Cognichip wants AI to design AI-powered chips, and just raised $60M for the venture

Advanced silicon chips have accelerated the development of artificial intelligence. Now, can AI return the favor?

Cognichip is building a deep learning model to work closely with engineers as they design new computer chips. The problem it’s trying to solve is one the industry has struggled with for decades: chip design is too complex, too expensive, and too slow. Advanced chips take three to five years from conception to mass production; the design phase alone can take up to two years before bodybuilding begins. Consider that the latest line of Nvidia GPUs, Blackwell, contains 104 billion transistors – that’s a lot to do.

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

“These programs are now smart enough that just by directing them and telling them what the desired result is, they can generate good code,” Aalaei told TechCrunch.

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

The company came out of the blue last year and said on Wednesday it had raised $60 million in new funding led by Seligman Ventures, with notable participation from Intel CEO Lip-Bu Tan, who invested through his venture firm Walden Catalyst Ventures and will join Cognichip’s board. Umesh Padval, managing partner at Seligman, will also join the board. Cognichip has now raised $93 million in total since its founding in 2024.

However, Cognichip cannot identify the new chip designed for its system and has not disclosed any customers it says it has been working with since September.

The company says its advantage lies in using its model trained on chip design data, rather than starting with a general-purpose LLM. That required access to domain-specific training data, which is not trivial. Unlike software developers, who share vast amounts of code freely, chip designers guard their IP closely, making the kind of open-source trove that often trains AI coding assistants less available.

Cognichip has had to develop its own datasets, including manufacturing data, as well as license data from partners. The company also developed processes to allow chipmakers to safely train Cognichip models on their proprietary data without disclosing it.

When proprietary data is not available, Cognichip relies on open source alternatives. In one demo last year, Cognichip invited electrical engineering students at San Jose State University to test the model at a hackathon. The teams were able to use the model to design CPUs based on the open source RISC-V chip design – a freely available design that anyone can build on.

Cognichip competes with existing players like Synopsy and Cadence Design Systems, as well as a crop of well-funded startups. Among them: Alpha Design AI, which raised a series A of $21 million in October 2025, and ChipAgentsAI, which closed an extended series A of $74 million in February.

Padval said the current flood of money in AI infrastructure is the biggest he has seen in 40 years of investing.

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