NSF awards $36M to decarbonize computing

The National Science Foundation will be awarding $36 million in funding to three energy-efficient computing projects, part of a larger agency program aimed at reducing the carbon footprint created by modern computing needs.

Announced on Thursday as part of the NSF’s Expeditions in Computing program, the three projects feature leading collaborators from academic institutions: Harvard University, the University of Massachusetts Amherst and the University of Texas at Austin.

Each project focuses on decarbonization efforts, or reducing the amount of carbon emissions generated through energy use required to fuel computer infrastructures. 

The project helmed by researchers at the University of Texas at Austin will work to leverage artificial intelligence and machine learning systems to help optimize computing performance and reduce the energy consumption within cloud computing softwares. 

Researchers at Harvard University will use the funding to support its Carbon Connect research initiative that aims to establish new standards for carbon accounting in the computing industry.

And the University of Massachusetts Amherst will work to develop a new computational decarbonization field focused on “optimizing and reducing the lifecycle of carbon emissions of complex computing and societal infrastructure systems.”

“We are thrilled to announce these visionary projects that will advance environmental responsibility and foster innovation in the field of computing,” said Dilma DaSilva, acting assistant director for the NSF Directorate for Computer and Information Science and Engineering. “Congratulations to these pioneering teams whose research will forge new pathways in computational decarbonization and in revolutionizing operating system design with machine learning.”

The Expeditions program was established within the NSF in 2008 and focuses on delivering transformative systems that could be applied to broader research communities and infrastructures. 

NSF Director Sethuraman Panchanathan discussed his agency’s ongoing research initiatives before the Senate Appropriations Subcommittee on Commerce, Justice, Science, and Related Agencies on Thursday as well, underscoring his agency’s contributions to the U.S. scientific innovation landscape as vital to national security efforts. 

“When we look at the early investment in an idea that revolutionizes an industry and yields incredible profit, we call that a wise investment,” Panchanathan said. “NSF continues to be an extremely wise investment for the United States. Every single American benefits every day from our commitment to unleashing the American spirit of innovation.”

He added that high performance computing research is one of the topics NSF will be focusing on in the coming year, as he testified on budget estimates for fiscal year 2025. In addition to the academic partnerships, as seen in the new Expeditions series funding, Panchanathan said that industry partners will play a big role in marrying emerging systems, like AI and machine learning, to high performance computing.

“It's a constant partnership with industry so that the latest and, in fact, the more advanced technologies are what we are deploying,” he said. There's tremendous computing needs right now on so many dimensions, whether it is climate mitigation or whether it is AI, whether it is quantum, the modeling needs are tremendous in every aspect of science and engineering. So it's important that we invest, but we also … need to partner with industry very closely.”

The new NSF funding follows broader Biden administration objectives to use the federal government’s vast financial resources to further sustainability efforts, including an April rule requiring federal buyers to procure sustainable products and services “to the maximum extent possible.”