Congress should pump $10B into supercomputing for AI work, think tank says

A think tank wants Congress to increase and diversity funding for supercomputing by reaching underserved states and minority and women AI researchers as well as two-year colleges. (Center for Data Innovation)

 

Center for Data Innovation called for $10 billion in federal funding over five years on Monday for high-performance computing to meet AI research demands.

The Washington-based think tank said the funds should targeted to regions in the US needing mid-range systems as well as minority-serving institutions with an eye to increasing HPC access for women.

Also, in its 40-page report CDI called for funding for HPC curricula at two-year colleges that offer transfers into four-year schools.

Significant AI research is being conducted in Alabama, Indiana, Utah and Georgia, but all four states lack sufficient HPC resources, Center for Data Innovation said.

“Limited access to high-performance computing is hampering AI researcher’s ability to develop new products and services that are vital for maintaining U.S. competitiveness,” said the report’s author, Hodan Omaar, a policy analyst at CDI.

 

supercomputing chart
Source: Center for Data Innovation

The report urges Congress to triple National Science Foundation funding for HPC to at least $500 million annually over the next five years and to triple the Department of Energy’s funding to at least $1.5 billion annually over that period.  NSF has seen a decade of funding cuts.

The report noted that DOE has increased investments for hardware, software and expertise for systems with the largest computational needs. However, there is still three times more demand than DOE has been able to support.

Other recommendations in the report include having NSF foster more public-private partnerships and getting both agencies to provide access to HPC experts to “improve researchers’ productivity.”

“Without sufficient access to HPC, researchers in a diverse range of science and engineering fields can’t address important research challenges,” Omaar added in a statement.

Center for Data Innovation is sponsoring a video webinar to the discuss the report and next steps for Congress on Thursday, Dec. 10, at 12:30 p.m. EST

RELATED: Trump bumps up quantum and AI research with private sector, national labs and academia