Future Wars: The Nexus of Technology and the Military

Panelists discuss the Silicon Valley’s role in the future of war and whether or not the Pentagon is successfully innovating rapidly enough to keep up with the technological changes facing the military. 

Copies of Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War will be available for purchase.

ROBBINS: Great turnout for considering how hot it is out there. Thank you all for coming. And, Anne, thank you for doing this. It’s probably even more hellish in D.C. So welcome to the— 

NEUBERGER: Exactly. 

ROBBINS: Yes, exactly. (Laughs.) Yes, exactly. Exactly how bad is it there? (Laughs.) So welcome to today’s Council on Foreign Relations meeting on Future Wars: The Nexus of Technology and the Military. I’m Carla Anne Robbins. I’m a senior fellow here and cohost of The World Next Week podcast for the Council. You all know our speakers. You have their bios, so I’m just going to give you a few highlights. To my right, Christopher Kirchhoff is a founding partner of the Pentagon’s Defense Innovation Unit X, tasked with bringing innovative commercial technology to warfighting. Is that a good summary of it?  

KIRCHHOFF: Sure.  

ROBBINS: OK.  

KIRCHOFF: I’ll take it. 

ROBBINS: During the Obama administration, he was also director for strategic planning at the National Security Council and the senior staff civilian aide to the chairman of the Joint Chiefs of Staff. He is the coauthor of the new book, Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War.  

Speaking to us from Washington, D.C., Anne Neuberger is deputy national security advisor for cyber and emerging technology. Previously, she served as the National Security Agency’s director of cyber security. She also led NSA’s election security effort and served as assistant deputy director of the Operations Directorate.  

And Raj Shah here is managing partner and cofounder of Shield Capital, and previously ran DIUx. He serves as a reserve F-16 pilot in the U.S. Air Force and has completed multiple combat deployments. And he is the coauthor of Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War. So now we’ve gotten two plugs for your book. (Laughter.)  

OK. Anne, Raj, Chris and I will chat for about thirty minutes, and then we’ll turn questions from our members. And as a reminder, this meeting is on the record. So, Raj, can we start with you?  

SHAH: Absolutely. Thank you, Carla. 

ROBBINS: Great. So in the fall of 2021, then Chairman of the Joint Chiefs General Mark Milley described China’s surprise testing of a hypersonic missile as “very close” to a Sputnik moment. And we all know that a Sputnik moment suggests that the U.S. was not only caught off guard but was also seriously lagging behind a technological and military rival. Are we lagging behind the Chinese that much in military technology? 

SHAH: Carla, thanks for—thanks for the CFR for having us here and for moderating. You know, there’s been a lot of discussion and consternation around the rise of China, their intentions, and our military capability. And certainly, there are some specific technology areas—hypersonics, things that can kill carriers, that they have some advantages on. But maybe we step back and think about, you know, what are the—what are the more strategic or broad categories that China is competing against us in a national security fashion? So one is just, like, mass. Like, how much are they investing? How many ships? How many airplanes? And their significant growth in their capability. And, of course, they only have to defend or think about it regionally.  

But the other one, that I think is particularly relevant to the work that Chris and I did, is their concept of civil military fusion. How are they integrating with their commercial sector and using that or having that prepositioned for national security? So, for example, the world’s largest manufacturer of small quadcopters is called DJI. And they are the best ones in the world. They make more. They probably have 90-95 percent of the consumer market. And they can snap them out, you know, by the—by the millions. There is no such thing as a true break between the private sector and the CCP, right? The CCP has representatives in the company and on the board. 

And so in a time of conflict, you could see them saying, all the next drones that come out will be going to the PLA, they will have facial recognition and, you know, munitions attached to them. And we don’t have that equivalent here in the U.S., for good reason. And maybe we can talk later about, you know, what I think our response should be. But you think about that in AI, you think about that in cyber, you think about that and all these other emerging technologies, they have an influence in the commercial sector that we don’t. And that could be a source of that imbalance.  

ROBBINS: So that’s part of their strength, but how much do we have to worry about how far ahead are they? And particular areas do you think we should be most concerned about? I mean, Sputnik moment’s a pretty strong, scary idea. 

SHAH: It’s a strong statement. You know, and, as I say, we shouldn’t let a good crisis go to waste. So, you know, how can we then use that to increase energy in our innovation sector to make the right investments and to counter that. Particularly, again, if you think about if, God forbid, we had to have a conflict in the Pacific, it’s a really long way from Japan and Korea to China. And those open waters with, you know, large ships, like aircraft carriers, are fairly vulnerable. 

ROBBINS: So, Chris, the war in Ukraine has been another major learning moment about technology in modern warfare. You’ve written about how at the start of the war the Russians were the ones weighed down by obsolete 20th century technology, but it now seems they’re adapting at a chillingly fast rate and with very low-cost technology. And particularly how Russian drones have forced the Ukrainians to pull back our very advanced tanks. Can you talk a little bit about adaptation and the sort of notion of low-cost technology versus the sort of things we’ve been investing in for such a long time, and whether that’s a major disadvantage we have to worry about. 

