A New Front
Can the Pentagon do business with Silicon Valley?
“Want to see the bin Laden compound?”
Tony Sanchez, director of special projects for a San Francisco tech company called Osterhout Design Group, holds up what looks like a place mat covered in zebra stripes. He’s standing in a Stanford University conference room, addressing a collection of people not usually seen in Silicon Valley: Some are undergraduates and technologists from area startups, but many are active-duty soldiers and veterans enrolled in the business school.
Sanchez hands out a set of smart glasses. When the people around the table put them on and look at the sheet, Osama bin Laden’s Abbottabad hide-out comes to life. It looks like something out of a 3-D video game. Turn the sheet, and the tan low-rise buildings rotate. Pull the sheet in, and the buildings get closer. Push it out, and they shrink away.
One of the soldiers takes a look. “Can you imagine?” he asks. “Our snipers would love that.”
About 30 people are participating in the workshop, part of a Pentagon project called Hacking for Defense that’s searching for new ways for the military to identify promising Silicon Valley technologies. During the Cold War, defense money funded much of Northern California’s nascent tech industry, and the military worked closely with universities and companies to develop electronics, microwave devices, semiconductors, and spy satellites. But the military did not stay connected to the venture capital–fueled tech industry that emerged in Silicon Valley in the 1980s. Until recently, the Pentagon didn’t see this as a liability: The United States enjoyed unmatched technological superiority on the battlefield. That advantage, though, is now dissipating. China and Russia have invested heavily in new systems. ISIS is using hobby-style drones for reconnaissance. Rebels in Syria are using iPads to aim mortars. Equipment like this was once prohibitively expensive. Now you can get a lot of what you need off the shelf.
Running the workshop is a compact man who stands ramrod straight and speaks in a gravelly voice. Not so long ago, Peter Newell was a colonel in the Army, commanding thousands of troops in Iraq. In 2013, after retiring from the service, he moved to Palo Alto to start a consulting firm, BMNT Partners, whose primary mission is to get the military and the tech community working together. A research initiative at the Pentagon called Technology Domain Awareness asked Newell to create the Hacking for Defense project.
Sanchez pulls out a small cardboard box. He shows how sensors in the glasses could read the box’s contents — the same way X-rays can peer into your body — while an app converts that data into a 3-D display on the lenses. The computing is done in a processor about an inch wide embedded along the top of the lenses that has as much power as a tablet computer. Bluetooth and Wi-Fi connectivity allow the devices to communicate with other glasses nearby.
The group discusses how soldiers could use the device in combat. A sensor on a rifle could track the reloading of magazines and feed that information to an app in the glasses, letting a soldier know how much ammunition he’s used. An accelerometer could measure the force of an IED blast to assess whether a soldier’s brain has been injured. Newell points out that if an entire squad or platoon was wearing the glasses, the team leader could use an app to figure out how to arrange his people. Another app could help a forward observer call in artillery. All of this, Newell says, would “remove that cognitive load from the guy doing calculations in his head.”
As much as Defense Department officials say they want better access to commercial technology, the way the Pentagon functions often makes this impossible. The military has spent decades configuring itself to work with defense contractors to build complicated systems that take years to produce, like fighter jets and aircraft carriers. With its cumbersome rules and processes, the Department of Defense is not set up to race alongside small, agile companies.
The Pentagon is beginning to realize it must operate differently. Some of the most advanced work in computing, big data, cybersecurity, energy, robotics, and space — all areas the military draws on — is being done by tech companies, not traditional defense contractors. Last year, the Pentagon kicked off a large-scale effort called the Defense Innovation Initiative. In April, Secretary of Defense Ashton Carter traveled to Palo Alto to announce that the department was establishing an office in Silicon Valley.
On the surface, Peter Newell, who is 53, does not seem like an obvious candidate to grasp Silicon Valley. An Army brat who enlisted in the National Guard before commissioning as an officer, Newell spent 32 years in the service. Early tours took him to Panama and Kosovo. He was on duty in the Pentagon’s National Military Command Center on September 11. During the battle for Fallujah in Iraq, he commanded a battalion, earning a Silver Star. In 2009, he was responsible for a swath of southern Iraq, tasked with intercepting fighters, weapons, and cash being smuggled in from Iran through marshes along the border.
Soon after that tour, the Army put Newell in charge of a unit called the Rapid Equipping Force. The unit had been created in the early days of the Afghanistan War. After senior officers saw a photograph of a soldier searching a cave for enemy fighters, some wondered why the military was putting lives at risk when it could send in robots. The technology existed, but the problem was bureaucratic. The military didn’t have a system in place to respond to unexpected needs on the fly; it can easily take more than a decade to research and develop a piece of equipment. The Rapid Equipping Force was created to get new gear to the field quickly.
