Latest in Podcasts and Multimedia

Armed Conflict Cybersecurity & Tech Executive Branch

Lawfare Daily: Christopher Kirchhoff on How the Pentagon and Silicon Valley Are Transforming the Future of War

Jack Goldsmith, Christopher Kirchhoff, Jen Patja
Tuesday, July 9, 2024, 6:00 AM
Can the Department of Defense innovate fast enough to maintain technological and military superiority?

Published by The Lawfare Institute
in Cooperation With
Brookings

Jack Goldsmith sat down with Christopher Kirchhoff, a former senior official in the Pentagon’s Defense Innovation Unit (DIU) and the co-author with Raj Shah of the new book, “Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War.” 

They talked about the origins and aims of the Defense Innovation Unit, how the defense bureaucracy fought it, and DIU’s successes and failures. They also discussed the pathologies of defense procurement, the relationship between technological innovation and military superiority, and whether the Department of Defense can innovate fast enough to maintain technological and military superiority.

To receive ad-free podcasts, become a Lawfare Material Supporter at www.patreon.com/lawfare. You can also support Lawfare by making a one-time donation at https://givebutter.com/c/trumptrials.

Click the button below to view a transcript of this podcast. Please note that the transcript was auto-generated and may contain errors.

 

Transcript

[Introduction]

Christopher Kirchhoff: The problem, however, is the long lags that have grown up in defense production that just leave this gap between the technology that was baked into the design and what is current at the time. And so you end up with this somewhat absurd situation where a 700 million dollar fighter takes off in 2016 with a processor in it that's slower than what we're all carrying around in our pocket.

Jack Goldsmith: It's the Lawfare Podcast. I'm Jack Goldsmith from Harvard Law School with Christopher Kirchhoff, a former senior official in the Pentagon's Defense Innovation Unit.

Christopher Kirchhoff: Our goals, first of all, were to prove that commercial technology was viable in military missions. That actually on the battlefield, you could take commercial technology directly from the shelf. And it could be effective.

Jack Goldsmith: We're talking about Kirchhoff's new book with Raj Shah called “Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War.”

[Main Podcast]

Chris, the name of your new book with Raj Shah is “Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War.” And Unit X is I guess a nickname for the Defense Innovation Unit where you used to work and what the book is about. So why don't you start off by telling us what the Defense Innovation Unit is?

Christopher Kirchhoff: Sure. The Defense Innovation Unit actually used to have an X by it. It was called Defense Innovation Unit Experimental originally. And Secretary Mattis gave us an upgrade a couple of years into running the office by removing the X as a symbol for making the office a permanent part of the Department of Defense.

So we passed the test and I'm proud to say that here, seven, eight years on, the Defense Innovation Unit is not only still standing, but it’s actually quite a bit larger than it originally was when we founded the office in 2016.

Jack Goldsmith: So what does it what is it, what was it founded to do?

Christopher Kirchhoff: So it was founded at a moment in time when it was becoming increasingly apparent to folks in the Pentagon, that the new technology ecosystem growing up in Silicon Valley was not only producing amazing gadgets that we’re all using in our daily lives, whether that's an iPhone or cloud computing or, social networks, but it was also beginning to produce things that clearly had potential applications in military missions.

And here I'm thinking of Google's self-driving car --- now, a number of Silicon Valley companies actually make flying cars or VTOL aircraft. Around 2016 it became apparent that the Pentagon was not really at all institutionally focused on this whole new technology ecosystem.

It was in fact still very much focused on working mostly with five major U.S. defense primes contracting mostly in really a completely different technology ecosystem. And what was interesting in 2016 is we had a secretary of defense, at that time, Ash Carter, who way back in 2001 spotted the trend that drove the new urgency behind having the Pentagon reach out to Silicon Valley. And that trend was that at the end of the Cold War, government research and development, which had really reigned supreme for a very long time, funding things like DARPA, the National Science Foundation, left the military in particular in a position where it typically invented the state of the art technology.

Folks in the Pentagon were used to having technology that were three, four, five, or even six generations ahead of what was on the consumer market. But that switched. And in 2001, Ash Carter authored an article, a very perceptive article, called “Keeping the Technological Edge,” that noticed the civilian technology system of production --- in other words that people producing technology for the consumer, was growing many times larger than the technology ecosystem that's producing defense only equipment. And what that meant for the Pentagon is that the locus of innovation had essentially shifted; that although the Pentagon lab still produced some amazing technology, the leading edge of innovation was increasingly moving into the private sector, and if the Pentagon wanted to stay on top and continue to bring the best technology to the battlefield, it would have to increasingly seek out that other technology ecosystem.

Jack Goldsmith: So let me unpack a few things there. Why did the switch occur? Part of it had to do with massive differentiation in research and development funds. DOD used to spend a lot more on that than it used to, didn't it?

Christopher Kirchhoff: What's interesting is the government expenditure, both overall on R&D at the federal level. and the Department of Defense research budget has maintained relative parity. It's been, more or less, kind of a flat line since the original drop down at the very end of the Cold War --- the so-called Peace Dividend.