KIRCHHOFF: Well, Ukraine is an incredible crucible for the future of war. And without question you can catch glimpses there of the future. And Raj and I were privileged to have a chance to visit Ukraine last fall, at the invitation of the general staff. And it was profoundly gratifying there to actually see in action some of the systems that we had, earlier in our careers at Defense Innovation Unit, helped start. There’s about thirty or thirty-five different systems by U.S. startups and tech companies presently operating in Ukraine. More importantly, there’s an incredibly diverse, and actually even more active, set of native Ukrainian innovators. Over 200 companies working in garage shops, some of which Raj and I got to visit, that are trying out technology that in many respects has actually surpassed the most advanced technology deployed by big-name defense contractors, and also by many of the startups that Defense Innovation Unit has backed.  

But I think, importantly, we’re seeing in Ukraine some really dramatic developments. And some of them have just happened over the last four to six weeks. And I’ll just mention a couple. One, and I think everybody here in this room should pay close attention to this, we gave the Ukrainians thirty-one U.S. M1 Abrams tanks. This is not only the most advanced battle tank in U.S. arsenal, it’s the most advanced tank in the world and in the arsenal of any one of our allies. And in the last six to eight weeks, because of advances in Russian kamikaze drones, the Russians have effectively destroyed a quarter of those M1 Abrams tanks. And as a result, the Ukrainians have had to evacuate them from the front.  

And if you step back and think about that, I mean, that’s an incredible moment in the history of war. That tells people like me and Raj, that are looking out for innovations in the battlefield, that we could be at the end of the century of mechanized—of man mechanized warfare that began in the First World War. You know, are we at a moment, kind of as mechanized warfare itself changed—or, replaced cavalry—where everything is changing? And if you think too about the investments that the U.S. military has made in so many of its major weapons systems—tanks right now being able to be killed by kamikaze drones, carriers and surface ships potentially being extremely vulnerable to hypersonic weapons—it could be that the arsenal of democracy, this incredible fighting force that five or ten years ago would have been the unquestioned victor on a battlefield, is no longer the force we hope it to be. 

ROBBINS: That’s a rather chilling assessment. (Laughter.) Anne, you’re—I want to talk to you, because you have done some very innovative things working with the civilian sector. But before that, I would love to hear your reaction to that assessment from Chris. You know a lot about technology. Is it that frightening? 

NEUBERGER: I very much agree with Chris’s perspective. So just take a step back for a moment. You know, it’s no surprise that the evening before Russia’s invasion of Ukraine the first really herald of their invasion of Ukraine was a cyberattack on a commercial satellite company providing communications to Ukraine’s military. And the only reason that didn’t have a devastating impact on the Ukrainian military’s ability to communicate was because of the innovative SpaceX approach. And what’s really innovative with SpaceX, and so different from the traditional military approach, to Chris’s point, is that we in the military have always focused and invested in exquisite, low number of capabilities—like the Abrams tanks. The SpaceX approach was, let’s deploy thousands of devices, expect them to be out there for four years, connect them in a way that all communications—(inaudible)—over to others, which made it resistant to the most advanced EW capabilities, in a way that our high-end, exquisite satellite systems are not.  

So that arc tells—is a good example of three parts of this. One, the integration of technology into warfare and the role they play. The disabling attack the Russians wanted to conduct was on a satellite system. We’ll come back to the fact that it was a commercial one in a moment. The fact that what enabled those communications to continue was private sector innovation, that had adopted a very different model but was also more resilient than traditional military communications. And then finally, the ongoing impact of electronic warfare, because militaries have become much more digitized but have not necessarily considered how to defend those digital networks. And the recognition that because offense is so much easier than defense, offense just have to find the signal, block it or jam it. Defense has to be able to resound—you know, be resilient against many different kinds of cyber, electronic warfare, laser attacks.  

We have moved to a digitized military but haven’t necessarily built a way to ensure it can be resilient and fight through degraded environments. And I think a big part of where we need to go are innovative efforts as a result that can leverage the innovative script. We have to recognize that a big part of why those innovations, notably a SpaceX, notably the cloud—where Ukraine moved a lot of their data in order to be able to continue to maintain their government services—came from private sector innovation. Leveraging that in the context of the way we fight wars today and in the future has to be our approach.  

And we have to also consider some of the complicated policy questions that raises with regard to companies like these, that are providers to militaries and civilian systems. That company that was hacked by Russia on the first—on the eve before the invasion also provided, you know, satellite connectivity to thousands of homes across Europe, who were also disrupted. So how we think about those issues are a part as well, along with the technology of today and tomorrow’s war. 

ROBBINS: So the word “exquisite” is one that Secretary Gates used a lot. It’s not the first time that people who get into the acquisition business, and certainly talk about military policy, have used it. And it’s a wonderful word, because when he would say it, he’d say, I understand the impulse here. We would love to spend billions of dollars and do absolutely the exquisite most fighter plane. But let’s produce something that we can actually turn out fast and actually does the job. And let’s move on and continue to—(off mic). What policy changes are you talking about, given the lessons that we’ve learned now from those opening hours of the war in Ukraine? You talk about resilient, Anne. You know, what are you guys talking about in the White House now?  

NEUBERGER: One, innovations and really promoting innovations, like DIUx, which Chris and Raj are here to talk about, with regard to how the military can work in a more effective way with private sector innovators to bring capability in. That’s the first part. The second part is how within our national security environment, within DOD and the intelligence community, how we leverage technology to be more effective. Think, for example, about AI.  