Newell had never heard of it. The organization was only seven years old and not widely known. But once he learned what it was, he got angry. Rapid Equipping Force teams had been in Iraq when Newell was there. The reason he was angry was that they had never approached him. “I’m working in a swamp,” he says. “I needed motorcycles, ATVs, and canoes.” Instead he was given a Navy unit with boats only suitable for rivers. If he had known about the Rapid Equipping Force, he might have gotten the equipment he needed. Lives almost certainly would have been saved. “To me, that’s not forgivable,” Newell says.
When I visit Newell at BMNT’s offices in Palo Alto, he tells me that soon after taking over the Rapid Equipping Force, he headed to Bagram and Kandahar. Iraq was winding down, but Afghanistan was heating up again. Newell met with commanders to find out what they needed. He gets up from the conference table and walks over to a bookcase. BMNT is located on California Avenue, one of Palo Alto’s many tech corridors, and its layout could pass for that of any of the area’s startups. It has an open floor plan and lots of white boards. But there are some notable differences. Most of BMNT’s staffers are veterans, and the walls are covered with pictures of them wearing fatigues and standing in Afghan fields or sitting in the back of Bradley Fighting Vehicles. In the entryway, an American flag sits in a modest picture frame — “a very public reminder,” Newell says, of the staffers’ desire to “continue to work for the good of our country.”
Newell grabs a picture from the bookcase and sets it in front of me. A soldier with a remote control is standing next to a Bobcat excavator, the kind you see on a construction site. The shovel has been replaced with a metal arm bearing five thick wheels. The Taliban, Newell explains, targeted soldiers on foot. IEDs were ripping them apart. Newell returned to the States and started hunting for some way to protect them. He discovered that a company in Britain had developed a remote-control system for Bobcats (“They used it in Fukushima”) and that an Army lab was developing small mine rollers that triggered and absorbed explosions. The Rapid Equipping Force combined the devices and created a vehicle it called the Minotaur, which could now clear a path for the soldiers. Nine months later, the machines started arriving in Afghanistan.
Newell headed the Rapid Equipping Force for three years. At one point, he came out to Stanford to consult with a research fellow about robotics. Insurgents were stashing supplies in Afghanistan’s extensive aqueduct systems. The military wanted to inspect them remotely. Newell’s host was Joe Felter, a former Special Forces officer who worked at the university’s Center for International Security and Cooperation. Felter had done a tour in Afghanistan focusing on counterinsurgency, so when he heard about Newell’s aqueduct problem, he understood it immediately. Felter suggested they head off campus and take a “garage crawl” through startups in the area. “What I found running around to these places,” Newell says, “was that there were exceptional capabilities out here and great ideas that I had no access to.”
When Newell first joined the Rapid Equipping Force, one of the things he sought to understand was why the issue of foot soldiers being blown up had never registered with the unit, even though reports had been coming in for months. “I asked them, ‘How did you miss this?’” Newell says. The organization saw its job as scouting for technologies. Newell thought that, instead, it should look for problems and then figure out how to crack them. “That set me on a two-year tantrum,” he says, to change the mentality of the organization.
Although a colonel, Newell reported straight to a four-star general. This gave him the freedom to try new approaches. He sent staff to business courses at MIT and then hired one of the instructors, Steven Spear, to find ways to accelerate the unit’s operations.
Toward the end of Newell’s term at the Rapid Equipping Force, he faced a decision: stay in the Army or retire? One night at San Francisco International Airport, as Newell prepared to return east, Felter suggested he move to California and continue what he had done with the Rapid Equipping Force, only now in the private sector. Going off on his own would be a risk — defense contractors are usually the employer of choice for retiring colonels — but Newell could see how it might work. “I had to convince myself I could do it,” he says.
BMNT Partners was formally inaugurated in early 2014. Felter and a former Marine and recent Stanford graduate named William Treseder joined as partners. (“BMNT” is military shorthand for “before morning nautical twilight,” the period before the sun rises but after light has pierced the darkness. For an army, this has historically been the best time to attack and, conversely, the most dangerous time to be on defense.) As with many startups, they began with little more than an idea and figured out how it would work as they went along. They’ve helped U.S. forces in Korea research how to blast through concrete doors in tunnels. They’ve taught innovation practices to the Marine Corps and the Pacific Fleet. They’ve facilitated design sessions for Special Operations Command on how to power an “Iron Man” suit. And they’ve aided early-stage startups in finding government funding. One thing that hasn’t changed is their core mission. “We’re trying to get better technology to our military faster,” Felter says.
