Chinese EDA Firm Says Agentic AI Cut Some Chip Design Tasks by 75%

Chinese chip design software company Empyrean Technology says it has reduced the time needed for some circuit design tasks by as much as 75% by using agentic AI.

According to company chairman Liu Weiping, one circuit layout task that previously required about four weeks was completed in roughly one week with the help of AI agents.

The claim highlights a broader push across the semiconductor industry to use AI for electronic design automation, or EDA, where engineers rely on specialized software to design, simulate and verify complex chips.

Detail	Reported information
Company	Empyrean Technology
Industry	Electronic design automation
Technology	Agentic AI
Reported design time reduction	Up to 75%
Example task	Circuit layout
Previous time	About 4 weeks
New time	About 1 week
New platform	Agentic EDA system under development

Agentic AI is being used for specific design tasks

Agentic AI differs from conventional assistant software because it can carry out multi step tasks with greater autonomy.

In semiconductor design, that could include helping engineers optimize layouts, test alternatives, check design rules or move between several stages of a workflow with less manual intervention.

Empyrean says this approach has already produced major time savings in at least some circuit design work.

The company is also developing its own agentic EDA platform, suggesting that the technology could become a more formal part of its future software portfolio rather than remain an internal experiment.

The 75% reduction should still be treated as a company reported result rather than a general expectation for every chip design project. Different workloads can vary widely in complexity.

AI is becoming a larger part of EDA software

Empyrean is not alone in pursuing this direction.

Several major global EDA companies have already announced agentic AI tools intended to automate or accelerate parts of semiconductor design.

Startups are entering the field as well, with companies working on AI systems that can assist with circuit generation, verification and optimization.

The attraction is clear.

Modern chips can contain billions of transistors, and design teams spend large amounts of time testing layouts, correcting errors and balancing power, performance and area.

Even modest improvements in those processes can reduce engineering time and shorten development schedules.

China sees EDA as a strategic part of chip independence

EDA is especially important for China because much of the global market has historically been dominated by companies based in the United States and Europe.

China has been investing in domestic alternatives as part of a wider effort to reduce reliance on foreign semiconductor technology.

That effort extends beyond chip design software.

Chinese companies are also investing in logic manufacturing, memory production, semiconductor equipment and new techniques for producing more advanced chips with deep ultraviolet lithography.

AI could help narrow some development gaps

Agentic AI will not remove the need for advanced fabrication equipment, but it could improve efficiency in parts of the chip development process that depend heavily on software and engineering work.

For domestic EDA companies, faster design workflows may help customers move through iteration cycles more quickly and reduce the time required to complete certain stages of development.

That could be particularly useful while China continues building a broader local semiconductor ecosystem.

The reported four week to one week improvement provides an early example of what AI assisted chip design might offer, but wider evidence will be needed to show whether similar gains can be repeated across large commercial semiconductor projects.

For now, Empyrean’s claim shows how quickly agentic AI is moving into EDA, an area where automation could have a direct effect on the cost and speed of developing future chips.