EnCharge AI is pursuing a fundamentally different chip architecture to break the energy and memory bottlenecks of running AI. Its analog in-memory computing approach performs calculations directly within memory using analog techniques, sidestepping the data-movement overhead that makes conventional digital AI accelerators power-hungry.
The company claims its accelerators can be up to 20 times more energy efficient than leading digital AI chips, a profile especially valuable for client computing and the edge, where power and thermal budgets are tight. EnCharge targets laptops, PCs, workstations and other devices that need to run capable AI models locally without draining batteries or requiring the cloud.
EnCharge was founded in 2022, spun out of Princeton University from the lab of CEO Naveen Verma, a professor of electrical and computer engineering. Co-founders include Kailash Gopalakrishnan and Echere Iroaga, and the company has DARPA backing reflecting interest in its efficiency claims. Its technical roots in years of academic research underpin the architecture.
In February 2025 EnCharge closed an oversubscribed $100 million Series B led by Tiger Global, with participation from Samsung Ventures and HH-CTBC, bringing total funding above $144 million. The company planned to tape out its first production chips for mobile, PCs and workstations and bring its first accelerators to market.
EnCharge's bet is that as inference moves on-device, energy efficiency, not raw peak performance, becomes the decisive metric, and that analog in-memory computing is the architecture best suited to win there.