Are you ready to bring more awareness to your brand? Consider becoming a sponsor for The AI Impact Tour. Learn more about the opportunities here.
Imec, a semiconductor research accelerator in Leuven, Belgium, announced the first fruits of the Sustainable Semiconductor Technologies and Systems (SSTS) research program aimed at reducing the semiconductor industry’s carbon footprint. Early participants include large system companies such as Apple and Microsoft and semiconductor suppliers such as ASM, ASML, Kurita, Screen and Tokyo Electron.
Today, the semiconductor industry plays a small role in global carbon emissions. But the industry is facing unprecedented demand for new chips. A big concern is that growing demand could increase the semiconductor industry’s relative proportion of carbon emissions, particularly as other industries reduce their carbon impact.
Semiconductor manufacturing requires large amounts of energy and water and can create hazardous waste. In the short run, the SSTS effort will help companies understand and optimize their systems to minimize carbon emissions. Eventually, better metrics could also help mitigate other problems caused by semiconductor production.
A growing concern in the semiconductor industry
Researchers have shown that nearly 75% of a mobile device’s carbon footprint comes from its fabrication and half of this comes from the chips. The SSTS team has already built a virtual fab to simulate the environmental impact of future logic and memory devices.
The AI Impact Tour
Connect with the enterprise AI community at VentureBeat’s AI Impact Tour coming to a city near you!
The SSTS team is also starting to enrich and simulate data with and check the representativeness of models used. Lars-Åke Ragnarsson, program director at SSTS, told VentureBeat that the early efforts are prototypes of simulated fabs rather than true digital twins. However, the hope is that future efforts will allow enterprises to weave these models into digital twins to help optimize ongoing manufacturing operations.
He stressed the importance of sharing emissions’ data across the entire manufacturing lifecycle and all steps in the supply chain. By developing metrics, everyone will get a better idea of what actions can improve sustainability versus lessen it. For example, today, some companies may quantify carbon footprint on a per-chip basis, while others do it on a per dollar of product basis. Unified standards will provide an apples-to-apples comparison on the merits of various adjustments.
Some different sources of carbon emissions include electrical consumption, direct emissions and water usage. The group is already showing the potential impact of higher efficiency pumps and recovering hydrogen. Down the road, they are hoping to identify high-impact processes for logic and memory technologies where manufacturers should focus efforts. They also want to develop guidelines for manufacturers, fabless companies, equipment providers and material providers that could help minimize carbon impact.
The project is still in its infancy. Currently, chip manufacturing only accounts for about 0.1% of global carbon emissions. But the concern is that without any new approach, they could grow by three times by 2030 and have an outsized impact on carbon emissions as other industries reduce their carbon footprint. “The goal should be to say at 0.1%,” Ragnarsson said.
They plan to add other environmental aspects, such as material costs, down the road. Other tools will make it easier to explore how new manufacturing and computing techniques such as molecular computing might impact the climate footprint. They also plan to develop tools to help share these findings and broader lifecycle data with other stakeholders such as politicians and citizens.
“We need to work together as a team,” Ragnarsson said. “This is not competitive. We need to work together to make this happen.”
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.