Swiss AI Initiative - Call for Small Grants

Swiss AI Initiative - Call for Small Grants is sponsored by Swiss AI Initiative (supported by ETH Zurich and EPFL). This opportunity supports mission-aligned projects and measurable outcomes.

Swiss AI Initiative – Large Grants ‒ FUNDING ‐ EPFL Swiss AI Initiative – Large Grants The call is currently closed. The Swiss AI Initiative focuses on developing foundation models and advancing responsible, efficient, and socially aligned AI technologies with significant impact for the Swiss and global community.

Through targeted funding, the Swiss AI Initiative enables cutting-edge research and the development of open science outputs in key areas of foundation model development, as well as high-impact application domains addressing critical societal challenges. Thematic areas and topic domains The call is open for all proposals that align with the goals of the Swiss AI initiative and require considerable computational resources.

Proposals should outline a research agenda that matches several of the following criteria : advance core AI capabilities (scalability, efficiency, alignment, safety, multimodal integration or evaluation) advance research in AI fundamentals or impact applications of AI benefit the Swiss or European or global ecosystem and societal context foster interdisciplinary collaboration, and/or collaboration across more than one group ensure ethical AI use and regulatory compliance promote principles of open science (including open-source, open weights, open data) foster participation of various stakeholders, such as public administration, SMEs, Startups, NGOs.

For the 3rd call, proposals in the following topic domains are welcome: AI for Science (including biology, chemistry, physics, astronomy and others) AI for Climate & Weather modeling Strengthening the interface of AI and democratic processes, including for example media, law and justice, or the public administration However, proposals in other areas will also be considered.

Eligible applicants must have an affiliation with an ETH Domain institution (ETH Zurich, EPFL, PSI, WSL, Empa or Eawag) and be employed in a permanent or Professor tenure-track capacity for the whole duration of the project.

Co-applicants requesting financial resources must have an affiliation with an ETH Domain institution , be in a position to receive funding and to carry out independent research in a permanent or Professor tenure-track capacity for the whole duration of the project.

Cross-institutional collaborations are strongly encouraged : computational resources can be accessed by co-applicants affiliated to academic institutions outside of the ETH Domain, such as cantonal universities, universities of applied sciences, and foreign public academic institutions.

Proposals may include additional partners ( not requesting nor accessing any resources ), such as administrations, NGOs, SMEs, or Startups Access to the CSCS Alps supercomputing infrastructure. (The compute hours amount requested can exceed 500k computer hours , though smaller amounts are also possible.) The core purpose of this initiative is the allocation of GPUh.

In addition to compute hours, applicants can exceptionally request funding for personnel and research costs , up to a maximum amount of 200k CHF . Engineering support from the Swiss AI Initiative team will be available for the funded projects (48 FTE-months total across all projects). Large project grants are allocated for a period of one year.

Please explore the Application toolkit and read carefully the application guidelines. All required documents must be submitted on the online submission platform . All documents must be submitted in English.

A short Declaration of Intent must be submitted by 16 March 2026 (17:00 CET). The submission deadline for the full proposal is on 31 March 2025 (17:00 CEST) .

The selection will be based on the following criteria, assessed by external reviewers: strength and complementarity of the team (as evidenced by prior collaboration, expertise diversity, and degree of inter-institutional or interdisciplinary collaboration) expected impact (as evidenced by expected value of open science artifacts, such as models, data, evaluations, platforms, etc., as well as planned technology transfer and outreach activities) feasibility of the project in a one-year time frame (as evidenced by initial preliminary results, experience managing large-scale AI projects, state of data collection, and plans for legal and regulatory compliance) Follow us in social networks Contact: [email protected]