But the biggest change has been the trend line in the commercial sector, which follows the path of a rocket. It's almost straight up. And by the time you get to 2016, when Raj Shah, my coauthor of Unit X, and I arrived at Defense Innovation Unit X, there was this extraordinary situation where the five major tech companies, Amazon, Apple, Microsoft, Google, and Facebook themselves each had a market capitalization that was larger than the entire U.S. defense industry combined. And more than that, Apple at the time in 2016 actually had enough cash on hand that, had Tim Cook wanted to, he could have actually just bought outright every major U.S. defense contractor. So that gives you a sense of just how much bigger that other technology ecosystem that the Pentagon is not really accessing, did not really have the means to access, had grown in the intervening years.

And that was the technology ecosystem that Defense Innovation Unit was purpose built to access.

Jack Goldsmith: And explain to us why the five primes, the five big defense contractors weren't in a position to supply what you think the kind of the commercial sector, as you're describing it, was able to do.

Christopher Kirchhoff: Yeah. It's a fascinating bit of political and economic history, how we ended up with these two completely different systems of production in the United States.

And we typify this actually with the cover image of Unit X, which depicts an F-35 fighter and an iPhone. And those two technologies are illustrative of the two different ecosystems, and the F-35, to give you a sense of why that ecosystem is so different than the consumer ecosystem is, in many respects, a technological marvel. It's a fifth-generation stealth fighter. It has a suite of sensors that is extraordinary. It has armaments that those sensors are linked to. It's nicknamed in the Air Force, the flying computer for the sensor fusion it does. And impressive as all that is the design for the F-35 was finalized, it was essentially frozen, in 2001, which is the year that the Pentagon issued a production contract for the F-35 to Lockheed Martin, and the F-35 didn't actually become fully operational until 2016. So 15 years elapsed between the design freeze and when the F-35 started rolling down the runway. And of course, in 15 years, the state-of-the-art technology literally flew past the F-35.

And not only was the design frozen in 2001, but the way that most complex defense programs are run is through a series of what are called requirements. So these are specifications that are worked out in advance, that are in the moment meant to be the list of capabilities that military planners most believe must be in a new weapon system.

The problem, however, is the long lags that have grown up in defense production, then just leave this gap between the technology that was baked into the design and what is current at the time. And so you end up with this somewhat absurd situation where a 700 million dollar fighter takes off in 2016 with a processor in it that's slower than what we're all carrying around in our pocket.

Jack Goldsmith: Okay. And there's a couple of more background points that I want to get out before we get to what the Defense Innovation Unit set out to do. What you just said touches on something that I've been trying to understand for a long time, and that is the reasons for the slowness and delay and kind of crustiness of the traditional defense procurement process.

Can you just say a few more words about why those companies weren't able to produce these massive military projects --- in this context, a fighter jet --- just say more about why it took so long to get the thing built and why they weren't able to incorporate cutting edge technology in the building of it. That seems, to many people that will be surprising.

Christopher Kirchhoff: Yeah, it's a striking fact, and it comes from a history that is understandable in the sense that when you're looking at building a fifth generation stealth fighter, this is a skill that really only a couple of companies in the world even, one, two, three, maybe four have.

And so you're not operating in a kind of classic free market. You're in a situation where you have just a couple sellers and one buyer, a sort of monopsony situation, as economists would call it, in which ordinary laws of supply and demand don't necessarily apply. And so in order to protect the interest of the American taxpayer, the Pentagon over the years has created a very elaborate system to manage a cost schedule and capabilities of large defense projects.

And this is important because when you don't have supply and demand in a free market, it's possible, as we've seen in many examples, for the taxpayer to be overcharged, for there not to be strong incentives for programs to keep costs down. So as a result, this whole institution of systems that governs not only how production occurs, not only the system of requirements that is generated that a contractor must meet with their design but also actually deeper than that, a system of accounting and auditing that is very bespoke, that is specific to the Department of Defense that is spelled out in defense audit guidelines, and the Department of Defense actually employs something like 3,500 or maybe even more auditors --- we have the numbers in the book --- who work from offices around the world to ensure a very high degree of compliance in all these systems that are being built. And it’s good in the sense that it does protect the taxpayer's interest, but it also creates even more friction in terms of actually just building technology, and more than that in terms of the other technology ecosystem that's grown up to service consumers in the consumer market, it creates a giant barrier to entry, because if you're a startup and you're looking to sell the 25 trillion consumer market --- so your business plan tells your investors, that's how large of a market I'm going after --- the defense market is first of all, pretty small. It's a few hundred billion dollars at the end of the day. And if you're going to go after that, and then someone says you can do that, but you have to set up a whole different system of accounting. That's not commercial standard. That's going to cost a lot of money. You have to have two sets of books and at any time you can get audited and, oh, by the way, you might even have to hire lobbyists as well to be competitive going after contracts in the system. That's how these two systems of production came to diverge and came to be very different.

Jack Goldsmith: Okay. And the last piece is, and this is an important theme in your book, the importance of technological innovation for defense and for American military superiority. And what is the concern here?

