Designing Chips Overnight Using AI
12:48, 15.04.2026
NVIDIA recently announced that it has managed to reduce the chip design process to a single night.
Key Aspects of Using AI in Projects
Bill Daley, a research scientist at the company, told a colleague at Google that NVIDIA is already using AI in certain stages of chip design. It has now been revealed that AI is used for verification, error handling, design option analysis, and other tasks. However, predictions regarding fully autonomous chip design are not yet very realistic.
For example, previously, porting a cell library to a new semiconductor process took approximately 10 months of work and a team of 8 employees. AI-based tools now complete this task in just one night.
Additionally, the company utilizes the prefix RL - an internal AI-based tool. AI can generate new designs, and key performance metrics improve by more than 20% compared to designs created by human engineers.
The company uses internal LLM models named Bug Nemo and Chip Nemo. These models were built using NVIDIA’s proprietary technology. The LLM is used by the company’s junior specialists and is also necessary for summarizing error reports and classifying them.
Hope you found this article helpful - what do you think? Like and subscribe to our blog to get more practical tips and the latest tech news from HostZealot.