Chinese military-linked universities (Science|Business, May 2023)
Why Europe barred China from flagship Horizon research programmes (Nature, February 2026)