Chipforge

About

Why Chipforge

Legacy EDA was built for a different era, and it left the door open. Chipforge is a distributed APAC team walking through it: independent hubs, sovereign-capable, built AI-native from the start.

Map of the Asia-Pacific region highlighting Chipforge's hubs

Partnerships

Built on the world's #1 AI research.

Two decades of hardware-AI research from NTU Singapore, the engine under Chipforge's AI.

#1

For AI research worldwide (ShanghaiRanking 2025)

20+

Years of peer-reviewed AI-for-hardware research

36

Papers at DAC, ICCAD and DATE

NTU Singapore

Exclusive, perpetual licence

Chipforge holds the IP licence to NTU's hardware-AI research portfolio, including the graph-reasoning techniques behind Smeltr. NTU is ranked first in the world for AI in ShanghaiRanking's 2025 Global Ranking of Academic Subjects.

NVIDIA Inception

Ecosystem member

Part of NVIDIA's program for AI startups: technology access, engineering expertise and go-to-market ecosystem.

Chipforge difference

No baggage to drag forward.

Born post-legacy: every old tool drags decades of assumptions with it. Chipforge started from nothing, and built only what AI-native silicon design actually needs. Our agents don't just generate: they run a self-correction loop against real simulation and formal proofs, learn from every result, and work off a live graph of the whole design, all on hardware you control.

The old way

Chipforge

Hand-written RTL
Generated from intent
Manual month-long debugging
A self-verifying Reflexion loop
Spec drifts from code
A live Design Intent Graph
A dozen point tools
One flow, intent to silicon
Tools that never learn
Agents that improve every run

Verified in days, not months.

Come help us reinvent chip design

We're a small team teaching agents to design chips the way the best engineers do, only faster. If that excites you, come build the forge with us.

Want to learn more?

Tell us what you have in mind, and we'll take it from there.