Chips are a key component of the AI ​​industry. However, the new chips are not on the market at the same speed as the new AI models and products.

Cognichip's goal is to create a basic AI model that can help bring new chips to market faster.

San Francisco-based Cognichip is working to build a basic AI model with physical information that can be used by semiconductor companies to speed up the development process of new chips. The company calls this approach “artificial chip intelligence” and hopes it can help speed CHIP production time by 50% and reduce associated costs.

This ambitious idea comes from the semiconductor industry veteran Faraj Aalaei, who has worked for a variety of companies including Fujitsu Network Communications and Centimeter Communications.

Aalaei told TechCrunch that the origin story of his company began in 2015. At the time, Aalaei was a member of the Silicon Valley Leadership Leadership Group, which often met to talk about what was bothering their industry.

He is increasingly worried about the situation in the semiconductor industry. He said he gave the group a speech on the decline in star investments from venture capital investments to semiconductor companies, which peaked at $200 in 2000 and only one or two per year by 2015.

“I essentially warn another CEO that it’s not good for us,” Aalaei said. “It’s not good for the semiconductor industry in the U.S., and all we need to do is fundamentally change it. If this trend continues, then we’re going to lose competitiveness. We’re going to lose the energy to bring new ideas to the table.”

He said it is not surprising that the companies are not attracting venture investors, considering how long it will take for new chips to go public. Then he sat on this idea for nearly a decade.

He went on to build Candou Ventures in 2016, through which he had to watch the rise of AI startups. When he realized that the advancements in generating AI had reached a point where he could be used to help solve the challenges that exist in certain semiconductor industries, he decided to launch Cognichip in 2024.

Since then, Cognichip has been invisible and has accumulated a team of AI experts from places like Stanford, Google, and MIT to start building. Aalaei said it would take at least a few years to build a model of “final performance”, but said it should be able to help before the company reaches that goal.

“When we get to that point, this artificial chip intelligence, we’re going to build a system that can actually be like an expert engineer,” Aalaei said. “Once we achieve this vision, you can actually do the same with a small number of people, and it’s much shorter.”

Cognichip is now involved with $33 million seed funding from Lux Capital and Mayfield, and is involved from FPV and Candou Ventures.

Navin Chaddha, managing partner at Mayfield, told TechCrunch that when he was introduced to Aalaei, he felt they “cut from the same piece of cloth.” Humans are still doing most of the work in the semiconductor industry. He said he thought it was right to bring AI into the mixing time.

"This is a major pain point, and the solution the company will offer will be painkillers, not vitamins in the semiconductor industry," Chad said. "If you don't have humans doing this job, can AI do it in a place where talent is lacking? Ranked first, great teams, and second, (they) solve a real pain point in a huge, trillion-dollar industry."

Aalaei said he hopes Cognichip can also help democratize access to building chips so that more semiconductor companies can get started and invest in land. Easier access also means smaller companies can also build more specific chips for professional or smaller models, he said.

All of this depends on when or if the company can reach artificial chip intelligence.

“What we are doing is not gradual change,” Alay said. “We are not building (electronic design automation) tools, but we are not trying to make some adjustments to the process. We are trying to set new goals for our industry and bring some significant changes.”