FemtoAI
Physical AI
Full-stack AI silicon and compression that fits large models onto low-power edge devices.
femtoAI builds a full-stack AI silicon and compression platform: the Sparse Processing Unit (SPU) accelerator, a software development kit, and pre-trained, pre-optimised models delivered through a developer portal. The approach draws on the neuromorphic concepts of sparsity and locality to cut power consumption by 10x and memory footprint by 10x, with minimal effect on performance, so larger models fit into low-power silicon - a claimed 10x model capacity and 10x energy efficiency. Unlike approaches that optimise only hardware or only software, femtoAI co-designs the two. Target applications include wearables, household appliances, smart assistants, smart glasses, baby monitors and robotics, with the SPU-001 accelerator available now and a developer community offering evaluation kits and pre-trained models. femtoAI was founded in 2018 by researchers from Stanford University's Brains in Silicon Lab and was previously known as Femtosense.
Research mapping
Physical AI / Edge Compute & Control Silicon (L5) / A
Evidence boundary
Original analysis and independent corroboration remain under development. Commercial scale and durable technical advantage remain under review.
What makes it defensible
Neuromorphic sparsity exploitation with hardware and software co-designed - 10x model capacity on 10x less power, which neither a chip nor a compiler can deliver alone.
Position in the stack: Silicon and toolchain supplier one layer beneath device OEMs. The moat is the co-designed stack of accelerator, SDK and pre-trained models, which is hard to replicate without both silicon and neural-network compression expertise.
Who buys it
Founding team
Sam Fok — Co-founder and CEO — Co-founded femtoAI (originally Femtosense) in 2018 out of Stanford University's Brains in Silicon Lab
SourceFinancing detail
- Last round
- Series A+
Investors: Fine Structure Ventures; SV Pacific Ventures; Acclimate Ventures; Allegro MicroSystems; Quest Venture Partners; Kleiner Perkins; J2 Ventures; Vanedge Capital; Amino Capital
Mapped funds and investors
Direct relationships supported by published evidence.
Reported holdings as-of
No sourced relationships mapped yet.
Disclosed investment history
No sourced relationships mapped yet.
Showing up to four relationships in each group. ETF weights describe the fund’s portfolio, not ownership of this company. Private investment history does not establish current ownership.
Evidence & media
-
femtoAI to Highlight its Sparse Processing Unit Edge AI Accelerator at CES 2026
What this source supports: femtoAI showcased its Sparse Processing Unit edge AI accelerator at CES 2026.
What it does not establish: Vendor claims; no independently verified performance data.
Source