Software and startups top priorities for DOD tech chief pick

Heidi Shyu in 2013 as assistant secretary of the Army for acquisition, logistics and technology in Kandahar with Gen. Dennis Via. (DOD photo by Spc. Isaac Adams)

Software and elevating small businesses in the defense industry base top the list of priorities of the Biden administration's pick to lead the Defense Department's technology efforts.

Heidi Shyu, who previously served as the Army's assistant secretary for acquisition, logistics and technology, debuted her priorities to modernize the military with a strong emphasis on software during her Senate Armed Services Committee confirmation hearing on May 25.

"In order to rapidly transition the latest software, we need to have an open architecture that isolates the software from the hardware then allows rapid user testing," she wrote in policy question responses.

Shyu told senators that DOD needs to "flip" its investment ratio to make development and procurement 70% of the costs for a new weapons system by buying more emerging tech like artificial intelligence, synthetic biology, and hypersonics, rather than sustaining legacy systems.

"Today, sustainment makes up 70% of total weapon system cost, with development and procurement making up 30%," Shyu said in opening remarks.

Shyu said if confirmed as undersecretary of defense for research and engineering, she would focus on such as creating cyber-hardened "networked systems-of-systems" to improve data collection and sharing, developing advanced materials to increase fuel efficiency and investing in secure, upgradable software.

Small businesses, particularly startups developing new technologies, were mentioned repeatedly throughout the hearing as being integral to the Defense Department's success.

However, challenges with the federal acquisition system and failure to translate prototypes into programs of record are often too much to overcome -- and Shyu wants to change that by ensuring there's a clear "transition path".

"Part of the reason there is a valley of death for technology is because a lot of the technology programs are being developed by small companies," Shyu said, "and unless you had the foresight two years ago to understand that the technology is going to be mature within two years time...by the time you get the money to buy that technology it's two years old now."

The technology tested in labs should also be tested in "relevant environments" and increase user feedback so that program managers are more comfortable betting on prototypes, she said.

"I saw a six-person company that's developed any type of fuel as input and the output is a DC-plug. Those are the types of creative, innovative companies we need to nurture. And they're struggling to figure out who to talk to in the DOD," Shyu said.

"This is a giant fortress and they have no idea where the door is."