Synopsys CEO Sassine Ghazi has stated that fully autonomous AI chip design remains out of reach, according to a report by DigiTimes. The assessment from the head of the world's largest EDA vendor sets a hard boundary on what AI can currently deliver in semiconductor development.

The statement carries weight beyond one company's outlook. Synopsys supplies the design tools — including logic synthesis, place-and-route and verification software — that underpin chip development at TSMC, Samsung Foundry, Intel and most fabless design houses. When the CEO of that toolchain says autonomy is not achievable today, the judgment applies to the industry's entire design stack.

What did the Synopsys CEO say?

Ghazi said that fully autonomous AI chip design remains out of reach. The claim came in remarks reported by DigiTimes on the current state of AI in electronic design automation.

The position is notable because Synopsys itself invests heavily in AI-driven design tooling. The company has shipped commercial products that apply machine learning to chip implementation, and its customer base includes the leading-edge foundries and fabless semiconductor firms that face the highest design complexity. Even so, the CEO drew a clear line: AI assists the workflow; it does not run it end to end.

In plain terms: the largest EDA vendor confirms AI still needs human chip designers.

Why does the claim matter for the market?

The distinction between AI-assisted and AI-autonomous design shapes expectations across three groups:

• Chip developers — engineering headcount and expert oversight remain necessary for complex designs, with no near-term replacement by autonomous systems.
• Investors — projections of wholesale automation of semiconductor engineering lack support from the vendor that would build such capability.
• Tool vendors — the competitive frontier stays at augmentation: faster convergence, better optimization, fewer iterations — not full autonomy.

Modern system-on-chip designs involve billions of transistors, multiple verification stages and tightly coupled hardware-software dependencies. Each of those stages currently requires human judgment, from architectural choices through signoff. The CEO's statement indicates that this dependency persists despite rapid AI progress elsewhere in the semiconductor supply chain.

The message for buyers of EDA software is simple: budget for AI that accelerates engineers, not AI that replaces them.

What does this signal about AI's actual role?

Ghazi's framing positions AI as a productivity multiplier inside existing design flows rather than a substitute for them. That aligns with how EDA vendors have productized machine learning so far: point optimizations inside human-directed methodologies, with designers retaining control over architecture, verification closure and final signoff.

The CEO's verdict also tempers broader industry narratives. Public discussion of AI frequently assumes expanding autonomy across technical professions. The statement from the top of the EDA market says chip design — one of the most computational and data-intensive engineering disciplines — still resists that trajectory.

For now, the design of a leading-edge chip remains a human-led process.