CAST's Bold Move: Suspending NeurIPS 2026 Funding in Response to US Sanctions Policy
The China Association for Science and Technology (CAST), a pivotal organization supporting scientific endeavors across Chinese universities and research institutions, announced on March 27, 2026, that it would immediately halt funding applications for scholars attending the 2026 Conference on Neural Information Processing Systems (NeurIPS). This decision came swiftly after NeurIPS organizers updated their submission guidelines to enforce stricter compliance with US sanctions, effectively barring participation from numerous Chinese entities. The move underscores escalating tensions in international academic collaboration, particularly in artificial intelligence (AI), where Chinese universities have emerged as global leaders.
NeurIPS, one of the premier AI and machine learning conferences, draws thousands of researchers annually. For Chinese academics, it has been a critical platform for showcasing breakthroughs and networking. However, the new policy linked to the US State Department's Consolidated Screening List raised alarms, as it encompassed over 800 Chinese institutions and companies, including giants like Huawei and SenseTime. CAST's suspension means no financial support for travel, registration, or related costs, redirecting resources to domestic conferences or international venues that uphold principles of openness and equality.
Understanding NeurIPS: A Cornerstone for Global AI Research with Heavy Chinese Participation
NeurIPS, formally the Conference on Neural Information Processing Systems, originated in 1987 as a small workshop and has grown into the world's largest AI gathering, typically accepting around 2,500-5,500 papers from tens of thousands of submissions. In recent years, Chinese institutions have dominated: at NeurIPS 2025, Tsinghua University led with the most accepted papers, followed by Peking University and others, with China surpassing the US as the top contributor country for the first time. Universities like Shanghai Jiao Tong, Zhejiang University, and the Chinese Academy of Sciences routinely feature prominently, reflecting China's massive investment in AI talent—producing over 50% of the world's AI researchers.
This dominance stems from national strategies like 'Made in China 2025' and the New Generation AI Development Plan, funneling billions into university AI labs. NeurIPS papers often count toward promotions, grants, and international rankings for faculty at top Chinese universities such as Tsinghua and Fudan. A ban or funding halt disrupts this pipeline, affecting PhD students, postdocs, and professors who rely on such visibility for career advancement.
NeurIPS' Sanctions Policy: From Announcement to Rapid Reversal
On March 25, 2026, NeurIPS released its Main Track Handbook for 2026, mandating compliance with US sanctions. The policy prohibited 'services'—including peer review, editing, and publishing—to Specially Designated Nationals (SDNs) and considered submissions from 'Non-SDN' Chinese entities on the broader US list. This was interpreted as barring 873 Chinese institutions, sparking immediate outrage.
By March 27, NeurIPS issued an apology on its website and X (formerly Twitter): 'In preparing the NeurIPS 2026 handbook, we included a link to a US government sanctions tool that covers a significantly broader set of restrictions than those NeurIPS is actually required to follow. This error was due to miscommunication... We deeply regret the confusion and concern.' They clarified only strict SDN/EAR compliance applies and updated the handbook. Despite the reversal, damage was done—Chinese bodies had already mobilized.
CAST's Precise Countermeasures and Broader Chinese Academic Backlash
CAST, representing over 8 million scientists and engineers affiliated with universities nationwide, stated: 'Effective March 27, 2026, CAST stops accepting funding applications for the 2026 NeurIPS conference... NeurIPS 2026 papers will not be recognized as representative achievements for CAST projects.' Funding is redirected to supportive venues.
The China Computer Federation (CCF), influencing university curricula and recommendations, urged a boycott of reviewing and submissions, removing NeurIPS from its top conference list. Other societies followed, with professors like Tsinghua's Huazhu Fu withdrawing as area chairs. This collective stance amplifies impact on Chinese higher education, where conference metrics heavily influence hiring and funding.
Direct Impacts on Chinese Universities and Researchers
Top AI powerhouses like Tsinghua (home to over 100 NeurIPS papers in 2025) and Peking University face hurdles. Faculty promotions in China often require international conference papers; NeurIPS devaluation hurts. PhD students lose funding for international exposure, crucial for global careers.
Collaborations with US firms/labs may suffer, as dual affiliations trigger scrutiny. Universities with Huawei partnerships, common in Shenzhen and Beijing campuses, are hit hardest. A survey by CCF hinted 30% of young researchers planned NeurIPS submissions—now pivoting domestically.
- Loss of CAST grants: ~¥500,000-1M RMB per attendee historically.
- Paper recognition void: Affects NSFC grants, 'Double First-Class' university status.
- Career ripple: Reduced CV weight for jobs at Baidu, Alibaba, or abroad.
Historical Context: US-China 'Decoupling' in AI Academia
This isn't isolated. US Entity List since 2019 restricted Huawei access to chips/tech, spilling into academia via export controls. Prior conferences like ICML faced similar scrutiny. Chinese universities responded by boosting domestic events like CCF's CCMLC, now rivaling NeurIPS in scale.
China's AI output surged 20% yearly, with 70% of global patents. Yet, US policies aim to curb military applications, citing national security. Balanced views: US experts worry tech transfer; Chinese see politicization harming science.
South China Morning Post analysis highlights fears of 'academic Iron Curtain'.Statistics Highlighting China's Stake in NeurIPS
China authored ~25-30% of NeurIPS 2025 papers (up from 10% in 2015), with Tsinghua alone at 4.73%. In 2024, 41 from Riken AIP, but Chinese unis topped. Loss could shift ~1,000 papers, denting global AI progress.
| Year | China % Papers | Top Chinese Uni |
|---|---|---|
| 2025 | ~30% | Tsinghua |
| 2024 | 28% | PKU |
| 2023 | 25% | Zhejiang U |
Stakeholder Perspectives: Voices from Chinese Academia
Tsinghua Prof. Huazhu Fu: 'Cannot endorse discriminatory policy.' CCF President: 'Violates academic freedom.' University admins worry about rankings—QS/THE weigh conferences heavily. Students fear limited exposure amid 'publish or perish'.
Positive spin: Accelerates self-reliance, boosting CCMLC (10k+ attendees 2025).
Future Outlook: Domestic Rise and Global Fragmentation?
Short-term: Fewer Chinese at NeurIPS 2026 (Sydney?), but policy reversal may soften. Long-term: China invests ¥100B+ in AI by 2030, prioritizing homegrown forums. Universities like BAAI host rivals. Risk: Slower innovation from silos, but gains in sovereignty.
Solutions: Bilateral agreements, neutral venues. For Chinese unis: Leverage ICLR, ICML less affected; focus EU/Japan collabs.
NeurIPS 2026 Handbook (updated) | CAST Statement (Chinese)
Actionable Insights for Chinese AI Researchers and Universities
- Alternatives: Submit to CCF-CCMLC, AAAI (US but less strict).
- Funding: Apply NSFC early; CAST prioritizes domestic.
- Career: Diversify to ArXiv preprints, GitHub for visibility.
- Compliance: Check affiliations; use personal emails if non-sanctioned.
This saga highlights academia's vulnerability to geopolitics, urging diversified strategies for China's thriving AI ecosystem.
.png&w=128&q=75)
