Pod Insights a16z Podcast Why Speed, Not Size, Will Define the Next War About this episode As global tensions rise, AI and autonomy are transforming how nations prepare for conflict.
In this episode, Horacio Rozanski, CEO of Booz Allen Hamilton, and Gary Steele, CEO of Shield AI, join Erik Torenberg to discuss how technology, speed, and public–private partnerships are reshaping America’s defense strategy.
They cover lessons from Ukraine and Taiwan, the rise of autonomous systems, and why the future of warfare will be defined by software, agility, and innovation.
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Speaker 00 We are in this really very challenging, very dangerous geopolitical moment, perhaps the most challenging and dangerous of the last 50 years.
0:08 Speaker 01 An autonomy will play a critical role. It's redefining the way conflicts have been fought in the past.
0:13 Speaker 00 Ultimately, it'll be the faster nut, the bigger that will win.
0:17 Speaker 01 It just takes too long for the government to buy things today and take advantage of things.
0:21 Speaker 00 Seventy-five, eighty years ago, a new technology platform could last thirty, forty, fifty years. And if it took you three to five more years to get there, you could still be ahead. That no longer works. The U.S. has to have a response that cannot be driven solely by the government, solely by the existing defense base, solely by the valley. It's really going to take everybody.
0:40 Speaker 01 It's life and death.
0:41 Speaker 02 Today, you'll hear a conversation from Horacio Rozanski, CEO of Booz Allen Hamilton, and Gary Steele, CEO of Shield AI, about how technology and partnerships are reshaping America's defense strategy. We'll discuss lessons from Ukraine, the future of autonomous systems, and why the U.S. needs to build faster, not just bigger, to maintain its edge. Let's get into it.
1:05 Speaker 02 Gary, Horacio, welcome to the podcast.
1:07 Speaker 00 Thank you.
1:08 Erik Torenberg? Thank you.
1:09 Speaker 02 Because of Booz and Shield and your other positions, you guys have a front row seat into what's happening in defense. I want to get to the big questions right away. If China were to take Taiwan, or to try to, are we ready to defend it? What does being ready to defend it even mean?
1:23 Speaker 00 I think the answer is yes. Indo-Pacom has been thinking about being ready to fight tonight for several years now. And the nation is ready. The political will under this administration is there. We're recording today on Columbus Day, so Happy Columbus Day. But it's also the day where the president is in Israel and Egypt announcing major peace agreement, which was driven by American power, American superiority, and our ability to project that into a peace-through-strength framework. And I think the same is true. In the Pacific, I don't think that the country is interested in conflict. I think the nation is interested in peace, but I think is certainly prepared and able to do what it needs to do.
2:07 Speaker 01 And what I'm super encouraged by is the ability to bring the world's best technology to the fight. And I think one of the reasons that we have a unique opportunity is to take that advanced technology and apply it to deterrence and to winning any conflict that we might get involved in. And I think we're in a unique position today, and I don't know that I would have said that three or four years ago, but I actually feel that way today.
2:29 Speaker 00 Absolutely right. Your guys' work on autonomy would be central to any effort in the Pacific.
2:36 Speaker 01 And I think we've all watched the various conflicts that have gone on around the globe. I think we're much smarter as a result of that, and we're going to apply all that knowledge and innovation to solving the problems that our military has, and I think we can save a lot of lives as a result of that.
2:50 Speaker 00 Yeah, the other thing I would say is I've spent time in Taiwan, and they are getting ready. They're moving fast, they're very determined, they're very focused.
2:58 Speaker 01 And they have a very futuristic view of how to apply technology in the right way to get the right outcomes.
3:04 Speaker 00 Hundred percent.
3:05 Erik Torenberg? Let's get deeper on this for people who are not as familiar into the specifics. Say more about what being ready means. You guys think we're ready, some people think we're not ready. What is the disagreement about, actually? And one thing you just said is, three or four years ago, you might not have said that we're ready, or it might not have been as confident. What changed? So maybe let's get deeper on what does it mean to be ready? Like, what are the actual specifics of that? And then, sort of, how has that evolved over time?
