Synopsys, one of the companies whose software the semiconductor industry relies on to design chips, says it has demonstrated a set of "autonomous engineering workflows" built with Nvidia technology — spanning, in the company's phrasing, everything from silicon to systems.
According to the announcement distributed by PR Newswire, the workflows cover the full arc of engineering work rather than a single isolated step. EE Times Asia frames the same news as Synopsys and Nvidia advancing autonomous AI agents for chip design and engineering — software that doesn't just suggest answers to a human engineer but carries out multi-step tasks on its own.
The sharpest claim comes from Stock Titan, which reports that Synopsys demonstrated AI agents that compress chip verification to hours. Verification is the painstaking process of proving a design actually behaves as intended before it is manufactured, and it is traditionally one of the slowest and most expensive parts of building a chip.
Running alongside the Synopsys news, Nvidia Developer published results claiming its Nemotron 3 Ultra leads open models on both accuracy and efficiency in agentic RTL coding. RTL — register-transfer level — is the code that describes what a chip's circuits do, so a model that writes it well is aimed squarely at the same bottleneck.
Why it matters: chip design is a gatekeeper for nearly every technology downstream of it, and if AI agents can genuinely shrink verification from a slog into a matter of hours, the pace at which new silicon reaches the market — including the chips that run AI itself — could shift.