Christopher Kirchhoff: I think the concern, and this is something that people at Defense Innovation Unit has studied for a very long time now since, since DIU has stood up, but it's now really becoming visible to anyone even casually watching the news. You just look at Ukraine and what's happening in Ukraine and, just to give you a data point from last month that's really quite extraordinary, the United States gave the Ukrainian military 31 of our most advanced battle tanks the M1A1 Abrams tank, which is an extraordinary piece of machinery. This is the tank that won the first Gulf War. And recently, the Ukrainians, at our request, evacuated all 31 of the M1A1's tanks from the front line because a quarter of them had been killed by Russian kamikaze drones.

In other words, really inexpensive drones were taking out the world's most advanced battle tank. Not only our most advanced battle tank, but the most advanced battle tank that our allies have in their stocks in their armaments as well. So that tells you that technology is moving fast enough to at least in this particular tactical context, overwhelm a core legacy weapon system that is a cornerstone of our armed forces and that is no longer effective in battle. It could be that we are nearing the end of a century of mechanized warfare. If that's the case, pf man-mechanized warfare, then we have a real problem because we've made a tremendous investment in the equipment that you need and the man you need to wage manned mechanized warfare. So it could be it could be the case as we're seeing in Ukraine and also seeing in a couple other places around the world that I'd love to talk more about later with the Houthi rebels and Hezbollah and Hamas. We're seeing really dramatic changes in the way war is fought. And we're seeing the introduction of low-cost systems that are typically autonomous and often powered by AI that are showing remarkable degrees of success against much more expensive and sophisticated weapons platforms that nobody thought could have been so easily defeated.

Jack Goldsmith: Okay. We will come back to those examples. So this is the basic background against which Unit X or the Defense Innovation Unit was founded. So tell us about how you and Raj came to lead the organization in, I guess what you call it, second wave. Tell us how you came to be there, how you found it and what you did.

Christopher Kirchhoff: Yeah, for sure. It's a funny story of sort of happenstance and opportunity, I came into the Pentagon as a young aid in the Obama administration in 2009, and before that I worked as an emergency wartime employee as a civilian in Iraq. And I came into the Pentagon not with a military history or strategy background, but actually with a PhD in political science focusing on science and technology studies. So my own training is in essentially the sociology of technology. And I naturally gravitated at that time to working on issues that involved emerging technology.

First of all, I'm a total technology nerd, so it was fun. And second of all, in 2009, among other things brewing, cyber was becoming a real issue, and the Pentagon was in the process of formulating its first cyber strategy, which I got a chance to work on. So ever since I began in 2009, I focused more than anything else on, on emerging technology.

And similarly, Raj Shah, after his initial service in the Air Force flying F-16s in Iraq and Afghanistan, joined the National Guard and then became himself an entrepreneur and founded a series of startups, including a very successful cybersecurity startup that he later sold to Palo Alto Networks.

And we were both of the view that there was this whole wave of new technology growing up in Silicon Valley. And just like Ash Carter's 2001 paper pointed out, unless the military became, in Ash's words, a fast follower of that technology, an institution able to quickly adopt it and put it to work, that the military would be in danger of falling further and further behind.

And so Raj and I got to know each other and also got to know Secretary Carter through different ways. And he had the great thought of pairing us together because we had complementary skill sets. I had worked policy jobs in the Pentagon and had a lot of relationships in Washington, whereas Raj really knew Silicon Valley and the players, and I couldn't read a balance sheet to save my life but Raj sure could. So the two of us got tapped with this extraordinary opportunity to, with a team of people, get Defense Innovation Unit, which had been started about six months or eight months perhaps before we took over its leadership, to the next level of capability, because it turned out in that office's first incarnation, it was missing a few essential tools that it really needed to succeed, and seeing that it was struggling, Ash Carter made the quite unique decision in Washington to revamp it in a major way and to be very public and open about failures that he was going to be addressing with new strengths.

Jack Goldsmith: But what exactly was DIU, we still haven't really gotten out what exactly it was supposed to do that was different from what had come before. We've talked about, you've talked around it a little bit, but just say exactly what its role was.

Christopher Kirchhoff: The mission of Defense Innovation Unit was to purchase hardware and software from startups and tech companies in Silicon Valley that do not sell to the military and to deploy it in military missions.

So it was to access that other technology ecosystem the military was not at all engaging and to see what in it could actually help advance missions the military was currently taking on.

Jack Goldsmith: Okay. So how did it go while you were there?

Christopher Kirchhoff: It was extraordinary. We showed up at a time when there were headwinds from two different directions.

The first set of headwinds was actually from the aftermath of the Edward Snowden disclosures, which really put a bad taste, understandably, in the mouths of many people in Silicon Valley after Snowden disclosed that the United States government had tapped, among other things, the interconnects of many of the major internet service providers and technology companies for the purposes of foreign intelligence surveillance, because that essentially made the tech companies to many of their customers, and particularly to their customers from abroad, look like they were extensions of the American intelligence system.

And many companies lost business over it, and it was certainly a surprise to a lot of engineers that their communications were being entirely vacuumed up by the NSA. Because of that, when Secretary Carter first flew to Silicon Valley to announce Defense Innovation Unit and his push into Silicon Valley as the first secretary to go to Silicon Valley in over 20 years, he was not allowed on Google's campus.