3:27 Speaker 00 Readiness is a combination of political will, capacity and capability in the region, and as Gary said, having the right technology positioned in the right place to do what it needs to do. Everybody's been looking. Well, the thing that struck me the most when I was last in Taiwan is how closely they had looked at both Ukraine and Israel.
3:48 Speaker 00 To understand what they needed to do. One of the things, for example, that they've done that I think is encouraging is they've extended the length of the draft of their military service so that people can be properly trained. And that's actually very helpful. And in addition, they've passed a budget and they're actually spending against the key technology priorities that matter the most. And the same thing is true for Indo-PACOM and for the US. There's a lot of capability now deployed from space to undersea that could make a decisive difference. And of course, that is going to accelerate over time. I think that this administration is really bringing new tools, new technology to bear,
4:30 Speaker 00 whether it's in C2, whether it's in autonomy, whether it's in AI use more broadly, that will enhance an already pretty awesome fighting force.
4:39 Speaker 01 And the thing that I've seen change fundamentally is the world shifted from a technology point of view, the venture community shifted, the opportunity shifted for entrepreneurs to begin to fund innovation that could be directed towards the defense of the US and to be used in circumstances like Taiwan. And so as the world shifted, the investments, the focus, a lot of the work that had been done traditionally for the commercial world has gotten applied to defense. And we're seeing that innovation curve and that innovation opportunity play out as it relates to Taiwan.
5:21 Speaker 01 And as Arisha was saying, if you look at what happened in Ukraine, everyone's been watching what's going on. There's been a lot of education where you've seen a very big shift in the way conflicts get fought. And that's educated the Taiwanese on how to place investments and think about innovation and technology in a fundamentally different way.
5:41 Speaker 00 Exactly right.
5:42 Erik Torenberg? Let's zoom out to just compare more broadly, US and China as it relates to sort of military readiness. What are our weaknesses relative to theirs? And what are some stuff that they do that we don't quite do yet or can learn from?
5:53 Speaker 00 I would say, obviously, in this context, their primary strength is proximity. And so we need to project power where this is much easier for them to do in this context. And so our entire strategy needs to be one of understanding that fundamental asymmetry, if you will. Our strength lies both geopolitically in the fact that we have true allies and partners that are going to fight with us. Secondarily, that our armed forces truly are second to none, both in capacity, ability and talent. And third, as Gary is saying, is the engine of progress of the American economy translated in this
6:36 Speaker 00 moment in technology continues to give us an edge. Now, you're going to hear me talk about speed all the time. China is not standing still. And so while we still have a clear technological edge on some of these fundamental technologies, we need to continue to accelerate the deployment, the use and the development of these technologies in the real world because they are accelerating themselves.
7:04 Speaker 01 I think that's the thing that the US government has to improve on. The speed with which they adopt new technology, the speed with which they consume and participate in that innovation process is going to be critical. And there have been improvements. It is getting better, but we still have a long ways to go. To be able to take things that are relatively new and think creatively about how do you get these things deployed broadly across the military to really empower them to do what they need to do.
7:32 Erik Torenberg? Exactly. How should we think about whether our defense industrial base can sustain a prolonged great power conflict with China? And what else should we be doing that we may not be doing at the moment as well as we should?
7:43 Speaker 01 The thing that has to happen is we have to unlock speed in the procurement process through the government. It just takes too long for the government to buy things today and take advantage of things. And so if you can unlock and break down some of those hurdles and drive dollars more broadly into the development of these capabilities that will create tremendous outcomes for our military, that has to happen and it has to get better.
8:08 Erik Torenberg? And what are the bottlenecks for that?
8:09 Speaker 01 It's really just the purchasing process and the program process that exists today. And there's been a very big focus on how do you improve that and accelerate that. I think the great thing about it is if some of these barriers can be removed. The government will get much more value from the dollar that they spend. It'll be good for the American consumer. There will be a lot more accountability in the dollars that gets spent.
