Shield AI, which wants to reduce American and civilian casualties in ground combat to zero, just raised $10.5 million from Andreessen Horowitz.
Fast Company, April 3, 2017 (approx. 2,600 words)
No one knows the challenges of the battlefield better than the soldiers who face them every day. As part of a Navy SEAL team in Afghanistan, Brandon Tseng got to know all too well the gut-wrenching existential uncertainty involved in kicking down doors in a war zone: Would there be an IED on the other side? A suicide bomber? An enemy combatant with a gun or grenade? An ally who might look the same at first glance? Or just an innocent little girl, caught up in a conflict that to her means only terror, grief, and despair?
A month after coming home, Tseng, 30, teamed up with his older brother Ryan, a serial entrepreneur, in the spring of 2015 to start a company called Shield AI. Their first product, due to be deployed by SEAL teams overseas later this year, will be “an autonomous quadrotor [drone] designed to collect intel inside buildings and structures, which are incredibly dangerous environments to operate in,” Tseng says. The idea is to “get eyes and ears ahead of special operators before they go in,” and thus reduce both military and civilian casualties. Shield’s ambitious goal? To get American and civilian casualties in ground combat to zero by 2030.
The Tsengs raised almost $600,000 in angel investment within a few months of launching their company, but their real break came when they hooked up with an organization known as DIUx—the Pentagon’s Defense Innovation Unit-Experimental, which is charged with matching innovative private-sector technologies to problems that need solving at the Department of Defense. DIUx helped Shield make contacts and generate interest from a number of branches of the armed forces. That interest helped the company raise about $2 million from venture firms in May 2016, and in August Shield was awarded a $1 million contract to provide five prototype drones, reportedly to the Naval Special Warfare Command, which oversees the Navy’s eight SEAL teams (among other units). The company currently has contracts with the Department of Defense and the Department of Homeland Security, and today announced a $10.5 million Series A financing led by top-shelf venture firm Andreessen Horowitz.
Going from napkin sketch to battlefield deployment in two and a half years is the equivalent of traveling at light speed for the notoriously sluggish Department of Defense. By comparison, it took the Army more than five years just to choose a new handgun to replace the Beretta M9, and it will probably wait another year or more before the pistols actually are used in battle.
Coming up with ways to solve the problems of the battlefield has traditionally been the task of Defense Department mandarins armed with long lists of rigid requirements and towering stacks of government acquisition regulations. But with America’s adversaries piecing together cutting-edge, off-the-shelf technology in ways that often challenge current military capabilities, rapid deployment of new technology like Shield’s has become a crucial part of the Pentagon’s mission.
That new focus on innovation has led the Department of Defense to launch or support a raft of new programs recently, designed to update its technology both in the back office and on the battlefield. Some of these, like DIUx, are now attracting the attention of Silicon Valley and other entrepreneurs, drawn by a desire to serve—and to be part of a defense budget of around $600 billion.
But as this new breed of patriot entrepreneurs like the Tsengs gets geared up to do what’s arguably the most important work of their civilian careers–supporting a life-or-death mission that affects the security of the country–the fragile innovation infrastructure the Pentagon has managed to build is in danger of collapsing under the weight of a capricious new administration and an institutionalized bureaucracy that has helped cause the armed forces to come close to falling behind in the first place. Silicon Valley would like to help us win the next war. But will the Pentagon let it?
TANGLED IN RED TAPE
Startups can afford to build stuff first and ask questions later–quite literally. The best young technology companies learn where they’re going and how to get there by getting a product into the marketplace and then paying close attention to how users react. Failure is only another teachable moment on the long road to overnight success.
But for military planners, failure means that people die. To prevent such failures, and ensure the DoD acts as a good steward of the American taxpayer’s money, a deep well of regulations governs how and when the Pentagon can acquire new military gear. Which can present a problem.
Tim Greeff runs the National Security Technology Accelerator (aka NSTXL), a non-profit consortium supported by $100 million in Pentagon contract authorizations. Similar to DIUx, NSTXL connects private-sector companies with military customers that need particular problems solved. DIUx and NSTXL provide ways to get new tech to the warfighter more rapidly than is possible through the Pentagon’s traditional acquisition process. “Especially now, with the rapid pace of technology, the acquisition process will [often] acquire something that’s no longer cutting edge” by the time it’s delivered, Greeff says.
