Full story
Funding and Vision
Cognichip raised $60 million in a Series A round led by Seligman Ventures, bringing total funding to $93 million.
“Cognichip is building a deep learning model to work alongside engineers as they design new computer chips”
Intel CEO Lip-Bu Tan will join Cognichip's board.
The ACI platform is a physics-informed foundation model built specifically for semiconductor design.
The company claims it can reduce costs by more than 75% and shorten timelines by more than half.
Tackling a Complex Bottleneck
Chip design typically takes three to five years with the design phase alone taking about two years.
Cognichip replaces sequential workflows with parallelism.
Embedding physics into the AI model allows it to calculate tradeoffs data-driven tools cannot.
The platform integrates datasets, models, and infrastructure into a unified system.
Industry Implications
Cognichip is engaged with more than 30 semiconductor companies including many of the top 20 global players.
“Cognichip is already engaged with more than 30 semiconductor companies”
The startup faces an uphill battle in a market dominated by Synopsys and Cadence.
Analysts caution that physics-informed AI is unproven at scale.
Skepticism and Challenges
Cognichip cannot yet point to a new chip designed with its system.
Adoption cycles in semiconductor manufacturing are often measured in years.
Incremental improvements from entrenched players may prove effective.
Cognichip's Approach
Cognichip's ACI system integrates physical constraints and manufacturing realities.
“This allows it to reason across the full chip development lifecycle, from architecture to verification and production”
It reasons across the full chip development lifecycle from architecture to production.
Engineers operate more like system architects while the AI handles execution.