8:35 Speaker 00 Right, I mean, just to back it up a moment for people that don't live in this world. The big issue is that the entire procurement process is built on minimizing risk.
8:46 Speaker 00 Not minimizing cost, not accelerating technology, not moving fast is minimizing risk. That actually was probably the right posture seventy-five, eighty years ago, where a new technology platform could last thirty, forty, fifty years, and you just had to get it right. And if it took you three to five more years to get there, you could still be ahead. That no longer works. And so the new process needs to do two things. One, once it's deployed in the hands of a warfighter, it needs to work. It simply cannot fail at that point. But when you work upstream of that, the process has to offer a lot more failure because without failure, there's no success. Without failure, there's no speed. You try four things, three don't work, one does. You double down on that.
9:31 Speaker 00 You try three more things. And that process, which is very common to the valley and is very common to a startup and is very common even to a company like us that works in technology, is anathema to the way not the Department of War wants to operate, but it's anathema to the way the laws and the regulations are written for the Department of War to operate.
9:51 Speaker 01 And I think that there's very good lessons that can be learned from the Ukrainian conflict, where they are in a fast path of iteration, where they bring things to the battlefield, they either work or they don't work, and they iterate quickly. And that path of iteration has actually been the path that has created sustainment and success for them. And we have to learn that same agile iterative process in the US government for the US government to deliver the right capabilities to the military and not be shut out of innovation.
10:24 Speaker 00 Right. Now, to be fair, it is getting better.
10:26 Speaker 01 It is getting better.
10:27 Speaker 00 It's better now than it was five years ago. It's better than it was ten years ago.
10:31 Speaker 01 But we have a long ways to go.
10:32 Speaker 00 We have it. Well, we need to move faster because, again... This is a competition with a very well-funded adversary with frankly, a different even moral framework, not just legal frameworks.
10:44 Speaker 01 Right.
10:45 Speaker 00 And so if we don't move at speed, eventually they're going to catch up.
10:49 Speaker 01 Ted, and how worried should we be about supply chain vulnerabilities?
10:53 Speaker 00 I think that's in the news right now, right? I mean, the entire conversation about critical minerals. We do need to be worried about that. A few years ago, we were working on a direct energy program inside Booz Allen. And I said, this is great, but I want you to prove to me that it can be made completely by things that are either sourced or procured in the US. Or if it cannot be procured in the US, it needs to be procured by an ally that is steadfast and that we would not lose in a conflict. Ultimately, the answer was that could not be done.
11:31 Speaker 00 And so now, so to me, this is the kind of issue that needs to ultimately get solved. I do think we need to be worried about it again. There's very strong efforts underway right now by the administration across energy, interior, commerce, defense and so forth to try and address it. But it's right now, it's an issue.
11:57 Speaker 01 And there's a lot of attention on this issue, not just within the government, but within the companies that manufacture and how they're working their own supply chains. But it doesn't change overnight. That's a long process. And the positive thing is there's progress being made, but it's going to take some time.
12:14 Erik Torenberg? I want to zoom, you mentioned Ukraine. I want to zoom out and ask first, just how has that conflict changed our understanding of war and what it takes to be successful?
12:25 Speaker 01 Yeah, I'll start. We've got a couple dozen employees deployed there, working alongside the Ukrainians to provide high value targets against the Russians. And I can speak personally for the experience we've had. We just get smarter every day. What product capabilities we need, it forces us to find ways to fix things overnight if we need to, because it's life and death. Like if we can make a difference in the next 24 hours, we're going to do that. And so I think the number one. Thing that we've seen is the agility that has been forced on to suppliers, supporting the military and the ability to change tactics quickly.
13:09 Speaker 01 And the simplest thing that I would say is that we operate in a GPS denied, comms denied environment. And so you've got to have capabilities that exist in your products that can operate in that environment because you literally never have, you never have comms.
