This founder is teaching chips how to recycle (their energy)

Details

Reversible computing — the idea that chips could run calculations backward to recover energy otherwise wasted as heat — has been theoretical since the 1970s. Vaire Computing, a startup cofounded by CTO Hannah Earley (31), last year demonstrated a chip resonator that achieved net-positive energy recovery for the first time, recovering more energy than the component itself consumes. Conventional chips erase intermediate computational states as they work, dissipating that energy as heat — Earley compares it to braking at every intersection. Vaire's resonator retains those intermediate states, allowing energy to be recaptured for reuse. The component is described as "a glorified pendulum" that stores recovered energy at the microscopic scale. Researcher Igor Markov (University of Michigan) called the result "interesting" but noted the technology is early-stage and will need a series of increasingly convincing demonstrations before attracting industry support. The startup emerged from Earley's Cambridge PhD work; no funding round or commercial timeline is disclosed. For AI, the relevance is indirect but potentially significant: AI inference workloads are among the fastest-growing electricity consumers globally, making chip-level energy recovery a meaningful long-term lever — if the technology scales.