AI’s ability to, for example, bring together in a multimodal way what we would traditionally keep siloed, geo-imagery, signals intelligence—so images and voice. And bringing that together rapidly to enable more rapid indications and warnings of a threat, to put together, you know, patterns of behavior that we know tell us that something’s about to happen. A high-level visitor going to a missile launch site, for example, we all know is an arbiter something to come. Putting those kind of patterns in, and using AI to better help us leverage intelligence to generate more highly and capable intelligence, is another example of that. And looking at the policy and legal implications of that can ensure we’re fully promoting that, resourcing it, and encouraging it in our processes.  

And then finally, it really is our international partnership. What the administration has done—President Biden has really put a big emphasis on allies and partners. That’s reflected in, for example, today’s NATO meeting, and in the rollout of, last year, a virtual cyber incident response capability at NATO. Where, for example, when one NATO member faced a major cyberattack from the Iranian government, Albania faced that, there’s now a coordinated way for individual NATO members, in that case six countries, to surge support and to surge capacity to help countries recover faster.  

So those are kind of three kinds of policy approaches that we’re working on. And I really want to note the efforts that Chris and Raj have done. To build those partnerships, to bring in innovation, and then scale them into the battlefield are hard to do and tremendously powerful in the context of the number—in the context of the battlefield that we’re facing in Ukraine. 

ROBBINS: Thanks. Thank you for that. 

So, Raj, you and Chris and DIU you had considerable success bridging the cultural as well as the contracting gap between the Pentagon and Silicon Valley. And you write about how it did encourage some change inside of the Pentagon, which you’re clearly frustrated hasn’t encouraged as much change. And you also write—I think you mentioned in The Atlantic article that you guys wrote—I did do my homework—that there were things that were provided to Ukraine that Ukraine needed in the first hours and days of the war, but that the Pentagon hadn’t even purchased. Some of the things that that you guys have had helped encourage—you planted the seed for it, but the Pentagon didn’t have it. They didn’t have an ability to roll it out. They didn’t have—can you talk a little bit about that, and what has held the Pentagon back from taking advantage of what you guys built? 

SHAH: Sure. Look, I think there’s been a sea change in this relationship between Silicon Valley and the Pentagon. And I use Silicon Valley broadly—not just that region but startups. If you think back to ten years ago when we ran into the job, this was right after the Snowden disclosures. So a lot of companies viewed the NSA and the government almost in the same light as they viewed a foreign adversary. The procurement process in the Pentagon, right, was an eighteen-, twenty-four-month timeline. And so most startup investors, venture capitalists, would advise their companies not to work with the government. Back to my first company I founded, most of our engineers were from the NSA. And the advice we got was, don’t sell to the—sell to the government. And so you had this—you know, you had this disconnect that was both philosophical as well as just sort of business related. It was the Pentagon was a hard place to do business with.  

But I think we fast-forward now, right, ten, twelve years later, and it’s a totally different world, right? There are—I’ve never seen more entrepreneurs that want to build companies focused on national security. There are more investors putting money into early-stage companies here. And the government itself—and I’ll cede here to Chris in a minute to tell a story about this—the government itself is putting a ton of money into fast contracting. And so, look, I think it’s very different than it was. And if I go now back to your opening question, what do we worry about with China, you know, I think the real change is technology is accelerating, and the group that is able to understand, ingest, and operationalize that new tech—be it a drone, be it some AI thing, some cyber—is the one that’s going to prevail.  

And so we win by betting on the American way, which is capitalism, free movement of talent, free movement of capital. And I think, ultimately, that will prevail over an authoritarian sort of, you know, command economy—not even a command economy, I shouldn’t say—but authoritarian view to how you incorporate that technology. Which I think is evidenced by the fact, you know, how many large language models are in the U.S. versus China. But, you know, there’s also been some pretty amazing contractual changes. And I’ll turn it over to Chris.  

ROBBINS: I’m glad that you’re so optimistic. But, forgive me, you wrote in The Atlantic: The United States is beginning to learn these lessons, but we’re not adopting them fast enough. And you also noted, but because the department had purchased very little of the arsenal developed by the DIU, it had almost nothing to send. That doesn’t sound all that optimistic. 

SHAH: Maybe I’m naturally an optimist. (Laughter.) It’s my fault. Look, there’s certainly challenges, right? And, you know, your—The Wall Street Journal had an article today talking about how some of our weapon systems, in addition to the tank, are faring. We have these Excalibur artillery shells. We have something called a ground based small diameter bomb. And both have, again, been withheld from the conflict because they are not survivable in this new electronic warfare. So we’re seeing sites of it, or glimpses of the future. And I think, you know, we—it’s incumbent on the Pentagon and people in the valley to learn those lessons, particularly on software and speed.  

ROBBINS: So how much—you know, we all know about the way the politics of this works. People build things in everybody’s congressional district. We know that the dominance of the primes, all of that. How much change is taking place in the Pentagon, and how much does it—how encrusted is it still? And if you were going to kick them to change, what would you be kicking them to do, Chris? 

KIRCHHOFF: Well, first of all, you need steel-toed boots if you’re going to be kicking things around the Pentagon. But we’re actually lucky to have here in the room the single greatest heroine of Defense Innovation Unit. Her name is Lauren Dailey. She’s seated up here at the first table. When Raj and I took the job and showed up in Mountain View, California that first week, we knew that the most important thing we could do was figure out a way to contract the startups on terms that were favorable to them. And if we didn’t figure out how to do that, nothing else would matter.  

