Chips are a fundamental component of the AI sector. But the new chips do not affect the market with the same speed as the new models and products AI.
Cognichip has a high goal of creating a fundamental artificial intelligence model that can help bring new chips to the market faster.
Cognichip based in San Francisco is working to build a model of AI of the physical informed foundation that can be used by semiconductors companies to accelerate the process of development of new chips. The company is calling this “artificial intelligence” approach and hopes that it can help accelerate the production times of the chips of 50% and also reduce the associated costs.
This ambitious idea comes from the veteran of the Faraj Aalai semiconductor sector, who has worked in various companies including Fujitsu network communications and Centillium communications.
Alaii told Techcrunch that the history of origin of his company begins in 2015. At the time, Alai was a member of the Silicon Valley Leadership Group, who often would have met to talk about what was suffering from their sector.
He was becoming worried about what was happening in the semiconductor sector. He held a presentation to the group on the strong decline in investments in risk capital in companies in semiconductors, who reached the peak of 200 agreements per year in 2000, he said, and went down to only one or two per year by 2015.
“I essentially warned the other CEO that this cannot be good for us,” said Alaii. “It cannot be positive for the semiconductor industry in America and that what we had to do is basically change things about it. If that continuous trend, then we will lose our competitiveness. We will lose the energy that brings new ideas to the table.”
Considering how long they take new chips to get to the market, it is not surprising that these companies did not attract venture capital investors, he said. So he sat on the idea for almost a decade.
He continued to found Candou Ventures in 2016 and through that fund he had the opportunity to attend the increase in startups. When it realized that the progress in artificial generative intelligence had reached a point where it could be used to potentially help some of the existing challenges of the semiconductor sector, decided to launch Cognichip in 2024.
Since then Cognichip Invisible Opera and has accumulated a team of artificial intelligence experts from places such as Stanford, Google and Mit to start building. Aalai said that it will take us at least a few years to build the model for “final performance”, but said that it should be able to help companies before it achieves this goal.
“When we arrive at that point, this artificial intelligence of the chip, we will build a system that can actually act as an expert engineer,” said Alaii. “Once you reach that vision, then you can actually do the same job with a fraction of people and in much, much shorter time.”
Cognichip is now emerging from furtivity with $ 33 million in seed funding in a round co-guided by Lux Capital and Mayfield with the participation of FPV and Candou Ventures.
Navin Chaddha, a CEO of Mayfield, told Techcrunch that when he was presented to Alaii, he felt he was “cut by the same fabric”. The vast majority of work in the semiconductor industry is still carried out by man; He said he thinks that timing is right to bring artificial intelligence to the mix.
“This is an important sore point and the solution that this company will provide will be a pain reliever and not a vitamin for the semiconductor industry,” said Chaddha. “If you don’t have humans who do the job, the IA can do it where there is a shortage of talents? Number one, great team, second, (they are) solving a real pain point in a massive industry from trillion dollars.”
Aalai said that he hopes that Cognichip will also be able to help democratize access to the construction of chips so that multiple semiconductor companies can start and investment on the ground. The easiest access also means that smaller companies can create more specific chips even for specialized or smaller models, he said.
All this will depend on when or if the company can reach artificial intelligence.
“What we are doing is not an incremental change,” said Alaii. “We are not building an instrument (Electronic Design Automation), we are not trying to change the process a little. We are trying to set a new goal for our sector and bring some serious changes.”
