On the count of three, Gov. Abigail Spanberger, Sen. Mark Warner, Newport News Mayor Phillip Jones and other officials lifted their shovels and drove them into the dirt. People seeking shade under a nearby tent applauded. The ritual marked the next phase of a multi-million dollar effort called the High Performance Data Facility officials say is the first of its kind.
“The HPDF will expand the nation's capacity to manage, move and prepare scientific data for the next generation of discovery serving researchers across the country at a scale we have never seen before,” said Gov. Spanberger at the groundbreaking. “And because Jefferson Lab leads that hub, Virginia is at the center of it all.”
The U.S. Department of Energy chose the Thomas Jefferson National Accelerator Facility Lab to lead its High-Performance Data Facility in 2023. The new DOE-funded project is expected to cost up to $500 million. Once it’s up and running in 2030, the data hub — outfitted with powerful computing infrastructure — will be a key component of the DOE’s AI initiative, Genesis Mission.
Friday’s groundbreaking was for the Jefferson Lab Data Center, which will house the future data facility. The state committed $49 million to design and fund the building.
Despite what the names may suggest, no data will be stored there.
“Moving data is super expensive and complicated,” Jefferson Lab Director Jens Dilling told WHRO ahead of the groundbreaking.
Instead, powerful computers in the facility will provide access to data generated and stored at the Department of Energy’s 28 research centers scattered around the nation.
Across all these sites, the data facility will manage exabytes of data. For scale, “if a terabyte can store the contents of all the libraries in Virginia, a single exabyte could hold the text of every book ever written in human history,” Jefferson Lab spokesperson Duane Bourne wrote in an email.
And the facility will be able to move all the data relevant to researchers’ questions up to 10,000 times faster than typical home internet.
If a researcher wants to know how the first life-forming molecules in the universe formed, for instance, Dilling said the person would need microscope data collected by biologists and telescope data collected by astronomers.
“One looks into the universe, one looks into biology,” he said. “If you want to find out how biology is generated in the universe, it would be useful to combine the data.”
But these data sets are usually separated by geography and silos between scientific communities.
“They have different scientific language, different scientific approaches of how they take data, how they pose a hypothesis, how they process data,” Dilling said.
The High-Performance Data Facility will use AI technology to bridge those gaps and combine data sets to solve mysteries that have stumped generations of scientists, Dilling said.
“What is at the core of nuclear physics, and can we solve that once and for all?” Dilling said, posing questions he hopes the new facility will help answer.
Darío Gil, the under secretary for science at DOE, called the facility “the place where our data, instruments and computing will meet and become usable to researchers nationwide.”
Jefferson Lab is developing the high-powered computing hardware with Lawrence Berkeley National Laboratory in California. And there’s still work to be done.
Researchers will have to standardize their data to make it easy for AI to identify and pull, said Sean Hearne, president and CEO of the Southeastern Universities Research Association.
“We need to get all this data so it's speaking the same language,” Hearne said.
And even though this data center won’t be like the sprawling data warehouses like those that have drawn criticism as they’ve proliferated across Northern Virginia, there are still the growing concerns related to AI that policymakers and leaders need to address.
“I've gone from excited to scared to hair-on-fire on AI,” Warner said at the groundbreaking. “There are great upsides, but there are also some real challenges.”
He said some of those challenges include national security, personal security, data centers and energy use.