So when we met the folks on the team that first week, Lauren came up to Raj and I and said, you know, I have an idea. And we said, tell us more. She said, well, was up late at night reading the National Defense Authorization Act—which is as big as a dictionary—(laughter)—and I found a sentence in Section 815 that could change everything. And so Raj said, oh boy, tell us more. And she said, well, actually, I’ve written a twenty-page paper describing how we can use this new authority to create a fast way to contract with startups.  

And so Lauren and I got on a plane to Washington. In short order, we met with the senior policy acquisition official in the Pentagon, the senior acquisitions official in the Pentagon—acquisitions lawyer, and the Department of Defense general counsel. And with the support of Ash Carter, in two weeks pushed through a new authority called the CSO, a new way of contracting through OTA, or other transactions and authorities, that allowed us to take that eighteen to twenty-four month process and shrink it down to as little as two weeks. And because of that innovation—and, by the way, this was codified as Department of Defense Policy in exactly two weeks. It goes two weeks from Lauren surfacing the idea to Ash Carter changing the directive. Fast forward eight years later, that method that Lauren invented has been used to buy $70 billion of technology by the Department of Defense. (Applause.) 

ROBBINS: I feel like the skunk at the garden party. (Laughter.) Seventy billion dollars over how many years?  

KIRCHHOFF: Eight. 

ROBBINS: Hmm. How big is the defense budget for acquisition? 

SHAH: Eight hundred (billion dollars). 

ROBBINS: Hmm, well, OK. Lauren, I think it’s fabulous. And I read about this in the book. And I think—I love the nerd. And someone who worked at The Wall Street Journal and actually reads the budget, I think it’s fabulous what you do. 

KIRCHHOFF: Before you go more skunk—(laughter)—let’s put one last—one last fact on the table. So the—arguably the biggest doctrinal change happening right now in the U.S. military is happening in the Air Force, under the leadership of Frank Kendall. Who has decided to take something we experimented with at DIU, a collaborative combat aircraft—so this is a stealthy, AI-powered drone—and to buy 10,000 of them for the U.S. Air Force to fly alongside a manned fighter aircraft, completely revolutionizing the way that we fight in the air domain.  

And right now, that contract is in its final stages. It’s down to two players. Now, you would think, you know, at first glance, that maybe this is, I don’t know, Lockheed and Boeing. Nope. The two finalists are Anduril, a company that was started with the help of early contracts from DIU, and General Atomics, another non-sort of-mainstream player. So the Pentagon is, I think, and Raj thinks, on the verge of making big changes, not only in the systems that it acquires, what those systems do, but the companies they acquire them from. 

ROBBINS: I want to turn it over to members, but I do want to ask Anne a final question here. So the Biden administration has been trying to balance the need for safety and innovation in AI. And this is obviously an area in which trying to get the balance in the relationship between government and the private sector is so essential. Are we—can you talk a little bit about the challenges of that balance, and how much that dominates your life? (Laughs.) 

NEUBERGER: You know, it’s interesting to do this. One of the reasons when Chris and Raj reached out to me and asked me to join them on the panel I smiled and was very happy to, first, to celebrate the work that they’ve done. Because having worked within the bureaucracy for eighteen years to try to bring innovation forward, promote the people who are doing it. You know, those battles are shared battles. And all of us who have been through those want to celebrate every victory on the journey. But it took me back, you know, really eleven years ago, to exactly this question.  

So at the time that Snowden—the Snowden incident occurred in June of 2013, I was running a sensitive part of NSA that dealt with a lot of classified, sensitive work with companies. Many of those had been established in a post-9/11 environment, you know, when companies came to the U.S. government and said, how can we help? We recognize—you know, we see global terror. We see leveraging ubiquitous, really American often, communications and technology services, because they were good and they were global. We want to help fight terror together.  

And over that ensuing ten years, those became strong partnerships. And when the Snowden incident occurred, you know, I was really at the front line of those first phone calls from companies coming in—you know, came in saying, what’s happened? Where do we go? And I called Chris, who was then, if I recall correctly, a special assistant to the chairman. And, as Chris was always ready to do, jumped in a car, drove up to Fort Meade. And talking through it that day, I think we saw that it was fundamental to change the relationship between government and the private sector. That relationship plummeted, as both Chris and Raj mentioned at the beginning, and, you know, U.S. companies, many of them, stepped back and said, we’re global players. And working with U.S. government, notably with the intelligence community and NSA, is not something—it’s putting our businesses and our internal cultures at risk.  

And I think what’s changed over the ensuing ten years has been a recognition that American values, who we are as a country, who we are as a democracy, is on the front lines. It’s in the battlefield, whether that’s in Ukraine, whether that’s in other conflicts around the world. And the role of innovation in protecting those liberties is aligned with our values, aligned with our laws, absolutely. That always comes first. And aligned with the greater government of willingness, to your question, to be transparent—here’s how we operate; here’s the laws and policies that guide our work—is a part of rebuilding, you know, that partnership.  

You know, a quick anecdote. When I was at the National Security Agency, I recall one of the big steps in trying to rebuild transparency, you know, the agency worked with the Department of Justice and declassified and released, you know, a ninety-one page policy guideline of how NSA protects U.S. persons. It was an initial very—you know, and it showed just how heartfelt it was. It also showed how inadequate it was, because nobody’s going to read a ninety-one page policy guideline, right? We have to, as government, as the military, the intelligence committee, come to where Americans are, come to where our private sector is. Talk about how we operate. Talk about the missions we fight. Talk about what we face around the world as part of rebuilding that partnership.  

