What EnCharge AI does
EnCharge AI is a semiconductor startup developing analog in-memory computing (IMC) hardware and a matching software stack aimed at AI inference on “edge-to-cloud” deployments. The company’s core technical thesis is to reduce the energy and data-movement costs of inference by performing compute inside memory arrays using analog charge-domain techniques based on intrinsically-precise metal capacitors, which it says improves analog processing signal-to-noise tradeoffs compared with traditional IMC approaches.
More
EnCharge positions its solution as fully programmable for modern AI models, with system integration intended to let organizations run models closer to users (laptops, workstations, and edge devices) rather than relying exclusively on cloud compute. Commercially, EnCharge’s announced first product is the EN100 AI accelerator, described as an analog in-memory computing accelerator available in two form factors: an M.2 card for laptops/workstations and a PCIe card for workstations. EnCharge’s launch materials emphasize 200+ TOPS of total compute power and an 8.25W power envelope for the M.2 form factor, along with software support intended to integrate with common ML frameworks (the company specifically cites PyTorch and TensorFlow) and to support an early-access program for developers/OEMs. As of the latest public materials available on its site, EnCharge has raised multiple rounds including a $21.7M Series A (announced December 14, 2022) and an oversubscribed $100M+ Series B led by Tiger Global (announced February 13, 2025). In 2023 it also announced an additional $22.6M institutional round. In 2025, EnCharge also announced key executive and research leadership hires following its Series B, reflecting a stated shift from development toward commercialization. EnCharge’s business model is not explicitly spelled out in licensing terms in the sources accessed here; however, its product and early-access messaging suggests a hardware-plus-software platform approach—selling accelerator hardware in defined form factors while providing a supporting compilation/optimization ecosystem and integration resources for AI inference workloads.