In 2017, the Defense Innovation Unit began working with several start-up companies to develop tools to automatically find and fix software vulnerabilities in the U.S. military's highly advanced weapon systems.
Building on capabilities developed during a 2016 Defense Advanced Research Projects Agency demonstration, "Project VOLTRON" has shown its tools can detect previously undiscovered vulnerabilities in aircraft software within minutes, according to DIU.
The need for such tools became critical in 2018, when the Government Accountability Office reported DOD weapons testers found "mission-critical cyber vulnerabilities in nearly all weapon systems" under development in the previous five years.
But nearly three years after the project's launch, VOLTRON remains in the prototyping stage.
DIU officials said they are close to transitioning the project to a production contract. They are working with "communities of practice" within various DOD program offices who like what VOLTRON's tools have demonstrated in tests.
But they are still looking for those program offices to put money down on a production contract.
"For these small teams that see the value in it, because it makes their work more efficient and allows them to be more effective, they may have challenges convincing their leadership that this is worth investing in," DIU product manager Army Lt. Col. Ben Quimby, DIU cyber product manager, told Inside Defense during a December tour of DIU's offices in Mountain View, CA.
And program leaders, focused on problems already laid out in assessments like the 2018 GAO report, may not be eager to buy tools that uncover even more vulnerabilities.
"They already have a huge backlog of issues they know about," Quimby said. "So you walk in and say, 'Hey, we might be able to help you find even more stuff that's wrong with your program.'"
In many ways, Project VOLTRON is representative of the challenges facing DIU and DOD's broader efforts to deliver commercial, disruptive technologies to the military.
DIU has initiated dozens of prototyping efforts with small, nontraditional companies over the past five years. But when the opportunity comes to move beyond experimentation to more expensive, long-term contracts, the Pentagon often balks.
A report published by the Reagan Institute Task Force in December states that "beyond initial strides in narrow circumstances," the U.S. government has not shown a willingness to award major contracts to nontraditional contractors.
"DIU remains a small element of the ecosystem, and the DOD lacks experience in integrating commercial products into its programs of record," the report states. "The coin of the realm for integrating dual-use companies into the DOD ecosystem is a large program of record."
Meanwhile, two years after the 2018 National Defense Strategy declared "new commercial technology will change society and, ultimately, the character of war," investors are still waiting for DOD and its $700 billion annual budget to make a bigger splash in the commercial technology world.
"I've been on the venture side of defense investing for maybe four years now -- the pace and quantity and quality of conversations are improving," James Cross, vice president, research analyst and portfolio manager with Franklin Equity Group, said during an interview with Inside Defense at his office in nearby San Mateo, CA.
"But I think everybody's recognized it's time to find something more than connective, conversational tissue," he added. "We need to build business relationships."
Beyond an experiment
As many analysts are calling for the adoption of commercial technology to compete with the Chinese military, DIU touts itself as the only DOD organization focused solely on commercial outreach.
"There's not a single other entity within the DOD, arguably within all of government, that is able to have that level of reach into the commercial space and be able to culturally deliver an engagement product to [companies] that is easy for them to navigate and understand and therefore engage with it," Tom Foldesi, DIU's director of commercial engagement, said in a December interview at DIU headquarters.
Established nearly five years ago by then-Defense Secretary Ash Carter as the Pentagon's new beachhead in Silicon Valley, DIU has weathered early criticism and steadily grown as an organization. In 2018, the "experiment" tacked on to the end of its name was removed, signifying its enduring place within the department.
DIU works across five technology portfolios: artificial intelligence, autonomy, cyber, space and human systems. In addition to its Silicon Valley headquarters, DIU has offices in Boston, Austin, TX; and the Pentagon.
The organization’s aims are twofold. It wants to use commercial technology to solve military problems, and by doing so, show commercial companies -- and the venture capital firms that finance start-ups -- that there’s a viable path to success by working with DOD.