Having to give the speech he gave at Stanford University, which had become a kind of neutral ground. So that was headwind number one. Headwind number two came from Sand Hill Road. It came from the venture capital firms who fund the startups whose technology we were asked by Secretary Carter to access.

And, in general, in 2016, the view on Sand Hill Road, and again, not unfairly, is that the Department of Defense is an awful customer for startups. Why? Most large contracts led under the federal acquisition rules take 18 to 24 months from beginning to end. And if you're a startup, you can't wait that long to close a deal.

You have to show your investors profit before you're going to get additional rounds of funding to keep your company alive and to keep growing your products. And then even if you got lucky and you won a FAR contract, it's not at all clear that you would automatically win the larger follow-on production contract that typically takes place.

It could be that you're out maneuvered by a firm that knows how to work the system better. It could be that somebody else from a larger company, a more capable company swoops in. So for that reason, the general vibe on Sand Hill Road was that if you want to work with the government, we're just going to withdraw our support from you. We simply won't support you. So we had two urgent missions or urgent jobs to do right off the bat. The first was to convince investors and founders of startups that the Department of Defense could actually be a good customer on their terms, not on the Department of Defense's ordinary acquisition terms.

And then the second one was ideological. It was convincing engineers that have a deep aversion at times to the U.S. military and to the hard power of the U.S. military, that working with the military can be done in ethical and moral ways, and actually advances values that, that are important to all of us.

Jack Goldsmith: Okay. So how did you overcome this? There's a third headwind I want to come back to. And that is the push back from the defense bureaucracy, but we'll come to that.

Christopher Kirchhoff: Oh yeah, that was lots of headwind and a series of stiletto stabs in the back.

Jack Goldsmith: Yeah, which I think have persisted for a long time, but we'll get to those. Let's just talk about the Silicon Valley side. So you talk in the book about how someone on your team reading deep into the National Defense Authorization Act found some authority that who knows where it came from to basically fast track the acquisition process.

So that was one part of the strategy, right?

Christopher Kirchhoff: It's an extraordinary story of what one individual alone can do to an extremely large institution to completely change how it thinks and acts. And the story comes down to Lauren Daly, a lovely colleague of ours who, when we first met her on our first days on the job was 29 years old.

She's from an army family. Her dad was a tank commander and her way of serving was to become a civilian in the Pentagon and to become an acquisitions expert. And she was the first full time acquisition professional to wind up at Defense Innovation Unit shortly after it was founded. And because Lauren is a complete acquisition nerd, she was up late at night reading through the dictionary sized National Defense Authorization Act and looking for new authorities that Congress had granted the department on acquisition.

And she had this lightbulb moment reading one particular sentence in section 815, which gave a new set of permissions to the department in a different kind of contracting called OTA or Other Transactions Authority that had been around actually since the early space program --- NASA used it to contract with mom and pop suppliers during the Apollo era --- but with this new provision in the NDAA allowed the department to do is to take an OTA contract, which you can let in just a matter of days and has really a lot less of the obligations and strictures of a far-based contract. It allowed the department to take a pilot that was successful in its first OTA contract and immediately move it to selling at the production level.

So in other words, you could take a drone, you could prove one drone was successful, and then you could sell 10,000 the next day to the army without having to renegotiate the contract. And because this could be led a contract it 10 to 14 days and was actually quite flexible, it was the first tool, the first acquisition tool, that allowed the department to meaningfully interact with startups on terms that was commercially acceptable to them. And not only did Lauren find this, but with the support of Secretary Carter, in very short order we flew Lauren to Washington --- I went with her and we met with the senior acquisition policy official in the Pentagon to talk about this new authority and to talk about Lauren's plan for codifying it into a new contracting program --- we met with the head acquisitions lawyer and then we met with the Pentagon general counsel. And in just over two weeks, we got the puff of white smoke and created a whole new way for the Pentagon to buy technology from startups that in the intervening years has been used to buy $70 billion worth of technology for the Department of Defense.

Jack Goldsmith: That's an amazing story. How did you overcome the, what I'll call the Snowden problem?

Christopher Kirchhoff: That took longer and it's important to disaggregate Silicon Valley. I think the firms that themselves were directly affected by the intelligence actions that were taken were probably --- or those that had high degrees of foreign market share were most affected commercially by the Snowden revelations.

There are other pockets of people though, Silicon Valley is a large community, who were quite willing to work with the military and curious about it. So it really depended on the company and the set of people. The second year that Defense Innovation Unit was in existence there was a major protest about a project that we were helping move forward called Project Maven, where 3000 Google employees signed a letter of protest because they were enraged and upset at allegations that weren't the whole story, that Google was involved, at least the initial headline claimed, in an offensive drone program on behalf of the Department of Defense. The story is much more complex. It was a machine learning experiment to figure out whether machine learning algorithms could label data in surveillance footage that was used for force protection and defense not offense. But nevertheless it speaks to the tensions in Silicon Valley that exist between engineers and programmers and product managers who came to work in tech companies because they wanted to build products that made the world a better place.

They did not come to Silicon Valley assuming they would ever be working on a project related to the military and certainly not a project that would make the world a more lethal place. Fast forward to today, it's, I think, become increasingly apparent to everybody that we live in a world that is more fragile than we thought, that has more security problems than we would ever wish with a land war on NATO's border playing out and the Middle East in real crisis.

