Six AI startup projects selected for new Launchpad funding
© 2025 EPFL
EPFL and UBS are proud to announce the six inaugural grantees of the Startup Launchpad AI Track – powered by UBS, Switzerland’s first grant program dedicated to AI. Selected from over 50 proposals, these projects will receive funding and tailored support to accelerate their journey from prototype to market readiness.
To translate EPFL’s breakthrough research into real‑world impact, EPFL has partnered with UBS to identify and support promising AI startup ideas across the EPFL ecosystem. Launched in June 2025, the programme supports these bold ideas with up to CHF 50,000 in grant funding, alongside tailored mentoring, state-of-the-art infrastructure, and access to UBS experts and industry networks.
Each selected team will accelerate their transition from prototype to commercialisation, advancing EPFL’s vision of building a leading hub for transformative AI innovation.
Segmeridian
Radiologists spend hours manually outlining organs on MRI scans while imaging demand soars and specialist numbers remain stagnant. Segmeridian is an AI-powered 3D interface that replaces laborious slice-by-slice labelling with one-click organ segmentation, reducing work from hours to minutes while maintaining accuracy.
The software runs locally within hospital systems, improving only via anonymised model updates. Pilots are already underway with hospitals worldwide, beginning in women’s health, a field where annotated datasets are especially scarce.
Contact: Pablo Garcia-Amorena Garcia, Founder
Anyway
Many organisations want to adopt AI, but two major barriers often stand in the way: persistent concerns around data privacy and the high cost of specialized hardware. Anyway solves both. Rather than outsourcing sensitive data to external services, Anyway allows companies to deploy powerful AI models entirely within their existing infrastructure.
Its plug-and-play software securely links an organisation’s existing machines and combines them into a high-performance compute cluster, allowing powerful AI models to run without data ever leaving the organisation’s systems. No privacy risk. No costly hardware. No unnecessary complexity. Designed to be fault-tolerant and easy to maintain, the system is already being tested with partners in banking and public administration. By combining privacy, simplicity, and performance, Anyway enables organisations to unlock AI’s potential securely and efficiently.
Contact: Geovani Rizk, Co-founder
Baio
The early stages of drug discovery are slow, costly, and often based on trial and error. Baio has developed an AI platform that designs synthesizable, target-specific drug candidates in a fraction of the time.
With their Large Drug Discovery Model (LDDM), the team can generate new molecules that fit a given protein target while ensuring they can be manufactured and validated quickly and cost-effectively. Their approach has already shown promising results and is being explored in collaborations with major pharmaceutical companies.
Baio aims to compress years of lab work into weeks, speeding up the route to cures for some of the world’s most pressing diseases.
Contact: Dr Arne Schneuing
AquaShield
Water leaks are a costly and persistent problem. Too often, they’re hidden behind walls or buried underground, typically detected only after severe damage. AquaShield has created an AI-powered detection system to identify and locate leaks in real time. It uses non-invasive sensors to analyse water flow data, pinpointing anomalies before they cause any damage. The team has already tested the system in collaboration with MIT and will soon run a live pilot in Germany.
Designed to be affordable, scalable, and easy to install, AquaShield aims to make proactive water monitoring standard in every building.
Contact: Marguerite Benoist
OPTIION
While renewable energy is transforming the power grid, it requires intelligent optimisation to operate efficiently. OPTIION uses machine and reinforcement learning - systems that continuously adapt to data - to predict electricity prices, solar generation, and grid conditions. It then helps grid-scale battery systems decide when to charge, discharge, or hold power – optimising performance and profitability. The system is currently being piloted with a major energy provider in Switzerland.
OPTIION’s technology adapts to rapidly changing grid and market conditions, enabling smarter, longer-lasting energy storage that supports a stable and sustainable power system.
Contact: Pål Forr Austnes, Co-founder and CTO
Windlens
Wind energy is one of the leading renewable energy sources and central to the global energy transition. However, wind producers sell their output hours or days in advance – and any gap between forecasted and actual energy production reduces competitiveness and profitability.
Windlens has developed an AI-powered forecasting model that improves wind prediction accuracy by 15–45%. Their technology provides farm-specific forecasts thanks to localised models trained on each site’s unique layout and conditions. It allows for up-to-date forecasts every 15 minutes, enhancing grid stability and enabling wind producers to operate efficiently and reliably at scale.
Contact: Matteo Peduto
About the Startup Launchpad AI Track – Powered by UBS
UBS has a long history of supporting innovation, and EPFL is proud to be working with them on this new initiative to support early-stage AI ventures. This partnership enhances EPFL’s ability to support AI-focused startups at the earliest stages by introducing resources that bridge the gap between academic research and private investment. With aligned goals and complementary strengths, EPFL and UBS are helping shape a more dynamic, innovation-driven future for Switzerland.
Contact: Marius Conti, Program Leader