13:24 Speaker 00 That's exactly right. And that's something that is often lost. And, you know, if you look at where Booz Allen is playing most effectively, I'll highlight two areas. One is on at the edge. We're somewhat unbeatable at the edge, especially when it comes to communications, cyber, and, and being able to interoperate across these various technologies. And, and that is something that the Ukrainians have had to do on a shoestring. But, but we as a nation have become very good at. And, and we're learning how to iterate faster and how to, they've gotten very good at incorporating commercial technology, real time into what they're doing and iterating it once it's
14:08 Speaker 00 inside the mission as opposed to trying to do all of it before it comes in. The other area that has been fascinating to me at least is this concept of open source intelligence. You know, intelligence that can be gathered from people's cell phones, commercial satellite, social media feeds and the like. In lots of cases, we have found that the OSINT is as good, or perhaps not quite as good, as, as, as what comes in on the classified channels, but it can get there faster. And again, in a time where, where speed is of the essence, that's been a huge mind shift for a lot of people.
14:50 Speaker 01 And I think stating the obvious, but it, it plays out every single day and it played out in Israeli conflict as well, is the importance of autonomy and drones. And so this idea that You know, you're reducing the risk of life and that you're deploying these autonomous systems, whether they're low-cost attack drones or surveillance drones, whatever it might be, that is having a huge impact on the outcomes in that conflict. And you see the use by the very effective use by the Ukrainians, you also see the effective use by the Russians. But it's redefining the way conflicts have been fought in the past. And autonomy will play a critical role as we think about broader conflicts around the globe.
15:32 Erik Torenberg? Can you share more about that in terms of describing in greater detail what it used to be before this development in terms of what dictated which side would win or what was most important to get right? And now in this sort of post, now that we're entering sort of more drone-focused wars, what's most important and what changes there in terms of what's needed to win?
15:55 Speaker 01 Yeah, well, in the past, you had the systems were always human man systems, putting in human life at risk every single day for every aspect of the conflict. And today, what you see is a variety of autonomous systems or drones that are playing a very critical role. Now, it plays, it pushes more responsibility on technology for technology to work well, to make these systems operate in a way to get the right mission outcome. But it's taking the human and putting the human in a place where they're not always at risk and you're not always flying the most expensive warfighters, the most expensive
16:37 Speaker 01 systems, and you're doing it in a more cost-effective, mission-capable way.
16:43 Erik Torenberg? Is America ready for this future of warfare? Say more about our current position here and what's needed to get forward.
16:52 Speaker 00 The answer, short form, the answer is yes. I think that our military understands that we're now fighting across five domains simultaneously. It's no longer just air, land, and sea. It's also cyberspace and actual space.
17:06 Speaker 00 And they have to be able to operate in an integrated way across all of those domains. That in many cases, you have to be able to operate autonomously on all of those domains. Especially the further that you get from the human, the more the autonomy systems have to be highly accurate and highly reliable. Think about space, the latency issue alone, if somebody is shooting at your satellite, there is not enough time for the satellite to start a whole communication chain about what do I do? The satellite is going to have to react real time. And so all of these things are understood, all of these things are in train. And there's significant investment against all of them.
17:49 Speaker 00 You know, we've been very lucky at Booz Allen, we're the forefront of cyber, we're the largest provider of AI to the federal government and we're builders. So we're building a number of edge systems that either are soldier born, communication platforms, or we build an amazing technology. This is why we have this partnership, we can build on Hive Mind and the incredible things that Shield is doing to accelerate this future. Are we ready? Yes, but it's a future that is changing by the day.
18:21 Erik Torenberg? Let's actually go deeper into both your respective companies. Booz, of course, one of the biggest contractors for the government, been around over 100 years. Why don't you give a little bit of the history of how it became one of the biggest partners to the government and where it focuses now?
18:35 Speaker 00 So, you know, people get surprised when I tell the story of who we are now. So if you look at Booz Allen now, we are in the business of injecting technology into the most critical national security challenges of our country. Like I said, number one in AI, number one in cyber, and very strong set of edge technologies, just to name a few. About 75% of what we do is national security. So then the question is, how did you get there? Because certainly you weren't any of those things 100 years ago. And if I were to describe to you the story of the last 25 years really quickly, I would say we've been through three phases. The first one was to become very mission centric.
