Advancing Switzerland's leadership in responsible AI research

ETH Zurich and EPFL are releasing a large language model (LLM) developed on public infrastructure. Trained on the “Alps” supercomputer at the Swiss National Supercomputing Centre (CSCS), the new LLM marks a milestone in open-source AI and multilingual excellence.

Swiss AI Initiative

The Swiss AI Initiative is run by the Swiss National AI Institute. It was started in December 2023 and seeded with an initial investment of over 10m GPU hours on Alps (by CSCS) and a grant of 20m CHF by ETH Domain. The initiative is sustained with third party funding (industry partnerships, donations, etc) and grants.

It is the largest open science/open source effort for AI foundation models worldwide. By leveraging its integration with the ETH AI Center and the EPFL AI Center, the initiative benefits from the critical mass of both institutions as well as the expertise of a total of over 800 researchers (incl. 70 AI-focused professors) from over 10 academic institutions across Switzerland.

The frontier AI research is enabled by one of the world’s leading AI supercomputers (with over 10’000 GH200 GPUs) managed by the Swiss National Supercomputing Centre (CSCS). Regular compute calls are available to bring together researchers from different organizations.The initiative provides artifacts such as transparent and open software, models, and data releases, enabling their trustworthy use by various Swiss stakeholders, including SMEs and start-ups.

Compute Grants & Calls for Collaboration

The objective is to pool resources across leading academic institutions. To facilitate research on frontier AI foundation models, we are holding regular calls that are also open to researchers in Europe and beyond. For 2025, the 10-20 million GPU hours should be distributed. The large compute call – targeted at projects requesting more than 500k GPU hours – is held twice a year. Applications for small compute grants are continuously accepted.

Proposal deadlines & more information

Open call for small projects (~50k GPU hours), rolling reviews. Instructions for small project submission here.
Open call for large projects (>500k GPU hours), twice per year. Instructions for large project submission here.
- Call open for submission: August 8, 2025
- Deadline for mandatory Declaration of Intent: August 25, 2025
- Deadline for full Proposal: September 8, 2025

For any questions or clarifications, please do not hesitate to contact grants@swiss-ai.org.

Projects

The projects below are a highlighted selection of currently running projects funded through the Swiss AI Initiative.

SwissLLM: The Open LLM from the Alps

Open-source, multi-lingual 70B LLM model trained from scratch on 15 trillion tokens.

ETH Zurich: I. Schlag
EPFL: A. Bosselut, M. Jaggi

Academic Leads

Integrating Pedagogy into LLMs

Reliable pedagogically sound LLM-based tutors for university-level engineering education in Switzerland.

ETH Zurich: M. Sachan, G. Kortemeyer
EPFL: T. Käser, P. Jermann

Academic Leads

Foundations of Multimodal Learning

The goal of this project is to investigate the foundations of multimodal learning to tackle complex real-world problems.

EPFL: A. Zamir, L. Zdeborova
ETH Zurich: S. Tang, A. Krause
University of Geneva: F. Fleuret
NYU: S. Xie
MIT: P. Isola

Academic Leads

Democratizing LLMs for Global Languages

LLMs for underserved language communities via ethical data collection, training multilingual models, and designing benchmarks that reliably evaluate regional and cultural knowledge comprehension in multiple languages.

EPFL: A. Bosselut, M. Jaggi
ETH Zurich: I. Schlag
UZH: R. Sennrich
NUS: B. Low
U of Toronto / Vector Institute: C. Raffel

Academic Leads

Clintextualization

Modular, multimodal, multilingual ecosystem of medical LLMs that can be adapted across diverse clinical environments through lightweight contextualization from real world.

EPFL: M-A. Hartley, K. Keitel
ETH Zurich: M. El-Assady

Academic Leads

To view the full list of all current and previously funded projects, go to the All Projects page here. Further projects will be added over time after their successful application for large and small compute grants.

Industry Partners

Our goal is that our research translates into benefits for society & economy. To democratize access to AI especially for SMEs and start-up, we put an emphasis on freely available resources. For advanced partners, we offer an Industry Partnership Program (IPP) with a single point-of-contact for ETH Zurich & EPFL. Strategic partners are also doubling down in multi-year collaborations with research projects, donations to fellowship programs, and jointly coordinated research wallets.

SME Circle

We aim to make as many resources available in the open as possible to make the latest AI technology more easily accessible for SMEs. Interested SMEs can join our growing community in the Swiss SME Circle:

Learn how to leverage the freely usable frontier models such as the Swiss LLM (Apache 2.0)
Access to educational resources and AI research updates
SME Circle Meet-Ups at AI+X Summit and AMLD
Online Meet-ups of SME Circle

SNAI Membership

For companies who regard AI as a key priority for shaping their future business, we offer a dedicated partnership program:

Single, dedicated Point-of-Contact and personalized support for both ETHZ & EPFL
Community activation & engagement (incl. presence at AI+X summit & AMLD)
AI Talent & Innovation opportunities through challenges and project offerings
Visibility & Communication as a SNAI Member

Strategic Partnerships

For companies that look for a long-term, multi-year collaboration with significant transformative impact to their business and ecosystem, we set up customized strategic partnerships. Including:

SNAI Membership
Custom research wallet to boost efforts such as the Swiss AI Initiative
Custom plan for donations to fellowship programs for ground-breaking, exploratory, and interdisciplinary research
Custom educational program
Industry Advisory Council (voting seat)
Visibility & Communication as a SNAI Strategic Partner

Organization

Steering Committee

The Steering Committee is driving the strategic direction, shaping the activity areas, and deciding on resource allocation. ETH Zurich and EPFL have both equal voting power in this committee.