The Future of Invisible Edge AI

Power-efficient Analog In-Memory AI for every sensor on the planet

Scroll

The Problem

Sensors everywhere. Intelligence nowhere.

The world is drowning in data from billions of sensors — but almost none of it is processed where it matters most.

100-500ms latency

Cloud AI

Data round-trips to distant servers. Latency kills real-time decisions.

Digital Chips

Power-hungry processors generate heat. Unsustainable for edge deployment.

Analog Rivals

Incomplete solutions. Missing the full-stack integration needed for production.

0BIoT devices by 2030

Almost none with local intelligence

Our Approach

The bicycle motor for AI

Analog in-memory computing performs matrix multiplications directly in memory — the core operation behind all neural networks.

Analog Matrix Multiplication

10×less power

Analog computation eliminates the digital overhead that wastes energy on every operation.

<1–3Wruns on coin battery

Deploy intelligence in sensors with no external power supply needed.

100%on-device

No cloud dependency. No data leaves. Privacy and speed by design.

Technology

Built from first principles

A vertically integrated stack — from silicon to software — designed for maximum efficiency at the edge.

AI Inference
Circuit Layer
Silicon Die
1

Proprietary Analog Circuit Design

Custom-engineered circuits that perform neural network computations in the analog domain — bypassing the von Neumann bottleneck entirely.

2

Accelerated Matrix Multiplication

Physics-based computation at the speed of electron flow. No clock cycles, no memory fetches — just instant parallel processing.

3

Full-Stack Hardware–Software Co-Design

From transistor-level design to compiler toolchains — a vertically integrated stack optimized end-to-end for edge AI workloads.

Team

Silicon-native team

Engineers who think in transistors and dream in architectures.

F

Founder & CEO

MS VLSI, ex-semiconductor R&D

H

Head of Analog & Digital

Full-Stack Silicon Architect

PhD Microelectronics, 15+ tape-outs

A

Head of AI

Edge ML Optimization

PhD Machine Learning

Advisors

Senior Advisor

45+ years in semiconductor industry

Former VP Engineering, Global Foundry

Other Technical Advisor

Analog IC design pioneer

100+ patents, IEEE Fellow

Deep roots in silicon — from transistor to tape-out

MS/PhD engineers from top research programs

Advisors with 45+ years experience in chip design

Market Timing

Why now

Four converging forces make this the perfect moment to build analog edge AI.

2023

Analog IMC Breakthrough

IBM demonstrates 14nm analog in-memory chip achieving record energy efficiency for AI inference. The science is proven.

14nmprocess node
01
2024–26

Edge AI Demand Explosion

The edge AI market is projected to grow 20×. Every autonomous system, wearable, and industrial sensor needs on-device intelligence.

$65Bmarket by 2030
02
Now

Affordable Mature Fabs

130nm and 65nm fabs are widely available and cost-effective. No need for bleeding-edge nodes — analog thrives on mature process.

130nmfab availability
03
2030

Industry 4.0 Sensor Boom

Smart factories, predictive maintenance, autonomous vehicles — all demanding real-time AI at the sensor level with minimal power.

41Bconnected devices
04

AI belongs at the edge.

We're building the silicon that makes it possible. Join us in redefining how intelligence reaches every corner of the world.