An MIT-led team is developing a wearable patch intended to track biomarkers continuously while drawing power from body heat. MIT described the four-year PYTHIA project on November 4, with heart failure and sepsis among its intended targets.
ARPA-H’s award record lists MIT as the prime awardee, Ahmad Bahai as principal investigator and funding of up to $37.9 million, beginning September 24, 2026. The agency describes a clinical-grade sensor for heart-failure biomarkers with energy-efficient on-chip AI. The award ceiling is not a statement that the entire amount has already been spent.
MIT says the proposed device will measure low-concentration molecules in interstitial fluid, using binding chemistry and microelectronics to make signals detectable. On-chip models would turn the continuous measurements into information for clinicians.
Electronic-component development and testing come before planned preclinical and clinical trials. The announcement describes a development program, not a cleared diagnostic device, a demonstrated early-warning benefit or a product currently available to patients.
NAOAT analysis
Continuous measurement changes the information problem. A single laboratory result offers a snapshot; a wearable could produce a trajectory. That trajectory would be useful only if the measurements remain interpretable as the wearer moves, the sensing surface ages and conditions change.
The power source and local processing are central to that ambition. A patch intended for sustained use must balance the energy cost of sensing and interpretation against the energy it can harvest. Putting a model on the chip makes that balance part of the device design, rather than solely a software question.
The program brings semiconductor engineering and clinical validation into the same development path. Its eventual importance will rest on the connection between a measured signal and an actionable clinical decision. Funding establishes the opportunity to test that connection; it does not establish the result.