So there's a much greater willingness now certainly than there was when we started, to do work here in Silicon Valley with the government in general and specifically with the U.S. military.

Jack Goldsmith: You described many successes, both in your time there and afterwards. Give us a couple of examples of paradigmatic successes of DIU and quickly addressing a problem, a discrete problem that the Pentagon has to which there's a technological solution that you were able to bring to bear remarkably quickly compared to the normal acquisitions process. What are a couple of examples?

Christopher Kirchhoff: Yeah. Why don't I share one success and one failure. I’ll start off with the success. There was a company, a startup, that we found out about, it was in stealth, called Joby Aviation. And it had offices in a hangar, just actually about a mile north of DIU's offices.

And it was building a VTOL air taxi, a propeller driven electronic aircraft that could take off vertically and then travel like a propeller plane. And it was a really remarkable set of technologies, bottled in something that more or less was like the Jetson's flying car.

And as soon as we met the founders of Joby, we realized that they didn't need any of our money. They were extremely well capitalized, but they were really interested in using a test range, a military test range, that we had access to because you could get on it very quickly. There wasn't a lot of paperwork.

And it allowed them to test quickly and also to interact with extremely knowledgeable parts of the special forces and the Air Force that were interested in their aircraft and what it could do and get them a lot of testing data. After working with Joby for a very short while it became apparent that their prototype aircraft had a number of potential important military uses from troop transport to resupply. And I'm proud to say today that Joby is an operational program and has a fleet of its aircraft in operations for the U.S. Air Force. So that's an example of a program that went from a testing and experimentation phase to full operations in a remarkably few number of years for military aviation.

Jack Goldsmith: That's unprecedented, wasn't it? Basically, that speed?

Christopher Kirchhoff: It was really remarkable. And it's a testament not only to Joby but also to the Air Force's own evolution. And early backing of wanting to join DIU in accessing this new ecosystem of technology and then going a step further and actually setting up its own set of institutions like AFWERX that are purpose built to do the same thing. So that's a good news story. And then I'm happy to share a story that didn't go quite as well. And it starts off with in many respects and an even more impressive set of technology. And this has to do with reconnaissance from space. Another young engineer at Stanford named Payam Banazadeh started a company called Capella Space, named after one of the brightest stars in the sky, And Payam had begun his career at NASA working to develop a satellite that was to use the solar wind to propel itself to the moon --- or spacecraft rather --- and one of the engineering challenges of this particular NASA spacecraft was to design a large solar sail in a very compact space ---something literally the size of a backpacking backpack --- and it turned out that Payam's mastering of that unlocked an advance, a key advance in something called Synthetic Aperture Radar that 70 years of engineers had not been able to achieve. So Synthetic Aperture Radar is truly one of those James Bond kind of technologies. It grew up after the Second World War. It was first deployed on the SR 71 Blackbird, and its superpower is, by using radar waves rather than the visual light spectrum from space, it can see through clouds and darkness. So you're not dependent upon it there being a clear day to use an optical satellite to take an image of a military target.

You can now operate in all weather and day or night. The challenge in SAR and the reason why so few companies try to launch SAR satellites, and why SAR remained a technology exclusively or almost exclusively developed and operated by governments and high-end spy satellites, is that it took a lot of power to create the SAR signals that would bounce off the earth from low earth orbit.

And Payam using his NASA engineering background figured out a way to pack all the power and a big enough antenna into a really small package to be able to launch SAR satellites at economies of scale that the commercial space revolution is driving. So in other words, really small satellites that you can throw on a SpaceX rocket and get up to orbit.

And so we met Payam at a time during the mid 2010s when North Korea was approaching a really scary point of its own nuclear missile program development, where it had figured out how to build fairly reliable intercontinental ballistic missiles, and the range of those missiles was increasing.

But worse than that for those that live in the United States, and particularly those who live in cities in the West Coast the North Koreans were beginning to develop these missiles in quantity. And the great fear was that our remarkably effective missile defense program could potentially be overwhelmed if North Korea was to launch many missiles at a time.

And if that, was going to be the tactical situation, then the only way to reliably defend the United States from nuclear attack from North Korea would be to take those missiles out, in Pentagon parlance, left of launch. And the way to do that is to have massive SAR satellite coverage of North Korea, so you can watch every inch of that country and understand exactly where those missiles are and where you would need to target them if you had to target them.

Jack Goldsmith: You should quickly explain --- I understand, but make sure everybody understands --- what you mean by left of launch.

Christopher Kirchhoff: So left of launch means you need to hit and destroy an intercontinental ballistic missile before it takes off, before it is launched. Once it's launched, your only effective defense is going to be a missile interceptor. And we have only so many of those. And so you have to hit the missile on the ground, which is an extremely hard military problem when your adversary, North Korea, is extremely good at hiding its missiles across multiple places, multiple storage depots, and multiple launch sites in the nation. So you now have a needle in the haystack problem from an intelligence perspective that you have to get on top of.

Jack Goldsmith: Okay. And so what's the, what was the failure?

