A company called Agentrys says it is pioneering what it calls Agentic Design Automation, or ADA, and plans to present the approach at the 2026 Design Automation Conference (DAC), according to SemiWiki.
The pitch builds on a decades-old idea. SemiWiki notes that Electronic Design Automation, or EDA, transformed the semiconductor industry by automating the engineering algorithms used to design chips. Agentrys frames ADA as a new paradigm on top of that legacy: rather than only running automated algorithms, frontier AI can now automate many engineering tasks by understanding specifications and generating work from them, according to SemiWiki.
The broader picture, though, is still unsettled. Semiconductor Engineering reports that AI in chip design holds "lots of promise" but leaves "plenty of unanswered questions" — namely where AI will actually succeed, who stands to benefit, and how it will reshape the design process itself.
Those questions matter because designing an advanced chip is one of the most complex and expensive undertakings in modern engineering, involving armies of specialists and long timelines. If AI agents can reliably read a specification and carry out real design tasks, as Agentrys proposes, the economics and speed of building chips could shift. But as Semiconductor Engineering makes clear, it is not yet settled which parts of that workflow AI can genuinely handle, or who wins when it does.
Why it matters: the tools used to design semiconductors underpin nearly every electronic device, so a credible move to hand parts of that work to AI agents could ripple across the entire technology industry — if the promise holds up.