Cognichip Raises $60M to Build AI That Designs AI Chips
Cognichip announced a $60 million funding round on April 1, 2026, led by Seligman Ventures. The company has raised $93 million total since its founding in 2024. The notable addition to the cap table: Intel CEO Lip-Bu Tan, who participated in the round and is joining Cognichip's board.
What Cognichip Actually Builds
This is not a general-purpose LLM applied to chip design. Cognichip trained a deep learning model specifically on chip design data. The distinction matters: domain-specific training on chip data is a different engineering problem than prompting an existing model with hardware constraints.
The company began collaborating with customers in September 2025. It has also developed procedures that allow chipmakers to train on proprietary design data without exposing it externally. That addresses a real concern for semiconductor companies whose process nodes and IP represent billions in R&D.
The Claims vs. the Context
Cognichip claims its technology can cut chip development costs by more than 75% and reduce timelines by more than half. Those are large numbers. The context makes them at least plausible targets: advanced chips currently take 3 to 5 years from conception to mass production, with the design phase alone running up to 2 years. Nvidia's Blackwell GPUs contain 104 billion transistors. The complexity is not hypothetical.
Whether Cognichip delivers on those claims at production scale is a different question. The company demonstrated its technology in a hackathon with San Jose State University electrical engineering students using RISC-V open source architecture. A controlled demo is a starting point, not a proof.
A Crowded Race
Cognichip is not alone. ChipAgents closed a $74 million extended Series A in February 2026. Ricursive raised a $300 million Series A in January 2026. Three companies, over $430 million raised collectively in roughly three months, all targeting the same bottleneck in semiconductor development.
Lip-Bu Tan's involvement is the most interesting signal. Intel has deep experience with the exact pain points Cognichip is targeting. His decision to join the board suggests he sees something credible in the approach, though board seats after investment rounds are not uncommon regardless of conviction.
Bottom Line
The semiconductor design problem is real, well-funded, and attracting serious attention. Cognichip has a specific technical approach, early customer relationships, and now a board member who runs one of the world's largest chip companies. The 75% cost reduction claim will need independent validation. For now, the company has enough runway to try.
Source: Techcrunch