And young companies often face nearly insurmountable hurdles to doing business with the DoD. The Byzantine process of submitting a bid or becoming a government contractor—especially one with the security clearances needed to work with intelligence agencies or special ops forces—can be prohibitive. The tangle of red tape on both ends of the process means that many innovative solutions are never even considered. But organizations like DIUx and NSTXL (as well as a public-private partnership called MD5—confusingly sometimes also known as the National Security Technology Accelerator) are now helping cut through the red tape, and letting founders like Zeuss, Inc.’s Brandon White serve their country in ways that might otherwise be out of reach.
As a kid, White wanted to be Chuck Norris. But as a committed high school athlete, he had already had three ACL surgeries by the time he got to college, and he was told that Special Forces was not in the cards. So White bowed out. “I’m really not built to be anything but Chuck Norris,” White told himself at the time. “I’ll probably drive everybody crazy.”
Instead, he became an entrepreneur. But the desire to serve his country never left him. “Traditionally, the thought is, if I don’t put on a uniform or I don’t take a government job, I’m not serving,” White says. But White found a different way to serve: working with DIUx to provide intelligence- and information-management software to the U.S. Special Operations Command, or SOCOM, which oversees joint special ops missions across the Army, Navy, Air Force, and Marines. Chuck Norris doesn’t need boolean search to find some terrorist ass to kick; Zeuss brings him the right information, when and where he needs it.
BEHIND THE CURVE
In the traditional Pentagon procurement process, startups like Zeuss and Shield would probably never even make it to the prototyping stage. Without a handler like DIUx or NSTXL, “going the DoD or IC [intelligence community] route can be suicidal,” White says, which is why most Silicon Valley boards advise companies to steer well clear. “They can’t circumvent laws,” says White, “but DIUx helped walk us through.”
The companies that have navigated the DoD’s maze of regulations in the past have profited handsomely–and have become exactly what President Eisenhower warned of some 55 years ago: an entrenched military-industrial complex that seems to work harder at defending its own interests than at creating new technologies in the national service. According to White, the Pentagon’s new innovation programs send a clear message: “They’ve effectively told all of those big contractors, you guys haven’t been keeping up. You put us behind the curve.”
The question for many entrepreneurs is whether the DoD can leverage Silicon Valley innovation without dragging its new partners into a similar relationship. “The government’s challenge will be, how do they intermix here without putting the very complex that has inhibited innovation on us?” White says. “The inclination will be for them to enact their system on our culture. But our culture will inherently punch them in the face.”
THE VALLEY OF DEATH
It’s possible the Pentagon needs a good punch in the face–or at least a wake-up call. For much of the second half of the twentieth century, the government outspent the private sector on R&D by as much as two to one. But by the turn of the century the tables had turned, and private companies were spending almost three times as much as government on R&D–putting the cutting edge solidly in commercial hands. The rapidly falling cost of new technologies also means that adversaries, from foreign governments to non-state actors like ISIS, have easy access to that cutting edge, often without the regulatory constraints that can hamstring the U.S. government.
It’s this reality that the Pentagon is now attempting to come to grips with. And while there are attempts afoot to overhaul the acquisitions process from the top down, it’s bottom-up initiatives like DIUx and NSTXL that are already having an effect.
But the real test of a more entrepreneurial defense establishment has yet to come. DIUx provides a way for the armed services to rapidly prototype new technologies, but moving those technologies into bigger procurement contracts may prove more difficult. DIUx has so far moved more than $45 million to the more than two dozen companies it has worked with, which make products that range from air-support drones to wargaming platforms, cybersecurity services, data analysis, and more. It hopes to bring the first of those companies out of the prototype phase and into procurement this spring.
“Beltway insiders call that the Valley of Death, going from prototype to production,” says Raj Shah, a former fighter pilot who is the managing partner at DIUx. The “superpower” DIUx plans to use to get its companies through that valley is a provision in the 2016 defense budget bill that lets DIUx prototyping stand in for a more traditional acquisition process. This means that customers within the DoD don’t have to “recompete” the sourcing for solutions developed by DIUx companies, but can move straight to the kind of big procurement contracts that can make a company’s fortune.
