Pentagon’s Top Science Official Adds to Tech-Breakthrough Wishlist
Heidi Shyu, R&D undersecretary, said she went looking for tech areas to trim—and found that some vital ones had been overlooked.
When Heidi Shyu was nominated to be defense undersecretary for research and engineering, the longtime Army acquisition executive set out to trim the long list of technology areas set as development priorities.
But she found that each of those areas—hypersonics, artificial intelligence, cybersecurity, and at least a half-dozen more—is vital to the Pentagon’s future efforts to deter China. Moreover, the list leaves some key areas off.
“Initially, I thought I could reduce the number, but I ended up adding to it,” the recently confirmed Shyu said in an interview.
That may mean more headaches for the defense industry, which has struggled to align its own research spending with the priorities list. The list has shifted over time, and it remains easier for companies to pursue modest research goals around current products than to reach for big science-and-technology breakthroughs.
But Shyu has ambitious plans for change. At her Senate confirmation hearing, she noted that only about 30 percent of Pentagon spending on a given weapons program goes to develop and procure it, while 70 percent goes to sustain it. She wants to reverse that ratio.
Here are some highlights from the interview.
Shyu said she agrees with critics who object to the high cost of the Pentagon’s multiple hypersonics efforts.
“I'm very focused on trying to figure out how we can develop affordable hypersonic weapons. And so part of that ties into: do we have the right materials? Do we have the right test facilities to enable us to do this?” she said.
The Defense Department is asking Congress for $2.865 billion in 2022 to fund hypersonics development across multiple services. In the most recent request, the Army and Navy doubled their hypersonics funding while the Air Force sought a 40 percent reduction. The Air Force is achieving success in its air-breathing, jet-launched hypersonics platform and the Army has actually begun deploying hypersonics missiles to troops.
At some point, once those research and engineering questions are answered, manufacturing will pick up and that will lower down the price per unit, she said.
Shyu said the Pentagon might prune the number of hypersonic projects currently in development to focus on the winners.
“Ultimately, you can play in the [science and technology] world, but in the end the service has to [put] it into their budget, right? So depending upon how successful the [science and technology] projects are, that dictates what will end up transitioning.”
Silicon Valley’s massive spending on artificial intelligence means the Pentagon can focus its own R&D on better understanding what AI tools are actually doing, how they operate, and how to deploy them safely.
“Industry is spending billions of dollars on AI ML [machine and learning] and billions on autonomy. What I want to do is make sure we converge on trusted AI ML, trusted autonomy… In my mind, we have to drive towards affordable, attributable, survivable unmanned platforms. To do that, it's gonna be critical to have trusted AI ML and trusted autonomy.”
That puts a greater emphasis on testing, safety, and careful design of employment than just rushing new capabilities out. That, in turn, marks a departure from the way most tech firms experiment and then deploy artificial intelligence and machine learning tools.
Shyu’s vision for future cybersecurity rests heavily on networked and rapid detection of threats rather than trust in firewalls.
“As I look towards where we need to hit in the department, in my mind, the sensors can no longer remain stovepiped,” said Shyu. “We now need to have the ability to literally work in the intersection of cyber and electronic warfare, information operations, and communication. We have to be able to sense and react very very quickly.”
Open architectures are becoming ever more critical, she said.
“What you don't want to do is lock into one architecture and so, as the threats evolve, you're stuck with a legacy architecture. So, if we have a modular open architecture...with a secure processor that's able to evolve at the speed of commercial technology, that's what's going to be critical.”
Shyu put in an endorsement for the CHIPS Act, which would establish incentives to bring more microprocessing manufacturing back to U.S. shores. But she said the Defense Department must learn to better use the chips they have.Some “real time operations,” she said, “literally buy multi-core processors and then they only use one core because that's how they know how to program the node. So I think we need to put some money into research. If you have a 16-core processor or 32-core, you should figure out how to utilize all the cores.”
What do you do with those cores? One use would be more on-board processing in low-bandwidth environments where high-tech adversaries are contesting communications. But Shyu also emphasized the need for better 3D visuals—for training, but also for command and control.
“I would like to focus on and push us towards the ability to develop interactive 3D operations via 3-D operations center, enabling geographically dispersed, distributed command and control in a low bandwidth environment,” she said. “So we can enable rapid mission planning and mission command. That will be incredibly powerful. The technology is here. It's not something we have to invent. It just needs a little push to mature and get into the hands of our users.”
Shyu said the Biden administration has not ruled out the pursuit of space-based weapons, which her predecessor envisioned as useful in downing enemy missiles and warding off Chinese or Russian anti-satellite systems.
“Everything is classified,” she said.
However, the administration’s space strategy will still focus on resilience and fielding more satellites at less cost.
Shyu said biotechnology, especially advanced materials, would be key to future operations. She noted an August DARPA demonstration.
“Within 48 hours, they were able to use basically water and sand and biology to create a helicopter landing pad. So that's awesome. That greatly reduces the logistics cost if you could create that type of capability, at the pointy edge.”
But Shyu said that the Defense Department must broaden its biotechnology focus to include things like biometric data on soldier physical activity and health, which could alert leaders to big changes in the wellbeing of the future force or even the public.
“A lot of the technology developing is trying to anticipate what is happening in the future, and experiment with a type of biotech that can help to...I call it sensing and avoiding surprises. If you have the ability to to look at—to be able to sense—either from your environment, or...your behavior, via your Fitbit or whatever, that you may be coming down with something because your heart rate is increasing or or indications of warning of potential infections…I think we want to continue to fund those.”
Though it may not become a formal modernization area, Shyu brought up several times how new materials would be key to breakthroughs in several of the above areas.
The key to moving all of these ideas from concept to reality will be getting technology into the hands of warfighters faster and making them a much bigger part of the research process, she said.
Deputy Defense Secretary Kathleen Hicks “has stood up an innovation steering group for me to chair. One of the things that we're doing is creating a rapid development and experimentation reserve pot of money. So we could literally take very promising ideas, experiment with them and see if [the experimentation] fulfills a capability gap that's defined by the Joint warfighting concepts,” she said. “We want to conduct joint experimentation so it isn't just one service testing it out and meeting their needs…We’re working very closely with all the [Combatant Commands] and all the Joint Staff.”