How a semiconductor metaverse could accelerate chip innovation

Modern semiconductors can include billions of transistors. Manufacturing these chips requires a complicated dance between tools for growing wafers, generating the individual transistor features, cutting these into chips and then packaging them for products like cars, phones and data centers. Many of these tools are made by different vendors that are all improving simultaneously. Everyone needs to ensure that improvements in tools made by one vendor work with others. 

Welcome to the ‘semiverse’

At the Imec Future Summits conference, Lam Research’s CEO, Tim Archer, discussed work on a new semiconductor metaverse, called the “semiverse,” that could streamline this process in a keynote speech. 

“We envision the semiverse as a hybrid physical, virtual environment where development and testing are done as a joint human machine enterprise, where humans and machines are actually each doing what they do best,” Archer said. 

Lam Research has been a leader in semiconductor etching equipment since it first introduced a plasma etcher in the 1980s. In 2017, it acquired Coventor, a chip process automation and design automation solution. Archer’s vision is to extend these capabilities across the manufacturing supply chain to improve performance and sustainability. 

A complete digital representation of the entire chip fabrication process in the semiverse will allow chipmakers to look at the interactions across processes throughout the fab to improve repeatability and optimize yield. It will also enable greater collaboration to help the industry deliver disruptive technologies faster and more often. 

“It will vastly cut the time it takes from technology ideation to commercialization,” Archer said. 

Going virtual

Lam, ASML and Imec are already collaborating on technology for scaling transistor features down below 2nm widths in real-world physical laboratories. The semiverse will help extend this kind of collaboration into virtual laboratories. 

The semiverse will also allow enterprises building chips, research labs and chip equipment manufacturers to connect digital twins of chips, equipment, processes and labs in a shared virtual space. It allows researchers to test new ideas for a few dollars of computational time in virtual labs that might cost millions in physical labs. As a result, startups and researchers could contribute more disruptive ideas to the ecosystem.

Archer acknowledges that the industry is still a long way from the full realization of the semiverse. But Lam is working with others to put the foundational building blocks in place. For example, emerging model-based engineering tools are already helping to virtualize more of the engineering and testing process. 

Archer believes the improvements in CPU performance, new AI techniques and improved data science workflows will accelerate the ability to build more sophisticated predictive models across many steps in the chip-making process. 

“That’s a game changer for what we think we are going to be able to do here in building the semiverse,” Archer said.