3 new startup projects secure grants from the Startup Launchpad

Quanexum, Scholé AI and Lemna Bio secure grants form the Startup Launchpad© 2025 EPFL
Three new startup projects from EPFL have received support from the Startup Launchpad. Quanexum develops automated measurement systems and transistor models for cryogenic electronics, making low-temperature circuit design faster, more precise, and more accessible. Scholé AI leverages artificial intelligence to create personalized training programs that help professionals quickly acquire AI and data skills through interactive and scalable content. Both startup projects have been awarded an Ignition Grant of CHF 30,000. Meanwhile, Lemna Bio, which harnesses AI-powered geometric transformers to accelerate and enhance biomolecule design for drug development, has received an Innogrant of CHF 100,000.
Quanexum
Quanexum – an EPFL startup based at Professor Charbon’s Advanced Quantum Architecture lab(AQUA) – has been awarded a CHF 30,000 Ignition grant by the jury at the EPFL Startup Launchpad.
Cryogenic electronics are poised to transform industry by allowing electronic systems to operate at extremely low temperatures (down to -270°C). This deep cooling enhances performance across applications such as high-performance computing, quantum computing, space exploration, and ultra-sensitive scientific measurements. However, developing these cryogenic electronics comes with significant challenges due to the lack of cryogenic transistor models and high-throughput characterization tools. Inaccurate transistor models complicate the low-temperature circuit design, which cannot be simulated and optimized prior to fabrication, while inefficient measurement systems make low-temperature electrical data inaccessible. On top of this, this complexity makes the development of cryogenic electronics resource-intensive. Quanexum looks to overcome these barriers through its two complementary technologies: automated cryogenic measurement systems and cryogenic transistor models.
Traditionally, cryogenic measurements are done manually – a time-intensive and often inconsistent approach. Quanexum’s automated system simplifies and accelerates this process, enabling researchers to perform low-temperature electrical tests quickly and accurately. Where typical systems operate down to -40°C, Quanexum’s technology supports measurements as low as -270°C. Their transistor models offer researchers simulation tools compatible with standard software, but with the ability to function at extremely cold temperatures. Together, Quanexum’s technologies enable researchers and labs to develop advanced devices that would have previously needed industrial-grade resources.
Quanexum will use their grant to accelerate the development and market readiness of their advanced technology – paving the way for broader adoption across a range of industries.
Contact:
Scholé AI
Scholé AI – an EPFL startup project based in Professor Käser’s Machine Learning for Education lab – has been awarded a CHF 30,000 Ignition grant by the jury at the EPFL Startup Launchpad.
According to the World Economic Forum, by 2025, 50% of workers will need reskilling – particularly in areas such as AI and data science. Despite the plethora of adult learning courses available, businesses are still struggling to close this skill gap. Existing platforms are often static, generic, and disconnected from real world job tasks.
Scholé is developing a new training platform which relies on AI-driven curriculum. Through large language models, content adaptation and knowledge graphs, their platform tailors training to the specific job-centric needs and goals of an individual. Each user follows a personalised curriculum where they can choose between a variety of learning modalities – from quizzes, to podcasts, to video. The content is presented in bite-sized blocks constantly adapting to the user’s progress. Scholé’s training is also scenario-driven, allowing users to practice their skills in a realistic and risk-free environment. The team will focus their technology on B2B companies in industries with a high need for AI and data upskilling. They already have pilots and partnerships in place with some of Europe’s largest companies.
With their grant, the Scholé team will focus on the synchronous fine-tuning of its LLM, knowledge graph, and curriculum recommender system to validate and demonstrate the real-world impact of their novel approach.
Contacts:
Lemna Bio
Lemna Bio – an EPFL startup hosted by Professor Dal Pararo’s Laboratory of Biomolecular Modeling - has been awarded a CHF 100,000 Innogrant by the jury at the EPFL Startup Launchpad.
With ageing populations and chronic diseases on the rise, the demand for effective medicines and therapies has never been greater. Yet, for decades, drug development has relied on a costly and time-consuming trial-and-error approach. Developing a single drug takes over 10 years, costs upwards of $2 billion, and faces a staggering 90% failure rate.
Lemna Bio believes that AI with a deeper understanding of the molecular world can drive a golden age of biomedicine. At the heart of their platform lies the geometric transformer - an innovative AI technology rooted in physics-driven understanding of the molecular world. By enabling better representation of molecules, and their interactions, their technology facilitates better & faster design and development of biomolecules tailored to specific functions.
Lemna’s vision and validated models have already attracted interest from leading pharmaceutical and biotech companies. While their primary focus is advancing human health, they have also received interest for applications in agriculture, highlighting the versatility of their platform.
Lemna will leverage their Innogrant to conduct large-scale benchmarking of their technology and deepen engagement with industry stakeholders.
Contact: