When Raj Shah was tapped to launch the experimental Defense Innovation Unit (DIUx) in 2016, he could barely secure a budget.
Raj was a former F-16 pilot and Wharton graduate. He knew that the kind of innovation that drove Silicon Valley could help close the yawning gaps he saw in US capabilities. But the rest of the DoD needed convincing. For years, he ran around the Pentagon trying to squeeze funding from career bureaucrats to build the technology he believed would shape the future of war, like drones and artificial intelligence.
Now, nearly a decade later, DIU boasts an almost $1B budget (as of EOY 2024). Raj left to run Shield Capital, a deep tech-focused VC that raised an inaugural $186M fund in 2023. The DIU – through its many iterations – has helped grow wildly successful companies, like Anduril, Shield AI, and Saildrone, and helped many more through the door of the Pentagon. Some might call that a major glow-up.
Tectonic sat down with Raj just before President Trump’s inauguration to chat about all things defense innovation. He told us about his experience at the helm of the DIU, where he thinks the DoD still needs to innovate, and what the US needs to do to deter (or defeat) China. Plus, he dug into the technologies he’s most excited about as an investor.
Note: This interview has been edited for length and clarity.
Tectonic: Is the DoD still broken? What are the most urgent things that need to change?
Raj Shah: Broken is a strong word. I think the DoD is the world’s largest bureaucracy and has an incredibly difficult task of deterring wars around the world and winning wars. It needs to provide the president with options to defend and protect American interests of peace and democracy around the world. It’s extremely challenging.
The DoD does some things really well. There isn’t an innovation problem in the ranks; The men and women who serve in uniform are some of the most creative, dedicated, and selfless people you’ll ever meet. These men and women will take whatever kit you give them and ensure that they don’t fail in their mission. They do it at low pay and in harm’s way.
We have some legacy capabilities that are also pretty amazing. We can reach out and touch the entire world. Force projection through aircraft and aircraft carriers and ships—-that’s working. Let’s first recognize that. However, we have this system that is designed for a world that is evolving faster than the system is. We are no longer in the 1950s and 60s where you have a bespoke industry of defense contractors. Where you build some amazing technology that allows you to have supremacy for 10, 20, or 30 years. Today, we’re in a world where commercial technology, such as cyber and AI, and commercial space are moving fast. Our systems are not built for that. So, in that sense, it is broken in that we can’t get the latest capabilities and technologies to our men and women in uniform fast enough.
Why? For a whole host of reasons. Software is just different than hardware. It needs to be built, bought, and maintained differently than how you would a long-term asset. So that is broken. The other part that is broken, of course, is how we budget. That is not solely the department’s problem, but also the legislative branch. We’ve had continuing resolutions for 30 of the last 49 years. We’re in one now. And that impedes the ability to innovate and move quickly. There is no certainty about how much money we’re spending. The way oversight is done doesn’t give the department flexibility to move and adapt to these changes. So you know, I would say in some ways we have the world’s best fighting force, but we can’t rest on our laurels, and if we don’t recognize and move more quickly, starting now, that might not be the case in just a few years,
What technology do American warfighters need most, right now?
I think it’s two things. One is the speed of change, meaning technology that can adapt to the realities of the battlefield. The Ukraine conflict has shown us the future of war. You see suicide drones, both in the air and on the ground, you see this complex electronic warfare environment, with GPS jamming and jamming of signals. You see war where things are changing so fast that the cycle time between what a bad guy or adversary innovates and our response is no longer years, or even months or weeks. Sometimes it’s days and hours. So how do you how do you do that? How do you move that quickly? I think that’s one thing our service members lack. That’s not just software, but also drone advances.
The second thing is around cost. There are some unique advantages to having something that is just cheap. Cheapness gives you mass, which means you can have lots of numbers of it. If an artillery shell costs $20,000, and now a drone costs $1000, you can replace things. It gives you new standards for how you think about reliability. If you launch 1000 drones and 100 of them crash on takeoff, maybe you don’t care, because they’re so cheap and will be replaced next week. That’s what our service members lack: Things that are very cheap and the ability to just adapt to the technology. You don’t want an aircraft carrier changing its form every week, right? Like we’re going to keep that thing for decades, but small drones, electronic warfare, sea drones, this is where we need a different approach.
Where is the US defense industry currently weakest? Where should priorities lie?
Software. The best software engineers in the world don’t go to the defense legacy primes. The best software engineers go to places like Google, Microsoft, and Amazon. Now, they’re going to the newer generative AI companies because it’s fast-moving. It’s at the cutting edge of technology. In particular, as we think about artificial intelligence and generative artificial intelligence, the pyramid is steeper for talent. Fewer people are truly world-class. You need to attract them and bring them into your organization. That’s the primary thing we lack.
And how can those people be better attracted?
One of the ways is to work with the companies they’re already in. On the positive side, I’ve never seen more interest from engineers, scientists, and young companies in working on national security. People realize that we can’t take democracy and democratic market values for granted, seeing what’s happened in Israel, in Ukraine, and threatening to happen in the South China Sea. Even companies like OpenAI and Anthropic have partnerships with young defense companies to serve the market. I think you just work with them and be demanding as the government for best of breed. If you want autonomy, analytics, and AI, I’d say that Valley-style companies will likely give you a better outcome.
Can the US actually catch up to China? If there was an invasion of Taiwan tomorrow, what would happen?
You know, there’s a famous Churchill quote where he says, “Americans can always be trusted to do the right thing, once all other possibilities have been exhausted.” In times of crisis, we have shown, time and time again, that we can muster a whole-of-nation effort. My hope is that we don’t wait until that moment of crisis to begin doing that.
In terms of Taiwan, it’s hard to say. How do you put enough strategic doubt into Xi Jinping’s mind and the minds of those in the military? How can we make them think that it would be foolhardy to try to invade? The ideal end state of a strong military is the deterrence of war. We need to continue to show with pure military capability how we could put Chinese assets at risk. But we can also use industrial policies to show that.
Why do I think that America has the right to win here? It’s twofold. First, the free flow of capital and talent leads to a better outcome. That’s the heart of entrepreneurship. So, having these partnerships between innovative tech companies and the Pentagon, I think, is very powerful. The second advantage we have is that we have friends and allies in the region. I think ideally these things combined create enough doubt for Xi and make him think that it would not be in his own self-interest to try to take Taiwan. That’s the ultimate goal. In general, any kind of large-scale war there would be bad for everyone, particularly given our reliance on advanced semiconductors from Taiwan.
So the DoD should show its dedication with more partnerships with these smaller defense companies?
Yeah, yeah, exactly right. And I think the department is doing that. The question now is how to do it at scale. So it goes back to what you say is broken. If we truly believe technologies like AI and autonomy are going to be decisive, what percentage of the budget and what percentage of people should be allocated to it?
I’d estimate right now it’s a low, single digit. It should be 5%, should be 10%, should even be 20%. This is the future. Let’s put a concerted effort behind it. But then again, it’s not just the Department of Defense’s prerogative. This is where Congress also plays an important role. The two should come together and say, “We need massive, massive reform.”
I think the current crescendo of interest is remarkable. I had the honor of getting to work for Secretary Ash Carter almost 10 years ago, when he was setting up the DIU and asked me to run it. At the time, it was a battle to convince people in the Pentagon that Silicon Valley-style companies were important and could be trustworthy partners. It was a battle to convince investors here in the valley to invest in defense-oriented startups. Neither of those two cases is true. Over the last three years, $140B of capital has gone into national security startups. Compare that to the hundreds of millions per year a decade ago. We’ve seen significant contract wins from startups. I think the tune is changing in the DoD. Now we just have to ask, are we really going to do it at scale? And are we going to reform the hard parts of the budget?
What do you expect from the second Trump administration in terms of defense innovation?
Look, I think it’s a significant positive opportunity. There is a group of reformers that will come into executive roles. You’ve got leadership in the Senate and the House that have spoken enthusiastically about reform. You’ve got Silicon Valley technologists, like Elon Musk, who are now in the new administration’s orbit and understand how to build this tech. I think there’s an incredible opportunity for the new administration to scale these changes, and I think they have a willing partner in Congress to achieve it.
And do you foresee any obstacles to defense innovation with the new administration?
There is no shortage of obstacles. That’s why it’s the government. You’ve got so many entrenched interests from, you know, bureaucrats that have been there a long time, from congressional districts that have to answer if there’s been a change or some technology moves. Of course, the enemy has a vote too. But that’s our system. Like, this is nothing new. And I think we can work through it. There are plenty of historical examples, again, of our gearing up, like in World War Two, pushing out the B29 and B17. Today’s equivalent is AI and small drones. We just have to have confidence, be willing to make changes, and be creative.
What areas of investment are you currently most excited about at Shield Capital?
There is no shortage of this convergence of national security and commercial technology. There are four that we’ve been particularly focused on. There’s artificial intelligence, of course, understanding all types of different data and making decisions at machine speed. Generative AI can improve workflows and help humans distinguish noise from signal, or signal from noise. At the end of the day, national security is a human endeavor.
There’s also cyber security. We’ve seen the recent revelations about what China has been doing in our electric grid, and what they did at the Treasury. So how do we protect our networks? How do we make sure that they are reliable, available, and trustworthy?