And I think we are in—that continues on the upswing. And a big part of the policy changes that Chris just outlined will then also make it easier. Because there’s a larger value question, and then there’s a practical, pragmatic question of can companies actually bring their tech to the fight? And because of things like our acquisition processes and bureaucracy, that’s often been hard, even for those companies who very much want to. 

KIRCHHOFF: Carla, and I will say that phone call I received from Anne was the single worst phone call of my career, because the next thing I had to do was walk into the office of the chairman and the vice chairman and tell them about Edward Snowden. Which I wouldn’t recommend having to do. But it’s an incredible arc, really, to have gone from that moment where Anne and I knew that everything would change with our relationship with Silicon Valley, to here we are after the invasion of Ukraine, after the Hamas attacks on Israel on October 7th, after the threat to shipping in the Red Sea by the Houthi rebels, after the rise of China. And today people in Silicon Valley have a very different understanding of how precarious our security is in the world, and how the U.S. government and U.S. military might be part of the solution. 

ROBBINS: So I want to open it up to members. You can join our conversation with your questions. If you could raise your hand. We’ve got mics here. And a reminder that this meeting is on the record. We have a hand up here and we have a mic. If you could identify yourself, that would be great. 

Q: Yes. Hi. Dick Foster, former senior partner of technology at McKinsey. 

So tell us about what’s happening in Russia and China. And how innovative are they? And what does it imply for us? 

ROBBINS: I tried to get them to answer that question, and they wouldn’t do it. So thank you. (Laughter.) 

SHAH: That’s great. Dick, we’ve not met but I have your book on innovation in my office from my old McKinsey days, so. (Laughter.) 

Q: Which book? 

SHAH: Both. 

Q:  Good. 

SHAH: You know, look, I think as we think about innovation, let’s talk more about China and the commercial side, right? I mean, they’ve got a couple of structural advantages. They have a lot of people. So if you look at the number of AI papers, there’s a debate on the quality, but there’s certainly more of them coming from China. If you look at data, right, they have the data advantage because the lack of privacy laws, if you will. Then they have access to a ton more data to train things on. And they’re spending a lot of money, right? Chris actually was part of this national—this strategic competitive project that analyzed that and flagged how much they’d spent.  

So you would think, with all of these advantages—and if we just maybe focus in on AI—they would be clearly dominant. But it doesn’t seem to be the case, as we look at generative AI. And so I think there are still some structural advantages here for us that we can take advantage of. It’s a great place to live. We can get immigration. So, you know, I think what we’re seeing is a lot of the core innovation still occurring here. But now let’s go from AI to drones, right, which is a little bit older. AI is new.  

You know, some of the earliest small drones were here in the U.S., right? We built those things. There’s virtually no commercial drone manufacturer left here in the U.S. And it’s not just the low-cost drone, right? Like, you could say, maybe it’s just a cost advantage. No. They’re actually building the best software for these drones because they’ve got an install base, people are using them. And so I think the fear—I think the worry for us should be, as these other new technologies come about are we the ones that are going to get to be able to scale them? 

KIRCHHOFF: You know, one further follow-on point on that. You know, it’s very clear that China has a manufacturing base many times the size of the U.S. and our European allies combined. And that’s a problem. I think the opportunity is that, you know, we’re just in the first inning of the convergence of autonomous systems and advanced artificial intelligence. And so if we’re the nation that presently has at least a modest lead in artificial intelligence, then in theory we have the opportunity as a nation to mobilize that and transform autonomous systems to be more advanced than our adversary’s autonomous systems. So I think in the next five years that’s probably one of the most important focuses for the U.S. military. I know Anne and her colleagues in the technology directorate of the National Security Council are very focused on this. And so where we go there will be extremely consequential. 

ROBBINS: Anne, would you like to jump in here? You come from a world of assessment. 

NEUBERGER: I would just add two quick points to really the excellent points Raj and Chris made. One, the partnership between the countries is incredibly concerning. What Russia has always had is cutting edge tech, but not an industrial base to scale it or innovate. And what China has definitely is the industrial base to scale. So the partnership we see—and I would note Iran is in the mix as well, notably in the drone area—is really concerning. It also has made it far harder for the U.S. and our allies, notably in Europe, to control sensitive technologies making its way to Russia. China’s role in that, through front companies and in some cases with the knowledge of the Chinese government, is deeply concerning as, you know, as we fight to control the advancement Russia’s making in its military through individual components that they are not able to produce, which China has stepped in and in several cases just provided. 

One other point I would note is that, you know, underpinning all of these autonomous technologies is data. And China’s presence in global telecom networks around the world and global smart cities applications around the world, millions and millions of cameras, enabling the training of facial recognition models, is also a noted advantage that they bring to this partnership that we’re watching very closely and are very concerned about. 

ROBBINS: Thank you for that.  

I’ll have another question right here. 

Q: Alex—(inaudible)—Management. 

You mentioned the Ukrainians have their homegrown expertise. And you also focused on the U.S. homegrown growth of technology. And how do we co-opt non-U.S. technology—finance, co-opt, help, grow, integrate into our defense systems? Because obviously they have experience that we won’t have, and they’re every day learning and adding to that database. 

