AI Chip design now surpassing human efforts
Artificial intelligence has already been being used for chip design. During it’s first integrations, it generated a lot of acadmic controversy. Despite the early pushback from the entrenched chip design players, new developments are showing that AI silicon development is here to stay by dramatically speeding up design and creating better chips than traditional methods.

Most recently, we see that AI is revolutionizing wireless chip design. It is dramatically reducing costs by cutting design times from weeks to hours, and it’s also generating unconventional designs that are faster and have unexpected circuitry patterns that are significantly outperforming traditional chip designs. Even the researchers aren’t sure how they are so much better.
We are coming up with structures that are complex and look random shaped and when connected with circuits, they create previously unachievable performance. Humans cannot really understand them, but they can work better.
lead researcher Kaushik Sengupta
Circuits are engineered to be more energy-efficient or operable across wider frequencies. Conventional algorithms sometimes take weeks to solve the complex and opposing structures. Some combinations do not have traditional solutions. Yet, new AI methods are solving these structures in minutes – and even solving problems that have previously been impossible.
This work presents a compelling vision of the future. AI powers not just the acceleration of time-consuming electromagnetic simulations, but also enables exploration into a hitherto unexplored design space and delivers stunning high-performance devices that run counter to the usual rules of thumb and human intuition.
Uday Khankhoje – associate professor of electrical engineering at IIT Madras
They also have discovered that AI started generating complex electromagnetic structures that are co-designed with circuits to create broadband amplifiers in a unique discovery. Beyond discovering new patterns, these tools are also improving designer’s lives by removing the more mudane, utilitarian work.
The researchers note that like many AI algorithms – they aren’t perfect. The AI often hallucinates faulty elements that don’t work are very inefficient and require human intervention. But the time saved far outweighs the cost.
Articles:
- Deep-learning enabled generalized inverse design of multi-port radio-frequency and sub-terahertz passives and integrated circuits” by Emir Ali Karahan, Zheng Liu, Aggraj Gupta, Zijian Shao, Jonathan Zhou, Uday Khankhoje and Kaushik Sengupta, 30 December 2024, Nature Communications.
DOI: 10.1038/s41467-024-54178-1 - https://scitechdaily.com/ais-strange-chip-designs-are-faster-smarter-and-game-changing/