AHyTEC - Turning low-grade waste heat into clean electricity
AHyTEC, an EPFL-based startup founded and led by Matteo Bevione from Prof. Giulia Tagliabue’s Laboratory of Nanoscience for Energy Technologies (LNET), has been awarded an EPFL Startup Launchpad Ignition grant of CHF 30,000.
Industrial processes release vast amounts of low- and ultra-low-grade waste heat (below 100 °C) that currently cannot be recovered cost-effectively. Existing energy-harvesting technologies are typically rigid, expensive, or reliant on scarce materials, limiting their scalability and economic viability in this temperature range.
AHyTEC is developing hydrogel-based thermoelectrochemical cells that convert low-temperature waste heat directly into electricity. The technology is flexible, low-cost, and based on abundant materials, enabling easy integration onto pipes, machinery, and curved industrial surfaces. Early prototypes demonstrate power generation from temperatures as low as 40–80 °C, opening new opportunities for energy recovery where no practical solutions exist today.
The Ignition grant will support the development of an industrial-grade prototype and pilot testing in real operational environments, with the goal of identifying high-value use cases and accelerating market readiness.
Low-grade waste heat is one of the largest untapped energy resources in industry. Our goal is to turn what is currently lost into clean, usable power for the future.
Matteo Bevione, founder of AHyTEC.
Contact: Matteo Bevione
KIIO - From reactive to proactive: smarter process monitoring for a resilient industry
Originating from Prof. Colin Jones’ Automatic Control Laboratory at EPFL, KIIO has received an EPFL Startup Launchpad Ignition grant worth CHF 30,000.
Chemical and other process manufacturing plants operate 24/7, and each hour of unplanned downtime can cost over $100,000. To keep plants running continuously, machines are monitored by sensors. However, studies show that over 80% of equipment failures are caused by abnormal processes rather than equipment end-of-life. By the time such faults are detected, it is often too late for operators to act. Today, process diagnosis is still largely manual: it takes a team of specialists to enumerate, test, and eliminate hypotheses one by one - a time-consuming and labor-intensive procedure that can take weeks or months.
This challenge is amplified by the fact that 20% of the chemical industry workforce will retire by 2030, taking decades of operational knowledge with them. New operators urgently require intuitive, explainable diagnostic support to ensure efficient and reliable operations. Currently, no effective solution exists for process-level monitoring and diagnosis to improve reliability. The result is lost productivity: chemical plants run at an average of only 41% overall equipment effectiveness, meaning that half of the potential capacity is lost to downtime, inefficiencies, or off-spec production - translating into hundreds of millions in annual costs for a typical large plant.
KIIO has developed technology that shifts monitoring from isolated assets to the full process by treating the plant as a connected network. By learning the relationships among sensor signals and tracking subtle changes as they occur, KIIO detects abnormal processes earlier - before they develop into equipment faults. The technology does not rely on fault libraries (a database of known failures) but learns directly from real-time data. Importantly, KIIO doesn’t just point out deviations from the norm; it also shows how a fault propagates through the process and helps engineers and operators understand where a problem may have started and how it can be mitigated.
The team will use the Ignition grant to create a working prototype and validate it in a realistic end-to-end scenario.
Contact: Mengjie Zhao
