EnCharge AI

Company spun out of Naveen Verma's Princeton lab, maker of the EN100 analog in-memory-compute inference chip. Its architectural pitch inverts the dominant analog-AI story: instead of resistive-memory (RRAM/memristor) conductance, it multiplies via lithography-defined switched capacitors, trading device-state precision for CMOS geometry precision. As of 2026-07-31 the EN100 is still gated behind an Early Access Program, and every published efficiency figure (200 TOPS at 8.25 W, "up to 20x" performance-per-watt) is company-sourced with no independent benchmark or stated comparison baseline located — the load-bearing open question for this cluster.

References

• claim-encharge-en100-switched-capacitor-in-memory-compute · claim-no-independent-benchmark-of-encharge-en100-efficiency-figures · claim-encharge-en100-20x-per-watt-claim-has-no-stated-baseline · claim-2021-isscc-verma-lab-chip-is-primary-precedent-not-en100-evidence
• question-verify-encharge-en100-efficiency-primary
• Related hubs: entity-naveen-verma