EPFL AI Center

What EPFL AI Center does

EPFL AI Center is an academic research hub within the École Polytechnique Fédérale de Lausanne (EPFL), a Swiss federal institute of technology headquartered in Lausanne, Switzerland. Founded circa 2018 and co-directed by Prof. Marcel Salathé and Prof. Pascal Frossard, the center coordinates over 100 professors and 400+ researchers across EPFL's five schools and two colleges, organized around strategic initiatives spanning AI for Health, AI for Science, AI for Robotics, AI for Education, AI for Sustainability, AI Fundamentals, and AI for Society. Its operating model rests on three pillars — Research, Education, and Innovation — combining frontier academic work with structured industry and societal engagement.

The center produces open-source AI research assets with measurable technical performance: MeditronFO (the first fully open medical LLM framework, with 46,000+ clinical guidelines and a 6.6 percentage-point improvement on medical exams), MiCRo (a brain-inspired Mixture of Cognitive Reasoners architecture presented at ICLR), Saturn and TANGO (open-source molecular design frameworks published in Nature Machine Intelligence and Nature Computational Science), Apertus (a fully open multilingual LLM co-developed with ETH Zurich), and an adaptive Deep Brain Stimulation system for Parkinson's disease (Nature Medicine). Operational programs include the Industry Partnership Program (Swisscom, Audemars Piguet), Industry Fellowship Program, Startup Launchpad AI Track (powered by UBS, supporting spinoffs like WasteFlow), the Geneva AI Initiative for humanitarian organizations, the Citizens' Assembly on AI, and the Applied Machine Learning Days (AMLD) conference.

Revenue is generated through a hybrid public-grant and industry-sponsorship model. Major funding sources include a 20 million CHF contribution from the ETH Domain for the Swiss National AI Institute (SNAI) created in 2024 with ETH Zurich, over 10 million GPU hours from the Swiss AI Initiative seeded in December 2023, the Medtronic Chair in Neuromodulation (CHF 4 million over 8 years), Innosuisse BRIDGE and SNSF grants, ERC Advanced Grants (9 awarded), and program fees from industry partners and startup track sponsors. The center serves enterprise partners (Medtronic, Swisscom, Audemars Piguet, UBS), academic collaborators across 10+ Swiss institutions with 800+ researchers, Geneva-based NGOs and international organizations, EPFL-affiliated AI startups, and the broader public through open-source releases and educational programs such as the AI Product Management Program (AIPM).

Name

EPFL AI Center

Legal name

EPFL AI Center (École Polytechnique Fédérale de Lausanne)

Website

https://ai.epfl.ch

Company type

Private

Founded year

2018

Operating status

Operating

Headcount range

1–10 employees

Short description

EPFL AI Center is an academic research hub within EPFL that coordinates over 100 professors and 400+ researchers across health, science, robotics, sustainability, and education, producing open-source AI models, industry partnership programs, startup support, and humanitarian AI capacity building.

Ownership category

akta.pro rank

Product category

AI Research and Education

NAICS

Colleges, Universities, and Professional Schools (61131), Educational Support Services (611710), Computer Training (611420)

SIC

Services-Commercial Physical & Biological Research (8731), Services-Prepackaged Software (7372)

akta.pro primary industry

End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA)

akta.pro secondary industries

AI Observability, Monitoring & Evaluation Platforms (Drift, Quality, Safety) (HDAEANAF), AI Governance, Risk & Compliance (GRC) Platforms (HDAAAMAA), AI & Machine Learning (Foundations & Applied) (EDAMACAG), AI & Emerging Technology Literacy (EDAAAIAJ)

Keywords

• AI research
• Machine learning
• Foundation models
• Academic AI
• Open source AI