While the prototype-to-production approach might sound like nothing more than common sense, it is in many ways new to the Pentagon, which is used to doing business through long lists of detailed requirements–requirements that may not actually capture a solution to the problem at hand. “We don’t ask for a set of requirements, but rather an understanding of what is their core problem and can we use nontraditional technology or orthogonal approaches to help solve it. That’s the key differentiator to how we work versus a traditional acquisition process,” Shah says.
The approach smacks of Silicon Valley’s lean startup principles, in which a deep understanding of the problem is a prerequisite to coming up with effective solutions. But getting Pentagon acquisitions officers to think in those terms is not easy.
“We as a government suck at writing problems,” says former Army Colonel Pete Newell. “We’re not focused on generating the problems, we focus on the solutions, and that’s how we’re blowing this.”
SHARP END OF THE STICK
These days, Newell can be found in the offices of BMNT Partners, a Palo Alto consultancy with a role similar to DIUx or NSTXL: connecting government customers with innovative private-sector problem-solvers. He brings a singular set of experiences to the job: his last role in the Army was as director of the Army’s Rapid Equipping Force, an office tasked with doing real-time military R&D and problem-solving from forward bases in places like Afghanistan and elsewhere. BMNT’s offices are decorated here and there with miniature models of some of the drones and robot minesweepers that Newell has had a hand in developing in the past. If anyone knows how to identify and describe the problems of the battlefield, it’s Newell.
Along with lean startup guru Steve Blank, Newell is working to update the process of doing military innovation and problem-solving both in government and among entrepreneurs. Together, Blank and Newell have developed a university course known as “Hacking for Defense,” which launched at Stanford in the spring of 2016; the course is now being offered at about two dozen universities around the country and has been adopted as the basis for commercializing innovations coming out of the National Science Foundation.
Based partly on Blank’s lean startup methodology but strongly influenced by Newell’s experience with the Rapid Equipping Force, H4D, as it’s known, gives small teams of engineering and design students the chance to work on actual problems faced by agencies within the Department of Defense, the intelligence community, and (through sister course Hacking For Diplomacy) the State Department. “I want my students to engage in some type of national service,” says Blank, who spent four years in the Air Force during the Vietnam War, including almost two years in southeast Asia. “Right now, the only way to do that is Teach for America or AmeriCorps, or maybe Code for America. But there’s no way to engage with the sharp end of the stick, whether it’s State or Defense, or the intel community.”
“We’re not talking about building battleships this way. But every component bolted inside that battleship for sure could be built this way.”
H4D gives students an up-close look at the business of national security. Course organizers work directly with government agencies to source problems that are actively in need of a solution. In turn, those agencies take advantage of the real solutions that emerge: something like half the 275 people who have come through the courses in the last year have continued to work with government in one form or another, according to Blank. The response on both ends has been overwhelming, he says, and the program is expanding rapidly. A Hacking For Energy course is underway, and planning has begun for both Hacking For Development, which will work with NGOs, and Hacking For Urban Resilience.
Newell and Blank are also seeking to popularize the methodology by training military and government planners in lean techniques through a nonprofit organization known as Hacking For Defense, Inc. DIUx, for its part, also hopes to influence on-the-ground acquisition strategy. To this end, it has published a guide to how other branches of the DoD and government can take advantage of the “commercial solutions offering” it uses to streamline acquisitions. Already, other organizations, such as the Defense Information Systems Agency, for one, are looking at how they can leverage similar mechanisms.
Entrepreneurial problem-solving isn’t the magic wand that will make all the Pentagon’s acquisition challenges disappear, of course. “You should not buy an aircraft carrier with us,” says Shah. But even aircraft carriers can benefit.
“We’re not talking about building battleships this way,” Newell says. “But every component bolted inside that battleship for sure could be built this way.”
“The DoD serves a very important mission, protecting the United States and her citizens. There are a lot of people that are interested in working on that problem,” says Brandon Tseng. “But in the past there were barriers. You’d have to work in a large defense contractor, and people weren’t sure what problems the DoD was dealing with. With DIUx, they’ve opened that opportunity up to many young startups.”
“We serve this mission because of my personal experiences, and the team’s sincere desire to help this deserving and often overlooked community of service members who put themselves in harm’s way, and civilians without the choice who find themselves in these situations. A lot of technology companies ignore these people. I think we can do a better job.”