Christopher Kirchhoff:  So the failure was, we knew that this was one of the great unsolved problems in national security. It was a top five problem that the president and the White House and the secretary of defense wanted to try and solve. If we could develop the right technological solution, and in meeting Payam and evaluating his technology, we realized that if we were able to deploy just a few of his satellites in the right orbit, we could really close the surveillance gap on North Korea.

So we took this idea to the Pentagon. The vice chairman and the secretary of defense got initially very excited about it, and they directed that $50 million be appropriated, given to Defense Innovation Unit X to make this deal, and we tell this story in chapter four of the book. And chapter four tells the full story, but the short version is that the money never arrived. And it never arrived for a set of very complicated political and institutional reasons.

The largest of which was that at the time DIU was pushing forward this alternative approach to SAR surveillance. Another major defense contractor with a spy satellite program in some crisis, a spy satellite program that was years behind schedule and billions of dollars over budget was up for reapproval on Capitol Hill.

And an extraordinary decision was taken to hold back Capella’s technology because of the worry that if it were to advance that it would unsettle the support for the other program on the Hill.

Jack Goldsmith: And so that's that's one example of many in the book, several in the book, where, and I was really surprised by this, it was striking. Unit X had the support of many secretaries of defense --- several, at several points, the director of Unit X was reporting to the SecDef. That implies tons of clout. And yet, there are several stories where the bureaucracy beneath the SecDef and even, and also in the armed services or maybe the appropriations committees in Congress, they seem to have antibodies to you. Your birth, the DIU's birth, seemed to trigger antibodies throughout the Pentagon and the defense establishment. And can you just talk about that in general terms? And it seems to be much better now. It seems that in recent years, DIU has become more broadly accepted in Congress and the Pentagon.

But can you talk about that bureaucratic allergic reaction more broadly?

Christopher Kirchhoff: Yeah, for sure. I'll tell two short stories about just a straight, full on allergic reaction for reasons that were stupid and damaging. And then I'll return to Capella maybe, and talk about some more of the nuances that make others’ judgments more understandable once you see the frame of reference they're coming from.

So the crazy the truly batch of crazy stories are that two days after Secretary Carter flew out on Air Force Two and in front of all of Silicon Valley and on the Pentagon channel, broadcast to bases worldwide announced us as the new leadership of DIU and in a very personal way vested his authority in the mission that we were carrying out on behalf of the Department of Defense, Raj got a phone call from a friend of his on Capitol Hill, and our friend who in the book called Deep Throat says, hey, I just got out of a meeting and you guys have a problem because somebody just whacked your budget. And Raj said, okay, how much did they whack a buy?

And Deep Throat said that's the thing. They whacked all of it. You've been zeroized. Now, zeroized it's a funny word. I didn't even know what it meant until it happened to us, but zeroized means the legislative branch of government has literally written a zero on your budget line for the next fiscal year.

And this is the prerogative of Congress, which is a co-equal branch of government with the executive and our founders ensured that as a co-equal branch, the way the legislature would exert authority over the executive branch would be through the power of the purse, through having to approve every single penny that the executive branch is going to use to carry out programs.

And they could do this and they did it to us. And so here we are, we've just been announced we're literally a couple months away from the next fiscal year. And unless we overturn that decision, we're dead in the water. The Department of Defense would be legally obligated under the Constitutional separation of powers to shut us down.

We would be no more. We were thinking, okay, maybe somebody has a, a better idea of how to spend the money or maybe there's a better approach or what do we not understand here? Because we're trying to do something that we believe is important and the right thing to do.

So we quickly go to battle stations, fly to Washington DC and meet with the two appropriation staffers who got out their government issued pencils and wrote a zero on our budget line. And they each had a different reason for why they killed us. The first represented a member from Indiana and she believed that more funding should be going to Indiana. And we tried to explain to her patiently that 92 percent of startups are founded in California, and as a result it shouldn't be surprising to her that the lion's share of defense innovation is very small budget is going to go to firms based in California.

But she had none of that. And her colleague himself, a former military veteran, when he got a chance to talk to Raj one on one, told him his real reason, why he gave us a zero and cut our budget entirely. And his reason was that he was all set to take a congressional delegation overseas and Ash Carter had denied him Air Force airlift for that. Rather than, take a fancy Air Force gulf stream to visit a foreign military base he had to fly coach and out of spite, he was going to kill Defense Innovation Unit because he knew that it was important to Ash Carter. And, oh, by the way, the second story is that as we were driving, or pardon me, as we were flying to Washington, and we got an email on our United Airlines flight that said your government card has just been declined at your hotel and you need to rebook your lodging.

And we thought, what the hell we just got these government cards. How could we, what's going on. And sure enough, the administrative team who worked for at that time the Pentagon's lead weapon weapons buyer, who was not particularly excited about DIU, had rather than transfer our credit cards onto the new administrative team handling Defense Innovation Unit, had just out of spite, cut them.

Here we are riding in an Uber over to Capitol Hill to get yelled at that we're having to pay for out of our own pockets, and then taking another Uber back to the Pentagon after being yelled at, paying out of our own pockets.

Jack Goldsmith: One criticism I've read of DIU is that it's made many important discreet contributions to discreet defense problems, but it hasn't done what people call producing results at scale and it hasn't transformed the mainstream acquisitions process that we started off with.