"In many cases, VCs have actually counseled companies to stay away from DOD, because the juice is just not worth the squeeze," Foldesi said. "We are here to change that paradigm. We want to be able to make the DOD as company friendly a customer as anyone else in the private sector."
DIU recently reorganized into two new engagement teams aimed at bringing disparate worlds together: the commercial sector and the DOD acquisition community. While Foldesi heads up the commercial engagement team, which meets with start-up companies and venture capital firms, a defense engagement team works with DOD program offices on their requirements.
"We're trying to take the noise that DOD procurement is for companies and turn that into an actual signal," Foldesi said. "So when you do business with DIU, you know that it is actually going to go somewhere, as opposed to just being potentially some other short-term program both in terms of time and resources and the amount of money."
Despite delays in transitioning programs like Project Voltron, DIU has moved several prototypes to large production contracts. DIU recently facilitated the award of a five-year, $95 million contract to C3.ai to provide "predictive maintenance" across DOD's aircraft fleets after the company successfully prototyped its machine-learning software with the Air Force.
DIU's goal is to move as fast as the commercial world operates. The aim is to negotiate an initial "other transaction agreement" with a given company in 60 to 90 days, and then prototype the technology within 12 to 24 months to determine whether it meets DOD’s needs.
"DIU is to be commended," C3.ai President and Chief Technology Officer Ed Abbo told Inside Defense in January. He said the initial "contact to contract" took just 90 days.
"They really did a lot of ground work to help us identify the right customers to engage and then obviously as we demonstrated the efficacy of the solution across different aircraft platforms with different characteristics, they were able to help us make the case to scale and make the solution across the DOD," Abbo said.
Between June 2016 and December 2019, DIU facilitated 154 prototype projects worth a total of $565 million in contracts, according to information provided by DIU. The projects have involved 142 nontraditional vendors, including 60 first-time DOD vendors.
And U.S. investors are taking note of the Pentagon's activity in the commercial start-up world. Last year, an informal "defense investor network" began meeting regularly to track emerging commercial technologies with potential military applications.
Foldesi said the network's formation is evidence of the "increasing momentum" around efforts to more closely link commercial companies, venture capital and DOD.
"I do think we're about to be on the path to a virtuous circle in this space around defense tech," he said.
'Transformative' goals, fiscal realities
DIU's latest objectives reveal the scope of its ambitions. In last summer's report, DIU outlined a new goal of facilitating at least five "transformative" projects every year.
"While DIU considers a wide variety of criteria when selecting projects, some of the most important indicators of transformative potential include technologies or methodologies that could scale across all of DOD; dramatically reduce loss of life and improve force protection; introduce new efficiencies and save taxpayer dollars in significant numbers; or inspire new operational concepts that enhance military capabilities," the report states.
The predictive maintenance project run by C3.ai is one example of a transformative project, as is Project Voltron's potential to improve DOD’s cybersecurity, according to DIU's report.
But as DIU seeks a larger impact and scale, it is running up against some limitations.
Mike Madsen, DIU's deputy director, said the annual DOD appropriations process poses a challenge to the organization's efforts to transition technologies beyond prototypes. The military services begin building their budget proposals more than a year before the Pentagon's request is sent to Capitol Hill. It then takes at least several more months for Congress to pass annual defense appropriations bills.
"What we found is when we execute prototypes, and especially if we have a successful prototype, well ahead of schedule, and the DOD partner was unable to plan for that in their budget, they have limited options," Madsen said on the sidelines of the Reagan National Defense Forum in December.
If they wait too long to move beyond the prototyping phase, Madsen explained, the technology can become outdated, or the company can even go out of business. DOD can ask Congress to reprogram money to fund the transition to production, but that process can take months, and requests are often rejected.
Madsen said DIU is currently studying several proposals to reform how defense innovation efforts are funded.
The effort is intended "not to revamp the entire appropriations process, but to take a look to see if there are any responsible, creative ways to allow for bridging that gap if things were done sooner," Madsen said.