SHAH: Yeah, I think it’s a great opportunity for us to learn, right? To see the future. When we visited, we got to go to an electronic warfare test range where they were flying live drones against some Russian electronic warfare systems and seeing what was working and what doesn’t. And they would fly something. It wouldn’t work. They’d land it. Some guys would write some code. And, you know, they’d send it back up that afternoon. And so, you know, that—I think that pace is what’s really interesting. And what I’m hopeful and optimistic—obviously I’m not working in the Pentagon now—is that, you know, we’ve got advisors there and folks that are absorbing these lessons and helping bring them to our industrial base.  

ROBBINS: Right there. Thank you. Yeah. 

Q:  Esther Dyson from Wellville. 

So I’m very much not trying to lead any witness, but we’re talking about Chinese and Americans and governments and private sector things. And then there are individuals like Elon Musk who, one way or another, set, implement, influence policy. He’s got his own relationships, good and bad, with Open AI, and so forth. What happens when individuals become something akin to international political players, and when sometimes international political players seem to be akin to unexplainable human beings? (Laughter.) 

KIRCHHOFF: Well, you know, Esther, we are, indeed, in a brave new world. And that brave new world is part of a generational shift in how technology is produced, in the sense that if you rewind the tape to the end of the Cold War, U.S. government labs like DARPA were two, three, four, five, six generations ahead of what was being produced in the commercial sector, on the whole. And that technology would trickle down eventually into the consumer market. And this is because of the amount of R&D funding that was at that time going to military and federal labs.  

The rise of the consumer technology economy and the $25 trillion consumer market it services now dwarfs, by many orders of magnitude, the defense R&D complex. To the point that by the time Raj and I got to Silicon Valley in 2016 Apple, Microsoft, Amazon, Google each had a market capitalization larger than the entire U.S. defense industry combined. And that disparity has actually only increased with Nvidia and with other companies in the years since. So we’re now living in a world where the locus of innovation has shifted, in a generation, from government labs to the private sector.  

So the government is now—instead of developing the secret sauces on its own, it’s having to go to the private sector to get it. And that’s a completely different dynamic. It not only required the Pentagon to reinvent, with Lauren’s help, how technology is bought, but now it’s going to require very different relationships between sovereign states and individual entrepreneurs who effectively control technology that only superpowers once did. 

ROBBINS: Can I just develop on that? Because that’s a fabulous question. And ask Anne a question. Anne, you had raised this—the role of private satellite network—Musk, and the role of communications in Ukraine in the early days of the war. And there had been some reporting, which I think not sure if it’s been questioned or debunked, that Musk at one point had had tried to turn off the communications because he didn’t like something that was going on in Ukraine. I mean, can we allow individuals to have the power of a nation-state like that? I think this certainly comes in as well on the question of AI, as well. Where is—if Musk had so much potential power over Ukraine, doesn’t that raise some pretty fundamental questions for the U.S. government? 

NEUBERGER: I won’t be to the specifics of that situation, but I think, really building on Chris’s point, as the locus of innovation has shifted from the U.S. government to the private sector, the unique, deeply sophisticated technologies are being—are available for both consumer use, commercial use, and military use. And the rules of the road for that kind of environment are new. And we’re working through and thinking through this. So in some cases, adopting the technology—you know, like, SPACECOM has efforts to try to build a larger constellation of low earth orbit satellites to adopt that kind of meshed network approach that’s both more resilient, enables innovation in a more rapid way, is one approach. But how we think about what those rules of the road are is a key issue that now the war in Ukraine has highlighted. And this example is a very prominent example of that. 

KIRCHHOFF: And you know, Carla, Anne, actually was—the leadership of Anne in the White House, and this is an under-told story of the Ukraine war, I mean, Microsoft essentially went on war footing to defend Ukraine’s internet infrastructure against Russian cyberattacks, in close coordination with the U.S. government. At the same time, Amazon used its Snowball portable terabyte hard drives to back up the records of the Ukrainian government and take them to a safe place, so that Russia did not have the ability to destroy Ukraine’s tax laws and property records. So this is placing U.S. corporations in the leading role in a global conflict, in a way that is unprecedented. 

ROBBINS: Which is a wonderful thing, because Brad Smith’s a great guy. But relying on, you know, the kindness of strangers is I’m not sure a great national security policy. 

SHAH: I guess, what I would say is that it’s not kindness but it’s incentives, right? What is the—what is the U.S. government? It’s the largest buyer of things, right? It has the power of the purse. So we can set the conditions to incentivize companies to be supportive. And it’s also, I think, a human condition, right? At the end of the day, these are organizations run by people. So how do you build that trust and understanding and sort of commitment to democracies over autocracies? And I think these examples that Anne and Chris just gave are encouraging that we’re moving in that direction. 

ROBBINS: Hmm. Woman right here, and then the gentleman behind her after that. 

Q: I’m Beth Hillman from the 9/11 Memorial Museum. 

Since you just mentioned democracy, I have to ask—and state actors versus private actors, and what’s happening here in the United State. The anti-authoritarian, you know, more regime of greater freedoms, that has been the advantage that you describe the United States as having. Should the civil service change dramatically, should the Biden administration end relatively soon, you know, what would be the consequence on the kinds of initiatives you’ve been promoting? 