Were those ambitions and is that a fair criticism?

Christopher Kirchhoff: I think it's a fair criticism in the sense that, still to this day, even three years now into the Ukraine war the overwhelming percentage of the Department of Defense budget goes to legacy weapons programs that are built in the original technology ecosystem of the military industrial complex. And only a very small amount actually flows to the new defense economy that is typified by companies like Anderol and SpaceX and Shield AI. But, if you rewind the tape back to 2016 when the budget that Congress took away and then begrudgingly gave back, that first full year's budget for Raj and I was $30 million. That's not even one wing of an F-35. It's a very small amount of money. So when we started out we had a couple of goals that were more modest than revolutionizing the entire Department of Defense. Our goals, first of all, were to prove that commercial technology was viable in military missions.

That actually on the battlefield, you could take commercial technology directly from the shelf and it could be effective. And then our second goal was to show that the Department of Defense could be a good customer to startups, that we could execute contracts very quickly on commercial terms that were favorable to a startup and win over many more startups to being willing to work with the government, which at that time they were in general, not.

And then our third goal, because we knew that we were small and we could only get so big so fast, was to inspire other parts of the military to also start working with the commercial technology ecosystem by using the tools that we had developed. And when you step back now, eight, nine years after we set those initial goals, I'm really proud at how the original team and those that have followed us --- there are so many people involved in this effort --- have really met. All of them. Lauren’s means of buying technology quickly has been used to spend $70 billion of the Department of Defense's money to firms that for the most part had never worked with the department before.

So that's a huge win. And if you look at the organizational chart of what's called the innovation steering group, the group comprised of all the different elements in the Department of Defense that are also pursuing this mission, there are now 54 different entities in the department in the military services, some at the combatant commands that are now very focused on accessing the Silicon Valley ecosystem and, turning full circle back to Defense Innovation Unit.

The mission of DIU is now, thanks to Congress and thanks to this last year's National Defense Authorization Act enshrined in law, as is the report of the director of DIU directly to the secretary. And oh, by the way Congress this year gave DIU's budget at just under a billion dollars. And today DIU's third director the former Apple executive, Doug Beck, who was the reserve commander at DIU when Raj and I were there, and he's a tremendous leader, is spearheading the most important initiative in the Department of Defense, the replicator initiative, which is to build autonomous systems at scale.

Jack Goldsmith: So I agree completely that you succeeded in your original goals and that running replicator and getting the billion dollar budget in the last year is undoubted signs of success and growing acceptance, especially in Congress. I wonder though, if you think running the replicator program is what success looks like for DIU?

Tell a few words about what the replicator program is.

Christopher Kirchhoff: Sure. So the replicator program is in reaction to the astonishing developments, not only in Ukraine, but in other battle spaces too. And maybe Jack I'll just share a couple.

Jack Goldsmith: Yeah. Why don't you share those stories now.

Christopher Kirchhoff: The attacks of October 7th, which have set off the worst wave of violence in the Middle East since the 1973 Arab-Israeli war, were precipitated by Hamas fighters coming over an incredibly sophisticated border wall, a set of layered defenses that the Israeli military had erected over decades between Israel and Gaza.

And how did they do it? Unlike the German army in World War II they didn't drive around the Maginot Line. They went right through it by using quadcopters to drop grenades on the generators powering surveillance towers that effectively knocked out the Israelis eyes and ears.

Another example of commercial quadcopters that cost almost nothing defeating a multi-billion dollar defensive line. At the same time that Hamas fighters poured over and kicked off this horrific wave of violence you now have Hezbollah using a combination of cruise missiles, rockets, and loitering munitions to effectively depopulate Northern Israel.

And this is a little bit of an under covered example. Story in the press with so much attention being focused in the south of Israel, but in the north you have 85,000 Israeli civilians that have had to leave their homes because the Israeli military cannot effectively defend their northern border against Hezbollah's attacks, and it's possible that in the coming weeks these Hezbollah attacks could even intensify. And then to make matters even more complex, in the Red Sea, you have the Houthi rebels supplied in part by Iran, using also a set of weapons including most recently, a few weeks ago, autonomous sea drones to attack global shipping.

And this matters if you're an American, because 12 percent of all global shipping goes through the Red Sea, and the minute that that shipping channel is seriously threatened, the global economy is going to take a big hit.

So what this tells me and other people at DIU, is the future we feared, where the proliferation of low cost microelectronics really does change the game, and in particular changes the game against advanced militaries who might have poured enormous amounts of money into advanced weapons platforms that are nevertheless vulnerable to these means of attack. We are now, whether we like it or not, in an environment where a great deal of our technological edge has been eroded and because we have only begun modest preparations for this new era of war, we are behind. We are quite a bit behind the power curve.

And that's why replicator is so important. Replicator is the first initiative spearheaded by the Deputy Secretary of Defense Kathleen Hicks and the secretary of defense to begin a game of catch up, to take these tactics currently being used against us and our allies and use them to go on the offense.

Jack Goldsmith: But scoping out more broadly though, so that's one program, that's one discreet and important program. But just looking at, we started off by talking about the DOD procurement and acquisitions process and whether it was up to the task of keeping the United States at the cutting edge of technology to maintain superiority over adversaries.