Meanwhile, DIU's budget poses its own challenge. According to DIU's annual report, 94% of the organization's projects between 2016 and 2018 were funded by DOD customers.
This is by design -- DIU wants to connect DOD program offices with innovative commercial companies -- but also by necessity. DIU received a total of $46 million for R&D funding in FY-20, a modest amount in comparison to other defense agencies.
During a February hearing held by the House Armed Services Committee's "Future of Defense" task force, Raj Shah, the former director of DIU, said the organization's innovation efforts are beholden to outside funding and decisions made by senior DOD leaders.
"There has been no shortage of senior level folks that have come to places like the valley and Boston that say, 'We want this innovation, we love it,'" Shah said. "But then when it comes time to fund relatively modest contract sizes, they say, 'Look we love AI, but we can't find $20 million for this project.'"
Shah advocated for drastically increasing the budgets of DOD's successful innovation organizations.
"My recommendation is that we've already experimented in several ways -- the ones that are demonstrating success, 10x their budget," he said. "10x their ability to bring these companies into the fold, and then you will get multiple orders of magnitude of private capital and support behind that naturally."
However, the Trump administration's fiscal year 2021 budget request released last month would reduce DIU’s budget to just $39 million across two R&D accounts as a result of Defense Secretary Mark Esper's "defense-wide review."
The reduction to DIU's budget would "realign funds for higher priority DOD missions," according to budget documents.
'Break through at scale'
Despite DIU's early successes in working with start-up companies, some reports indicate DOD's efforts to bring new firms into the defense contracting market aren't yet bearing fruit.
For instance, in February, the National Defense Industrial Association and Govini published a report on the defense industrial base that found new entrants into the defense contracting market declined by 14% between 2016 and 2018, despite DOD's increased use of novel contracting schemes intended to make it easier for nontraditional companies to do business with the government.
"This trend suggests that, despite the recent increase in DOD innovation funding and the growing use of faster, more flexible alternative acquisition pathways, the defense contracting market continues to lack attractiveness to non-traditional vendors," the report states.
Many critics of the Pentagon's technology adoption efforts point to the immovable prime contractors at the top of the federal market. But two companies, Palantir Technologies and SpaceX, are often held up as the potential vanguard in a new wave of innovative primes.
However, even Peter Thiel, the founder of Palantir, recognizes his company's success in the defense contracting world may be difficult to repeat.
"I personally funded Palantir, Elon [Musk] personally funded SpaceX," Thiel said in December during the Reagan National Defense Forum. "We got outside capital maybe six years into both businesses. And so if you have something where you have a six-year gap, and until then, it gets funded by someone who's putting in tens of millions of dollars as a sort of slightly quixotic project, that's a tricky template to repeat."
Thiel is also a co-founder of Founders Fund, a San Francisco-based venture capital firm. He said venture funds are looking to invest in companies they think can soon be worth more than $1 billion.
"You need to get some that can really break through at scale," Thiel said. "We're not investing in hotdog stands, which could be a great business, but it's not going to scale. We have to invest in companies that start small but have the potential to grow and scale to a world-class scale."
Cross, the Franklin-Templeton investor, said global technology firms like Amazon, Microsoft and even Google might be closer to the "new primes." Those companies have moved aggressively into the defense contracting world, spurred by DOD’s increasing adoption of cloud computing services.
Meanwhile, DIU remains the only DOD organization focused on "late stage" investments, Cross said, when start-up companies have proven out their products and are ready to scale.
Cross lauded DIU's efforts but echoed the findings of the Reagan Institute Task Force report's finding that the "coin of the realm" is DOD’s major acquisition category (ACAT) programs.
"I think we'll have success when every program executive office has somebody in the valley . . . every ACAT 1 has somebody in the valley, doing tech scouting on behalf of that program, instead of going through a second- or third-party organization," Cross said.