This article is published in Aerospace Daily & Defense Report part of Aviation Week Intelligence Network (AWIN), and is complimentary through Mar 20, 2025. For information on becoming an AWIN Member to access more content like this, click here.
National Geospatial-Intelligence Agency Director Vice Adm. Frank Whitworth spoke at the DoDIIS Worldwide Conference, Dec. 13, 2022, at the Henry B. Gonzales Convention Center in Texas.
The National Geospatial-Intelligence Agency (NGA) is processing imagery at rates 80% faster compared to one year ago, its director said March 10.
The agency is increasingly adopting artificial intelligence (AI) capabilities that are supporting such dramatic improvements in data collection rates, but more processing power is still needed, Vice Adm. Frank Whitworth said in a keynote discussion at the SATELLITE Conference and Exhibit in Washington.
“When you’re stuck with the same amount of compute, obviously, you might be facing a situation where you have a bit of an elbow, and it could go back up,” he said. Whitworth warned that the number of terabytes coming from space over the next eight years will grow “precipitously” as a swath of new satellite constellations come online.
As the NGA is looking for targets, adversaries are working hard to conceal those targets, he tells Aviation Week.
“That amount of data to then run through our model—that takes time. Running that inference with processing, compute power, takes time,” he said in an interview. “To have that drop 80% in one year—I like. That’s good news. I don’t want it to go back up.”
Whitworth attributed the drop in latency to efforts related to NGA Maven, the Pentagon’s flagship AI project that was launched in 2017 to integrate AI into military workflows. NGA took over the geoint aspects of Maven in January 2023.
He also referenced the NGA’s relationship with commercial technology companies, as well as “the ingenuity” of its staff and personnel at component commands as contributors to reducing latency, he added.
“But even things like compression and the way that some of the data is parsed, it’s all just getting better,” he said.
NGA has declared 2025 to be the year of AI adoption, he said. The agency has created three new roles meant to sharpen the focus on AI adoption throughout the organization: the chief AI officer, director of AI missions, and director of AI programs.
The agency is investing in accreditation and certification efforts like responsible AI training, and assessing the quality of AI training data and models, he said. He is also encouraging agency analysts to mentor the AI models in a “team approach.”
The goal is to have models that can identify a potential target while the analysts are asleep. “That’s when you’ve arrived,” he said.
NGA is using AI models that were stood up just in the past year to keep custody of certain elements “that we didn’t used to keep custody of,” Whitworth said. “And that’s a very powerful thing.”
NGA also needs to make sure that U.S. adversaries cannot manipulate the geoint data being collected, he said.
Related Content