Last year, the Pentagon put out a call for papers on which emerging technologies might have the greatest utility in the next couple of decades. This is not unusual. The Pentagon frequently asks private industry for advice. Defense contractors, hoping to snag lucrative contracts, are happy to put in the hours drafting white papers. This time, though, the Pentagon specifically said it wanted to hear from companies outside its “traditional orbit.” It’s not clear what kind of response the department received from the tech world; the Pentagon says it didn’t track submissions by company type. The deadline for submissions was extended twice, however, and a delegation flew to California to see how it could engage Silicon Valley.
Three years ago, while still at the Rapid Equipping Force, Newell met with an engineer at Google who was researching energy solutions for those off the grid. Outposts in Afghanistan ran on generators. Tons of fuel had to be trucked in through hostile territory, making resupply a life-and-death operation. Newell asked the engineer how much it would cost to have Google research the Army’s project. The engineer answered by drawing a giant circle on a white board. “That’s my budget,” he told Newell. Then he put a much smaller circle inside it. “That’s your budget,” he said. “I’m not interested in your money. I’m interested in your problem.”
“That’s when I learned about the fallacy of the Pentagon’s model when it comes to Silicon Valley,” Newell says. Defense companies hire themselves out. Tech companies don’t. They look for a problem that a lot of people have, develop a product that solves the problem, and then take it to market. “What the Google guy was telling me was: Your role in this ecosystem is to deliver problems,” Newell says. Earlier this year, when the forces in Korea asked BMNT to help with the tunnels, Newell asked that a lead engineer be sent to the States. When he arrived, Newell brought him to various startups and let the conversations flow.
The Defense Department’s equipping systems were designed for an era when it was a lot easier to predict needs 20 to 30 years in advance. (The Pentagon’s research arm, Defense Advanced Research Projects Agency, for example, focuses on long-term and conceptual challenges.) But today it’s impossible to know all the threats the U.S. will face in three decades or which technologies will have the greatest impact. After all, 15 years ago, who would have predicted that social media would give terrorists the ability to communicate and issue orders?
In July, the Brookings Institution came out with a report called There’s No App for That, which details the difficulties that tech companies face when trying to sell to the military — barriers it called “formidable, if not prohibitive.” Between laws on what the Pentagon can and cannot do with taxpayer money and in-house rules aimed at preventing fraud and error, the military’s acquisition system is opaque and confusing. “The DOD,” according to the report, “is simply perceived as a bad customer.”
Newell often emphasizes that BMNT’s mission is about more than just getting the latest gear to the military. He wants the Pentagon to internalize Silicon Valley’s approach to developing new products. “The way wars are fought now,” he says, “it does not matter how good the tech is that you show up to the fight with. It matters how fast you can change. We have to be in a position to recognize an emerging problem much faster and to deliver that problem into the hands of people who can actually combine the right technologies to solve it.”
For a long time, the Pentagon monopolized most of the advanced technological development, either by producing it in its own labs and in government-funded academic research centers or by contracting out to defense companies. Developing tech was expensive, and the Pentagon had the biggest purse. But as costs plummeted, the private sector exploded. Now it’s places like Silicon Valley — and its counterparts in China, Europe, and South Asia — that set the pace. New tools and devices are emerging so rapidly, the Brookings report said, that sometimes, by the time the military gets the equipment it ordered through traditional channels, it’s already out of date. “The cycle must be broken,” the report concluded, “lest we end up offsetting tomorrow’s threat with yesterday’s technology.”
That the Pentagon is opening a new Silicon Valley office is an acknowledgment that commercial tech is now ascendant. The new outpost is called the Defense Innovation Unit Experimental, a clunky government name meant to signify that the office is a work in progress. The tech world’s response has mostly been guarded, to the extent that the office has registered at all.
“In Silicon Valley, so many things are based on relationships and understanding the ecosystem,” one observer told me. “When you’re in the military, you’re by definition transient. Folks rotate through. To have long-term success, they’re going to need some people with intimate knowledge of the terrain, who have those relationships — or it could be a flash in the pan, a good idea that didn’t quite have the legs to make it.”
It’s hard to find a more natural candidate to work with the military than Ralph Osterhout, the 69-year-old founder of Osterhout Design Group whom the website Gizmodo once referred to as a “real-life Q” — James Bond’s gadget wizard. He developed high-tech scuba gear in his 20s and later pioneered microelectronics in toys. In the 1980s, he co-created the night-vision goggles used in the Gulf War. About six years ago, he began developing smart glasses designed for the military — the same ones that were demonstrated at the Hacking for Defense workshop. Processors were getting smaller, and tablets were taking off. Osterhout realized that for a soldier in combat, the best place for an app is in front of his face.
The firm’s offices are tucked into San Francisco’s South of Market district, the heart of the city’s tech community. At his desk, Osterhout wants to show me all the things the glasses could do and starts whipping through one slide after another. Soldiers are in a city, facing off against enemy fighters in a far building. Orange and blue lines trace the paths of the bullets. “Remember The Matrix?” he asks. “You saw the bullet paths in slow motion. Imagine you actually can. You see the contrail of the bullet moving air molecules out of its path.” Another slide shows a team of soldiers moving toward bombed-out buildings. A small hand-launched helicopter — think a Pringles can with a rotor — hovers above them and tracks the enemy, transmitting information to and from the soldiers’ glasses, which, among other things, allow commanders to follow the battle through the troops’ eyes.
The commander of the Pacific Fleet praised the devices at an industry conference last year. Various defense labs and units were testing the glasses until 2011, when Congress instituted across-the-board budget cuts. “Ever since, nothing has happened,” Osterhout tells me. The Brookings report talked about how the Pentagon’s unreliability deters commercial tech companies from working with the department. “There was always one more meeting, and the meetings just never went anywhere,” it quoted one tech executive as saying. Osterhout is moving ahead with his commercial clientele. Hospitals are exploring how the device could cut the need for monitors in operating rooms. Builders want to use the glasses to conduct augmented walk-throughs on construction sites. “I can’t afford to chase the DOD as a principal customer,” Osterhout says. “There is simply no incentive in the DOD to go fast.”
On August 28, Secretary of Defense Carter returned to Silicon Valley. At an event hosted on the grounds of the NASA Ames Research Center at Moffett Field, Carter told several hundred military, political, and tech leaders that the Defense Innovation Unit Experimental — or “my new startup,” as he called it — had opened for business. It was occupying space in an Army building across the street. When he’d arrived in April, Carter had been the first sitting secretary of defense to visit Silicon Valley in almost 20 years. Now he was back for the second time in just four months. One of his core goals, he said, is “renewing the ties, the bonds of trust between our national security endeavor at the Pentagon and our wonderful, innovative, open technology community.”
Taking four months to get a new project off the ground is swift for the federal government, but it’s slow in the tech world. The office’s funding is modest: $1.75 million for the first year and $5 million for subsequent ones. That’s a windfall for an early-stage startup but minuscule for the Defense Department, whose budget exceeds half a trillion dollars. The former head of a premier Air Force research lab is in charge of the project; a Navy SEAL with an MBA from Harvard is his deputy. The office will have 12 employees — “folks with entrepreneurial backgrounds and prior startup experience,” according to a senior Pentagon official.
The official says the office will act as a resource for entrepreneurs “to help point them to where they could make a difference in helping us solve national security problems.” This framing, though, makes one wonder how much the Pentagon still has to learn about Silicon Valley. Unlike defense contractors, tech entrepreneurs aren’t interested in “helping” the Pentagon do anything. They want to build products. If the military wants to buy them, all the better. To be fair, it’s probably too much to expect the Pentagon to transform its mind-set overnight. “If you look at how the DOD historically changes, it’s through a series of ‘Oh shit’ moments,” says Adam Jay Harrison, a research fellow at the National Defense University. “Pearl Harbor, Korea, the hostages in Iran.”
Carter likely knows this. A physicist who has spent his career going back and forth between academia and the Defense Department, he understands the Pentagon’s constraints. He also has long-standing ties to Stanford and understands Silicon Valley. Carter was in charge of procurement when vehicles in Iraq and Afghanistan were being blown up by IEDs. In early 2014, he wrote an article in Foreign Affairs about how bureaucratic inertia prevented the Pentagon from addressing the issue immediately. “Too many lives were lost,” Carter wrote, “because the Pentagon failed to keep up with a changing battlefield. Never again should it make the same mistake.”
Newell has been following the Defense Innovation Unit Experimental since its inception. He watched a livestream of the April announcement at his office. At Moffett Field, he had a seat in the front row. Pentagon leaders have asked him for advice. The announcement of a Silicon Valley office was an important signal from the top, he says. But entrenched forces inside the Pentagon are not to be underestimated. Newell also worries the unit may find itself overwhelmed. Startups usually begin small and quiet. Low expectations give them the freedom to stumble about until they figure out what they can accomplish. The public nature of the launch has put the office under a spotlight. “Everybody thinks they’re going to come in with a massive budget and do a bunch of things,” Newell says.
For his part, Newell has started to think about how to get the next generation of soldiers and techies to collaborate. Steve Blank, a father of the lean startup movement and a Stanford instructor, has created a Hacking for Defense class to be offered next winter. Military units will provide real-world problems, and graduate students in the business, engineering, and policy schools will work with soldiers, sailors, and other active-duty troops to solve them. Newell will be helping to design and teach the course.