Headquarters

HQ city

Lausanne

HQ country

Switzerland

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Services

Cost components

Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations

Revenue model

1. Research Grants and Funding: EPFL AI Center operates primarily through institutional funding, Swiss government grants (ERC Advanced Grants, Innosuisse BRIDGE grants), and Swiss National Science Foundation support for AI research initiatives
2. Industry Partnership Fees: Companies sponsor fellowship programs and partner with EPFL labs to access AI expertise, talent, and research collaboration, providing funding for specific research projects
3. Swiss AI Initiative Computing Resources: Initial investment of 10,000,000+ GPU hours from Swiss AI Initiative plus 20 million CHF from ETH Domain for large-scale AI infrastructure and research
4. Healthcare AI Research Funding: Clinical trials funding for MeditronFO medical LLM evaluation across multiple sites from Switzerland to Tanzania, including MED.USE project for emergency department optimization

Go-to-market motion5 records

Distribution channels7 records

Marketing channels8 records

Core offering

EPFL AI Center is an academic research institution that conducts foundational and applied AI research across health, science, robotics, sustainability, and education, supported by 100+ professors and 400+ researchers. It delivers this research through open-source AI frameworks and models (e.g., MeditronFO, MiCRo, Saturn, TANGO), educational programs (AIPM), industry partnership and fellowship programs, startup acceleration, and conferences such as AMLD. Its activities are funded primarily through research grants, government funding (Swiss National AI Institute), and industry-sponsored collaborations.

Product overview

EPFL AI Center is an academic research institution (not a product company) that serves as the hub for large-scale AI research at EPFL, Switzerland. The center offers a portfolio of research initiatives, educational programs, and innovation support rather than a unified product. Key offerings include: MeditronFO (fully open medical LLM framework) and MiCRo (brain-inspired LLM architecture) for AI research; Saturn and TANGO (open-source molecular design platforms); vision prosthetics AI models; Inside AI podcast for public engagement; AIPM (AI Product Management educational program); Startup Launchpad AI Track for AI ventures; Industry Partnership and Fellowship Programs for collaboration; Geneva AI Initiative for humanitarian organizations; Citizens' Assembly on AI for governance; and AMLD as a major European AI conference. The center coordinates over 100 professors and 400+ researchers across strategic initiatives including AI for Health, AI for Science, AI for Robotics, AI for Education, AI for Sustainability, AI Fundamentals, and AI for Society.

Differentiator

Problem solved

Functional benefit

Brands

• Inside AI Podcast: A podcast hosted by Prof. Marcel Salathé, co-Director of the EPFL AI Center, featuring thought-provoking conversations with brilliant minds from EPFL and the global AI community about technology transforming our world.
• Citizen's Assembly on AI
• Applied Machine Learning Days (AMLD)

Products and services

• Industry Partnership Program Structured collaboration program that enables companies to access EPFL's AI expertise, talent pipeline, labs, and infrastructure to solve real-world AI challenges through research or applied projects; partners include Swisscom and Audemars Piguet.
• Industry Fellowship Program Sponsorship program enabling companies to fund PhD and postdoctoral researchers conducting interdisciplinary AI research in partnership with EPFL laboratories, providing early visibility into emerging AI breakthroughs.
• AI Product Management Program (AIPM) Educational program offered as a graduate course (CS-500), a 3-day sprint, and a tailored corporate program that trains product leaders to turn AI/ML theory into real-world product impact.
• Startup Launchpad AI Track Acceleration track powered by UBS that supports EPFL researchers and recent alumni in launching AI startups with grants (up to CHF 10,000 Ignition Grants), co-working space, mentorship, and ecosystem access.
• Geneva AI Initiative Capacity-building program co-designed with the Swiss Data Science Center that helps Geneva-based NGOs and International Organizations integrate responsible AI into their missions via training (DASHI course), project scoping, and project funding.
• Applied Machine Learning Days (AMLD) One of Europe's largest AI conferences, organized by EPFL AI Center, providing a forum for researchers, professionals, and industry to exchange AI ideas and insights.
• MeditronFO World's first fully open framework for building medical large language models, including pipelines, training procedures, and evaluation methods, designed for transparent medical AI development and clinical deployment.
• MiCRo (Mixture of Cognitive Reasoners) Large Language Model architecturally divided into four specialized modules (language, logic, social reasoning, world knowledge) modelled on human brain regions, providing more controllable and interpretable AI reasoning; presented at ICLR.
• Saturn Molecular Design Framework Open-source generative framework for molecular design based on Mamba architecture using reinforcement learning with Augmented Memory for sample-efficient drug discovery.
• TANGO (Tanimoto Group Overlap) Open-source reward function for generative molecular design that guides AI toward synthesizable molecules using specific building blocks during synthesis route optimization.
• Object-Level Vision Prosthetics AI Models AI models using topographic neural networks to predict optimal brain stimulation patterns for evoking object-level images (faces, houses) in visual prosthetics; validated in live trials on sighted monkeys and presented at ICLR 2026.
• Adaptive Deep Brain Stimulation (DBS) System AI-assisted, real-time deep brain stimulation therapy for walking impairments in Parkinson's disease that uses neural decoders to detect locomotor states and adjust stimulation within seconds; developed with Medtronic and published in Nature Medicine.
• WasteFlow AI Waste Sorting System AI-powered copilot developed by EPFL startup WasteFlow that identifies and measures 66 waste categories from sensor and camera data with 98% accuracy for some plastics, helping sorting facilities optimize throughput and recycling yields.
• Active Visual Sensing VR Platform Custom-built freely moving virtual reality arena using DeepLabCut-Live marker-less tracking to study infotaxis behavior in mice; released fully open source for neuroscience research.

Quantifiable outcome

• 30% proposal selection rate for ERC Advanced Grants (9 of 30 proposals)
• +5 more outcomes

Named customers9 records

Segments9 records

Ideal customer profiles5 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability24 records

Feature9 records

Partnerships

14 partnerships are on record, tiered core and minor.

• MedtroniccoreStrategic or Co-development Partner · 1 January 2026Medtronic collaboration enabled development of adaptive deep brain stimulation technology by providing access to and refinement of neuromodulation technology targeting gait problems in Parkinson's disease. The partnership established the Medtronic Chair in Neuromodulation at EPFL with CHF 4 million funding over 8 years.
• SwisscomcoreStrategic or Co-development PartnerSwisscom partners with EPFL AI Center through the Industry Partnership Program, gaining access to world-class AI expertise, talent, and collaborative research opportunities in AI innovation.
• Audemars PiguetcoreStrategic or Co-development PartnerAudemars Piguet participates in Swiss AI Initiative and EPFL AI Center strategic partnership ecosystem, collaborating on AI research and innovation projects.
• Swiss Data Science Center (SDSC)coreStrategic or Co-development PartnerSwiss Data Science Center co-designs the Geneva AI Initiative with EPFL AI Center, providing AI expertise and resources to guide humanitarian organizations and NGOs in integrating AI into their missions.
• University of Geneva (Research Hub on Democratic Innovations)coreStrategic or Co-development PartnerCollaboration with University of Geneva's Research Hub on Democratic Innovations for the Citizen's Assembly on AI, organizing deliberative processes and producing governance recommendations for responsible AI deployment.
• DemoscancoreStrategic or Co-development PartnerDemoscan association designed and conducted the Citizen's Assembly methodology and facilitation, overseeing the deliberative process involving 40 randomly selected citizens of Romandie.
• ETH ZurichcoreStrategic or Co-development PartnerETH Zurich and EPFL co-created the Swiss National AI Institute with combined investment of 20m ETH Zurich plus 20 million CHF from ETH Domain, providing national perspective on AI research, education, and innovation across 800+ researchers.
• CSCS (Swiss National Supercomputing Centre)coreTechnology or IntegrationCSCS provides computing infrastructure for the Swiss AI Initiative enabling large-scale model training, evaluation, and open release of foundation models through the Alps supercomputer.
• MOOVE (Massive Open Online Validation and Evaluations)coreStrategic or Co-development PartnerMOOVE platform enables clinician participation in ongoing evaluation and improvement of medical AI models, ensuring development remains grounded in real-world clinical practice through direct auditing of training material and validated outputs.
• Blaze (EPFL Entrepreneurship Program)coreChannel Partner/ Reseller/ DistributorBlaze accelerator program combining up to six months of coaching with up to CHF 10,000 in funding for student entrepreneurs, supporting 70 projects since 2020 including WasteFlow startup.
• New York UniversityminorStrategic or Co-development PartnerAcademic collaboration on Saturn and TANGO molecular design frameworks published in Nature Machine Intelligence and Nature Computational Science.
• MIT (Research Collaboration)minorStrategic or Co-development PartnerEPFL research team revisited MIT's Platonic Representation Hypothesis, with constructive engagement from MIT authors before publishing findings challenging the convergence theory.
• Microsoft ResearchminorStrategic or Co-development PartnerEPFL Professor Bob West spent sabbatical at Microsoft Research, providing insights into AI agent capabilities and the evolving AI landscape during Inside AI podcast discussion.
• Baylor College of Medicine, Stanford University, University of Oregon, Carnegie Mellon University, Rice University, Columbia Engineering School, Tübingen UniversityminorStrategic or Co-development PartnerInternational research collaboration contributing to active sensing research in mice visual decision-making, published in Current Biology with multi-institutional authorship.