KIRCHHOFF:  Well, I can certainly say that the next U.S. president, whoever that is, faces fundamental decisions about the kind of military that we’re going to field and build. And those decisions are urgent, because global changes, as we’ve talked about in this discussion, are happening with astonishing speed. This is not something that we can wait five or seven or nine or ten years to solve. This is something that we have to take on now. And, you know, just to bring up a couple illustrative examples, you know, for one—the cost of one U.S. aircraft carrier, you can field, I think it’s, 17,000 sail drone—autonomous seagoing drones, with comparable intelligence, surveillance, and reconnaissance packages on them, that can operate for years on a time—at a time.  

You know, another question. How many Starlink terminals could we buy for fifty-four key security partners with that same amount of funding? So the next president faces incredible and important, and even grave, decisions over what kind of military the U.S. should build and rebuild. 

ROBBINS: And the next Congress. I mean, if you look at what has passed, you know, the House for decision on submarines versus joint strike fighters. I mean, most of the debate right now is not about drones and new technology. They’re still fighting about how many more nuclear submarines you want to have and how many more joint strike fighters you want to have.  

SHAH: We’re starting to see change there too. So, you know, when Chris and I had the opportunity to lead DIU, we spent the first six months just fighting for our budget, which is about 30 million (dollars). Not a lot. This last Congress has given DIU and the military a billion dollars to go and invest and build these types of technologies, particularly around autonomy and a program called replicating. I think you’re starting to see change. And, yes, it’s still a small percentage, but it’s directionally in the right way.  

ROBBINS: The gentlemen, here. 

Q: (Off mic.) 

ROBBINS: Can you wait for the mic, please? 

Q: John Levin, Levin Capital. 

I know the Air Force, and perhaps other elements of the Defense Department, are working actively on quantum computing. Is quantum too early? Is it relevant? Does the United States have a relatively strong position in quantum or not? Because that is a rebalancing thing that some people think is significant. 

SHAH: Quantum, no doubt, will have, you know, tremendous impacts on cryptography and other things. But, you know, in my kind of view, quantum has always been five years away. And so, but maybe this is we would defer to Anne. 

ROBBINS: That’s for Anne, yes. (Laughter.) Can you explain quantum computing to us? (Laughter.)  

NEUBERGER: You know, it’s so interesting. Building on this point, a potentially cryptographically relevant quantum computer—one that can break the cryptography that we use to shop online, to protect the nation’s secrets, would be a very powerful capability. We believe adversaries are a decade away. I’ll note that the National Institute for Standards in the U.S. will be releasing U.S. standards that are the result of a multi-year process later this month. So this is really exciting. We are moving to the next generation of cryptography, U.S. leading. A number of our allies and partners around the world will be using U.S. post-quantum algorithms to ensure that we’re protected around the world. 

One other really promising area that’s actually more mature in the area of quantum is quantum sensing. That could provide a way to navigate through a GPS-denied environment. We’ve piloted efforts with several militaries in other countries. And I think, particularly given what we’re seeing with GPS spoofing and denial in a number of combat areas, that is an area that we intend to accelerate, you know, with key partners, under the quantum working group that we have here at the White House. So watch this space. 

ROBBINS: A woman here at this table here. Thank you. 

Q: Lauren Wagner with Radium. 

How do you think about open-source systems and national security? 

SHAH: Open source in terms of software?  

Q: Exactly, yeah. People have said it’s amazing, it contributes to national competitiveness. Other people have said it’s very risky. How do you think about it?  

SHAH: You know, I don’t—it’s not a problem that’s probably unique to the military. I think large enterprises and corporations also have to grapple with this. If you’re not writing all the code, how do you ensure the security of it? How do you ensure what’s there? But what I think the advantages is it’s fast, right? There’s a lot of lot of change. I think the other corollary, though, which I would maybe expand, is this idea of open mission systems, which is something that the military is trying to do. Which is to say, can you separate the hardware from the software, right?  

Some of these acquisition processes that Chris described are really great for buying an aircraft carrier, for something we’re going to keep for fifty years. But it’s really not good for software, which changes—you know, our iPhones get updates almost daily now. So can you disaggregate the hardware from the software, where we have one system that buys things we’re going to keep for a long time, but they have an open system where you can try different types of software. And the cost of making a mistake is low, because you can always go back and change it to the previous revision. 

KIRCHHOFF: I know Anne will have a view on this question as well, but to bring back the audience here to the test range that Raj and I got to visit. You know, it wasn’t just about drones themselves that were using AI and new sensing technologies to be able to try and navigate through GPS-denied environments. There’s a whole stack, actually, of technologies based on open source that Ukrainian companies and drone operators are drawing upon. So for example, one of the coders that we met at the test range had used open-source software to be able to fuse sensor data from multiple drones, including video feeds, that you have effectively a kind of command and control picture that is built off these drones, and surveillance drones, and different EW feeds. So you’re—in essence, you know, the Ukrainian companies have figured out how to almost create the full stack of kind of NORAD, AWACS—you know, the battle management system that it took us a generation to create, simply using open-source software and Android hardware. 

ROBBINS: Here. 

Q:  Thank you. Rebecca Patterson. Appreciate the conversation today. 

