Cognichip wants AI to design the chips that power AI, and just raised $60M to try

The most advanced silicon chips have accelerated the development of artificial intelligence. Can AI now return the favor?

Cognichip is building a deep learning model to collaborate with engineers in designing new computer chips. The problem she is trying to solve is one that the industry has been living with for decades: chip design is enormously complex, ruinously expensive and slow. It takes three to five years for advanced chips to go from concept to mass production; the design phase alone can take up to two years before the physical installation begins. Consider that the latest line of Nvidia GPUs, Blackwell, contains 104 billion transistors – that’s a lot to sort out.

In the time it takes to create a new chip, says Faraj Aalaei, CEO and founder of Cognichip, the market can change and turn all that investment into a waste. Aalaei’s goal is to bring the kind of AI tools that software engineers have used to accelerate their work to the semiconductor design space.

“These systems have now become intelligent enough that just by guiding them and telling them what the desired outcome is, they can actually produce beautiful code,” Aalaei told TechCrunch.

He says the company’s technology can reduce the cost of chip development by more than 75% and shorten the timeline by more than half.

The company emerged from stealth last year and said Wednesday it had raised $60 million in new funding led by Seligman Ventures, with notable participation from Intel CEO Lip-Bu Tan, who is joining Cognichip’s board of directors. Umesh Padval, managing partner at Seligman, will also join the board. Cognichip has raised a total of $93 million since its founding in 2024.

Still, Cognichip cannot yet point to a new chip designed with its system and has not disclosed any of the customers it says it has been working with since September.

The company says the advantage lies in using a proprietary model trained on chip design data, rather than starting with a generic LLM. This required accessing domain-specific training data, which is no small feat. Unlike software developers, who share large amounts of code openly, chip designers closely guard their IP, making the kind of open source resource that typically trains AI coding assistants largely unavailable.

Cognichip has had to develop its own data sets, including synthetic data, and licensing data from partners. The company has also developed procedures that allow chipmakers to safely train Cognichip’s models on their own proprietary data without exposing it.

Where proprietary data is not available, Cognichip has turned to open source alternatives. In a demo last year, Cognichip invited electrical engineering students from San Jose State University to try out the model during a hackathon. The teams were able to use the model to design CPUs based on the RISC-V open source chip architecture – a freely available design that anyone can build on.

Cognichip competes with established players like Synopsys and Cadence Design Systems, as well as well-funded startups like ChipAgents, which closed an extended $74 million Series A round in February, and Ricursive, which raised a $300 million Series A round in January.

Padval said the current capital flow into AI infrastructure is the largest he has seen in four decades of investing.

“If it’s a super cycle for semiconductors and hardware, it’s a super cycle for companies like [Cognichip]” he said.