Commercial space continues to be an exciting area, from sensing to on-orbit operations. As costs of launch continue to drop, with SpaceX driving that, that will be an exciting area.
And then finally, robotics and autonomy. So, drones on the sea, underwater, and in the air. All of these have important commercial applications too. Some companies may start in the government sphere first and then go commercial, and others may do it the other way around. Those are the areas we’re looking at, mostly. There are also some emerging areas of technology in energy, materials, and bioscience that we are also keeping an eye on.
And are there any companies or tech in particular that you’re especially excited about right now?
Well, we love our portfolio. There’s a whole slew of great companies in our portfolio. But even just going across those four technology areas, we’ve got companies that are doing advanced manufacturing for space, a company called Apex. We have companies doing sensing on orbit, Albedo and Hawkeye. We have one called Starfish that’s doing some on-orbit remote operations. On the cyber side, lots of defensive capabilities. And then on AI, we’ve got a few companies that are building next-level infrastructure. So I think there continue to be a range of interesting companies.
What is your view on defense tech valuations? Do you think that companies are fairly valued?
One of the things I sometimes worry about is that in the valley and the investment world since you know, the dawn of time, there have always been bubbles. People get really excited about an area because there truly is underlying value and opportunity, and so it attracts more capital, which is a good thing. But you know, sometimes naturally they get over-inflated, particularly by people that or groups that don’t understand this market. So as much as we love defense and national security, and are focused on it, it’s still a tough market. You’re not going to totally change how government works, things just take longer, and there’s a unique approach to how things get done. So, I do fear that there are tourists entering this market and driving up some valuations, and I don’t know if it will be sustained. The natural end state to a bubble is deflation or a burst. Then people leave, those tourists leave, and that can have deleterious effects on companies. I’m hopeful it will be a smooth outcome, but there’s definitely in some stages some valuation behavior that is difficult to explain.
A new generation of defense tech startup founders is emerging, many of whom lack prior military or government experience. Does this represent a competitive advantage or a potential challenge?
That’s an eternal question, right? Is it better to know a ton about an industry and disrupt it, or to know nothing about it and be naive and disruptive? You’ve got countless examples of of both. Look, I think the good news is that even if you don’t know this industry well, you can learn about it. There are lots of really great people coming out of government and the military that will join your team. If you’re really good at a particular thing, like a particular technology, but you don’t know the customer base, you can add people to your team who can help. I think you’re seeing that across the board, particularly in this generative AI space. So I don’t think it’s a barrier.
Of course, there’s always risk. There’ll be some great companies, and there will be some not-so-great companies. And I think as the best entrepreneurs go out and look for capital, having investors that have deep expertise in this area is key. It takes a decade or two to get that kind of expertise; it can’t be developed overnight. Some firms will be more helpful to their founders than others. There’s tons of risk. You have to treat your investors very, very carefully. As a former entrepreneur myself, once you have an investor and a board, they’re there for the duration. So choose wisely.
What is the path to exit, liquidity, or profit with these companies?
That’s still to be written. That’s one of the challenges in this market. If you’re really successful, then there’s the public markets, or even private markets and you stay independent. We’re seeing several companies that are at that level, and that’s wonderful. I think part of the challenge is that there hasn’t been as large of an ecosystem of acquisitions for companies that don’t quite reach the stratosphere. This is where, again, evaluations run ahead of progress and there may not be a soft landing. When your buyers are historically thinking about EBITDA multiples as how they value a business, that doesn’t really work for a fast-growing, young startup that may not be able to spell EBITDA.
How has your perspective on innovation changed going from the DIU to VC?
Yeah. Well, there are two angles. Temporally, I was at DIU a long, long time ago. It took a lot of convincing on both sides to get people excited and to find that match between capability and need. I think that’s a lot easier now. But in some senses, there hasn’t been that much change. It’s just fundamentals. In both cases, you want a great team that’s tackling a really important market and has a unique approach to the technology. That’s the same in the government, at DIU, and in the private sector, like at Shield. It comes down to, as with most things in life, the people. Are they motivated? Are they smart? Are they hungry, are they ethical? Those are the people that we love to partner with.
It’s an incredible time for founders to build businesses supporting national security. We’re at a time in the world and in the market where I think you can do well and do good at the same time. It’s really important in an increasingly turbulent and dangerous geopolitical environment for our best engineers to be working on this. They can be very satisfied and fulfilled by the change and advancement they’ll bring. But it’s also a great market opportunity. You can create important businesses that create value for founders and their investors. You can do well and do good. I encourage folks to take that leap and build a company. And, you know, give us a call. Happy to talk to a founder.