Staying on national security, one thing I’m thinking about with both cyber and AI is the benefit cost, building alliances, having a centralized policy where we’re sharing with our allies to strengthen those relationships, because we’ll need them, versus a decentralized process. Because even with our allies like India, when we see Modi going to Russia giving Putin a bear hug, you know, governments change, people change, people have different incentives—to use your word earlier. I’m curious where you all stand right now as we go forward on building governance safety measures around AI and cyber, un-centralized versus decentralized?  

SHAH: Hmm. Well, that is a great point, especially this week, right, the 75th anniversary of NATO, the conference occurring right now. I think, you know, that’s the other advantage that a place like America has, is we have friends. You know, and multiple friends. And so close interoperability and having standards with them is super important. NATO’s taken some interesting steps here in terms of commercial technology. They’ve put together a billion-dollar euro fund that twenty-two partner nations have invested in, to make direct investments in young technology startups today in Europe, and hopefully later here in North America. So I think your point is well taken. I think it’s really important. And there are signs of us moving that way. 

ROBBINS: Anne, on something like AI, in which the EU has got out ahead on rules regulations for AI—probably got out ahead faster than it understood where the technology was going on generative AI—how much is of our relationship with our—some of our closest transatlantic allies is seen as competitive on a regulatory, on a governance role? And how much of it is—do you feel that we’re working together with them on that?  

NEUBERGER: You know, it’s an interesting question. So governance aspect, with regard to, for example, you know, looking at lessons learned from cyber as we look to AI in a governance context. You know, all U.N. countries signed up to a set of norms with regard to cyberspace in 2015 and reaffirmed them in 2019. What’s been so challenging is what are the consequences for countries that violate those norms? And there’s been an increase over the last couple of years of countries, because there is power in numbers, calling out activity that violates those norms, as part of taking first steps to then what are those consequences?  

And I’ll note that NATO, for example, did its first attribution when a NATO member, Albania, was attacked by Iran. Its government systems were taken down in cyberspace. The country’s government was not able to serve its citizens. NATO attributed that to Iran and the United States supported Albania as a consequence, as well as designated the actors in Iran. So that set of steps—both having the governance and then how you enforce it—there are lessons learned in how complex it’s been in cyberspace as we look to AI. 

With regard to your question from a regulatory perspective, which is where are the—what are the requirements we put on AI companies to ensure that they have the trust of the public, that they’re adequately red teaming systems, that they’re testing them, that they maintain an ability to maintain and update them as new vulnerabilities may be learned of? And just thinking about how attacks against the model can be—can be—how models can be resilient against the attacks on the system. You know, we and the EU have really had different regulatory approaches. The U.S. approach has often put innovation first and security second. I think in the EU it’s put security first, innovation second. And our country’s relative innovation bases reflect that.  

AI gives us the opportunity—and the way the president’s talked about it is safe and responsible AI. We’ve got to be able to do both from the beginning, because the technology is so powerful we have to consider building in security and safety along the way. That’s why the president negotiated agreements with a set of companies on how they’ll do red teaming, how they’ll do testing within the executive order with regard to how models—how companies must share information regarding models above a certain size. And we’ve been doing a lot of work thinking about what needs to be the rules before models can be deployed in critical infrastructure, because of the significant risk that could bring to the operations of the infrastructure our citizens rely on. And that’s where the U.S. and EU models, I think, are coming closer together with regards to what trust and safety has to be built in, and areas where we have a very low tolerance for disruption. 

ROBBINS: Thank you. So we’re coming to the end. I apologize to—I know there are more questions out there. I just wanted to give Raj and Chris and Anne a chance to—one minute—a one minute—a point that you didn’t get a chance to get out there.  

Chris, you want to go first? 

KIRCHHOFF: Well, again, I would just take a moment to step back and look at this extraordinary moment that we’re at. And in a way, as tragic as the developments in Ukraine and in the Middle East are, they are the best opportunity to have a real life wake-up call for the stakes of the decisions that are before us. And it’s my greatest hope that we continue moving in the direction that things seem to be swinging, which is towards taking more risk on innovation technology, rather than taking even more risk by doing things the way that we’ve been doing them for a while.  

ROBBINS: Anne, last thought. And we really appreciate your taking the time. 

NEUBERGER: Of course. You know, it’s said that there are no atheists in foxholes. I suspect there are no bureaucrats in foxholes either, right? There’s a reason that we see so much innovation coming from small countries in rough areas of the world without a lot of resources—Estonia, Israel, Ukraine. There’s a lot we can learn from that. So the challenge to all of us—and no matter whether in the private sector or government, it’s all of us—to bring that foxhole mindset to work every day. Whether we are a company producing innovative tech, and to think about how can this help with the difficult situations we see our allies face? And if you’re in government, what rule or process—there’s risk of doing; there’s risk of not doing—what rule or process can be waived to get things faster, in a more timely way, to those who really need it to defend who we are around the world? 

ROBBINS: Thank you. Last word to you, Raj. 

SHAH: I’d say, you know, as we think about betting on the American way, I would just say, you know, individuals matter. The DOD is a three-million-person bureaucracy, but now even as a story we just have Lauren, and many, many others. Like, individuals can make a—make a difference. And so I think it’s incumbent on all to try to do—try to do something.  

ROBBINS: Thank you so much. I’d like to thank Anne Neuberger, Christopher Kirchhoff, and Raj Shah. And the book is there. And thank you, Mike Froman, for hosting us. And thank you everybody for coming. (Applause.) 

(END) 

This is an uncorrected transcript.