Scale indicators19 records

Recent moves7 records

Expansion highlights6 records

Direct peers

• Allen Institute for AI (AI2): AI2 is a non-profit research institute producing open-source foundation models (OLMo, etc.) and science AI. Directly comparable on open LLM research (the OLMo base model is the underlying model for MeditronFO), open science stack, and mission-driven research.
• Mila (Quebec AI Institute): Mila is a university-affiliated AI research consortium. Comparable mission as a multi-institutional AI hub with foundation models, science applications, and large industry partner roster.
• UC Berkeley BAIR: Berkeley AI Research is one of the largest US university AI labs covering foundations, robotics, vision, and NLP, with strong industry partnership and open research culture mirroring EPFL AI Center.
• ETH AI Center: The sister Swiss federal institute's AI center and co-founder of the Swiss National AI Institute (SNAI). Directly comparable scope, talent scale, and co-funding model; co-leads Swiss AI Initiative alongside EPFL AI Center.
• Vector Institute (Toronto): Toronto's Vector Institute is a publicly funded AI research hub with industry partnership and graduate programs. Similar structure to EPFL AI Center: government-backed AI institute with corporate sponsors and research-fellow models.
• MIT CSAIL: MIT's largest interdisciplinary AI research lab. Closely comparable to EPFL AI Center as a university-based AI hub spanning foundational ML, robotics, vision, and language, with similar public-private partnership models.
• Alan Turing Institute: The UK's national institute for data science and AI, university-based. Comparable as a publicly funded multi-university AI research hub with applied-data focus and government backing.
• Stanford HAI / Stanford AI Lab: Stanford's HAI and SAIL drive cross-faculty AI research and industry engagement. Highly comparable: multi-school university AI hub with industry partnership programs, foundation-model research, and policy institutes analogous to EPFL AI Center.
• Carnegie Mellon University AI initiatives: CMU houses the first US PhD program in AI and runs multi-school AI research. Comparable as a university-based AI hub with deep robotics, vision, and language-model research.

Emerging players

• MBZUAI (Mohamed bin Zayed University of AI): MBZUAI is a graduate-level AI university backed by sovereign funding. Comparable as an externally funded research-first AI institution investing in foundation models and talent pipeline, though geographically focused on the Middle East.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/EPFL_AI_Center
• YouTubeyoutube.com/@EPFLAICenter
• LinkedInlinkedin.com/company/epfl-ai-center

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about EPFL AI Center