And in light of the examples you just gave, where are we on that? Probably not in a very good position, are we?

Christopher Kirchhoff: I think we have a long way to go. There are some new developments that are really exciting. The first is that there is now much more of a consensus, both in the Pentagon, at the White House and in Congress, that this pivot to the consumer technology ecosystem is truly important, particularly with the arrival of AI, which is growing up purely in that ecosystem and not at all in the defense technology ecosystem. The second thing that's happened is thanks to the great work of DIU and those other innovation entities and the $70 billion that's gone out the door since 2016, there is now a great willingness of venture funds to back startups that are looking to target from the beginning the defense market.

And indeed, in a really important respect, Defense Innovation Unit has become its own unicorn factory, helping to create new companies, product companies, that are as agile as any other tech companies, but are developing really astonishing technology for the defense market. And just to take one of them, Anduril, which I think is important to focus on because it's a company that has taken on such impressive approaches to both software and hardware, Anduril just advanced to the final competition stage in the contract for the Air Force’s CCA program or Collaborative Combat Aircraft, which is itself a revolutionary program. So this is a program to put 10,000 supersonic stealth drones in the air to fight alongside with fifth generation fighters like the F-35 and other drones and airborne weapons platforms. And so imagine this: the Air Force is saying that we would like robotic drones to take over the lion's share of air war conflict. And we're going to turn to a new set of companies to develop the hardware for this and potentially, Anduril, being one of the two final performers, General Atomics being the other.

So if you just look at that alone, I think that's a great example the leadership that Frank Kendall has provided the Air Force, Frank being the secretary of the Air Force, to introduce the biggest change to air war doctrine literally since the Wright military flyer. And to do it with the help of a Silicon Valley unicorn company, that Defense Innovation Unit helped get off the ground.

Jack Goldsmith: So here's my last question. And it's a variation of stuff I've already asked. What does success in this whole area look like? Does the idea that DIU is going to foster companies that end up competing with primes for bigger projects? Is it going to be that? That DIU helps foster integration between commercial sector and the primes. Does it involve changing the acquisition process much more to look like the acquisition process DIU has been fostering? What is success in the future among all these issues that we've been talking about look like?

Christopher Kirchhoff: The man who gave Defense Innovation Unit X its name is Deputy Secretary of Defense Bob Work, and he was unquestionably its most fierce backer in the early days.

And he grew fond of observing that the Department of Defense doesn't necessarily have an innovation problem. It has an innovation adoption problem. In other words, it has a cultural need to get much better at adopting advanced technologies that it has within arms reach. And so I think what success looks like here for DIU and for the armed forces as a whole, and for the nation, is going to be cultural change, not only at the Pentagon, but even at the primes.

In the whole system to look towards adopting technology much more rapidly in the ways that startups build and adopt technology, not in the way that technology was developed for the F-35. And to do that in a way that creates such overwhelming superiority that the United States military, as it has for much of our recent history, is able to very effectively deter conflict from ever breaking out in the first place, because war is not only horrific and economically destructive, but with today's technology it's even more horrific.

It's not something we would ever want. It's something that we should avoid at all costs. And so my greatest hope is that we will use the incredible innovation potential of this country to deter conflict. That I think, Jack, is what success will look like.

Jack Goldsmith: Christopher Kirchhoff. Thank you very much.

Christopher Kirchhoff: Thanks for having me.

Jack Goldsmith: The Lawfare Podcast is produced in cooperation with the Brookings Institution. You can get ad free versions of this and other Lawfare Podcasts by becoming a Lawfare material supporter through our website, lawfaremedia.org/support. You'll also get access to special events and other content available only to our supporters.

Please rate and review us wherever you get your podcasts. Look out for our other podcasts, including Rational Security, Chatter, Allies, and The Aftermath, our latest Lawfare Presents podcast series on the government's response to January 6th. And check out our written work at lawfaremedia.org. The podcast is edited by Jen Patja, and your audio engineer this episode was Cara Shillenn of Goat Rodeo.

Our theme song is Alibi Music. As always, thank you for listening.


Jack Goldsmith is the Learned Hand Professor at Harvard Law School, co-founder of Lawfare, and a Non-Resident Senior Fellow at the American Enterprise Institute. Before coming to Harvard, Professor Goldsmith served as Assistant Attorney General, Office of Legal Counsel from 2003-2004, and Special Counsel to the Department of Defense from 2002-2003.
Christopher Kirchhoff, an expert in emerging technology, helped create the Defense Innovation Unit, which he continues to advise. During the Obama administration, he was the director of strategic planning for the National Security Council and senior civilian adviser to the chairman of the Joint Chiefs of Staff. Chris obtained an AB degree from Harvard College and a PhD in social and political sciences from Cambridge University
Jen Patja is the editor and producer of the Lawfare Podcast and Rational Security. She currently serves as the Co-Executive Director of Virginia Civics, a nonprofit organization that empowers the next generation of leaders in Virginia by promoting constitutional literacy, critical thinking, and civic engagement. She is the former Deputy Director of the Robert H. Smith Center for the Constitution at James Madison's Montpelier and has been a freelance editor for over 20 years.