Neurons on a Chip: How Archean Sciences Is Rethinking Drug Development and Computing at Once

TechNexus invested $100,000 in Archean as part of the University of Illinois's Cozad New Venture Challenge

More than one in three people live with a neurological condition, the leading cause of ill health and disability worldwide. The drug candidates developed to treat conditions like Alzheimer's, Parkinson's, and MS fail in clinical trials at a rate above 90 percent, largely because the preclinical models used to test them don't reflect how the human brain actually works. Archean Sciences , a startup spun out of the University of Illinois Urbana-Champaign, is building a better model. The technology it is using to do so — real human neurons grown on a chip and put to work as a computing substrate — may end up reshaping two industries at once. Archean’s innovative approach was on display at the University of Illinois Urbana-Champaign's Cozad New Venture Challenge, where the company won the $100,000 grand prize — funded by TechNexus Venture Collaborative. TechNexus's investment was the competition’s first-ever six-figure grand prize. TechNexus was joined by Illinois Ventures and Origin Ventures, who each made a $10,000 investment in Archean. Archean was founded by Austin Ellis-Mohr, a PhD graduate in electrical and computer engineering, and Mattia Gazzola, a MechSE associate professor at the University of Illinois. Their core technology: real human neurons grown on a chip, connected to electronics that can read and interpret the cells' electrical activity in real time — and do so autonomously, at scale. The immediate application is drug development for neurological conditions. The numbers that frame the problem are striking. More than a third of the population suffers from a neurological disorder. And somewhere between 90 and 99 percent of drug candidates that enter clinical trials fail — often after years and billions of dollars have already been spent. Much of that failure traces back to the limitations of today's preclinical models: animal studies that don't fully translate to humans, and in vitro cell tests that offer only crude readouts. "If 90 to 99 percent of things

By Jim Dallke at TechNexus Venture Collaborative