19:19 Speaker 00 We are obsessed with success in the mission. This was born out of our work in cyber early on, our counterterrorism work post 9/11, and our counter IED work in Iraq and Afghanistan. It really put us at the center of these missions. The second thing that happened is, once we were there, we realized the power of technology to transform these missions real time and make ourselves much more effective, which is why we got into AI in 2017, when people were really talking about it, into autonomy earlier than most, into C2 systems that operate at the edge and that embed all sorts of artificial intelligence into it from the waveform to the way they operate.
20:01 Speaker 00 And then the third thing has been this speed centricity. I'm obsessed with speed. I think that ultimately it'll be the faster, not the bigger, that will win. And I believe that for the country, I believe that for our company, which is why we're forming these amazing partnerships, because we don't want to reinvent the wheel. We want to work on top of the best technology out there to extend it, to help it be more successful, to integrate it with things that we're building. And that's exactly what our two companies are doing as we speak.
20:32 Erik Torenberg? And why don't you give a description of the history of Shield and how it got into autonomy and where you're focused now, and then let's get into the partnership.
20:41 Speaker 01 You bet. So Shield is a 10-year-old company. We focus on next-generation autonomous systems. We have two product lines today. We have a software product line that delivers a software development kit for autonomy developers to be able to extend the core platform to make a variety of things fly. And one of the things that has really differentiated us is our time to flight. We can get something up and running in a very short period of time. We started in air, but we are now doing work in maritime and will move cross domain with that broad platform. And then we also have an aircraft product line. Today, we offer a Group 3 drone primarily used for surveillance and targeting. And it's unique in that it's vertical takeoff and landing.
21:26 Speaker 01 It allows us to be deployed in a lot of complex situations that traditional runway-based Group 3 drones have not. And we're continuing to invest broadly in next-generation aircraft. And you'll hear more about that from us in the coming weeks as we release new products in that particular area. One of the things that's been really exciting about the relationship with Booz Allen is being able to take that software SDK that we offer and work collaboratively to deliver great mission outcomes for mutual customers. And our focus is how do we help organizations extend the platform to fly whatever they want to fly?
22:06 Speaker 01 And Booz has been an incredible partner in delivering those kinds of outcomes. And as we look around the globe, one of the things we hear in the international markets in particular is the need and desire to have sovereign autonomy. So take that core platform, but then extend it into the specific aircraft, weapon systems, et cetera, that a particular country may want. And they want to be able to do it their own way and not have it be... it just needs to be unique to them. And that's what's been a big differentiator for Shield, and it really creates this mutual opportunity between Shield and Booz Allen.
22:44 Erik Torenberg? That's really awesome. You've talked about this all-of-nation approach. Why don't you discuss what that means and how come you should get involved?
22:52 Speaker 00 I feel very strongly that we are in this unique, really very challenging, very dangerous geopolitical moment, perhaps the most challenging and dangerous of the last 50 years. With the rise of China, with what Russia is doing, Iran is now somewhat diminished, but hopefully forever, but who knows?
23:13 Erik Torenberg? For now.
23:14 Speaker 00 But for now. And North Korea is always lurking, and there's other actors around the world. And the challenge is not just that they exist, it's that they are collaborating. And you see it in the battlefield, you see it in Ukraine, you see it all over the place. And so it is... So the US has to have a response that cannot be driven solely by the government, solely by the existing defense base, solely by the valley. It's really going to take everybody. And in my mind, it operates in multiple levels, but at the core of it is these unique partnerships,
23:54 Speaker 00 where companies are looking for points of commonality to move faster, like we are, and are going to market together. And in some ways, we have both pushed the "I believe" button that says, if we do this together, we can do great things. We did not spend months and years writing business plans. The business plan, part of the business plan just happened as we were getting ready for this podcast. But this is how we need to iterate on a basis of both common purpose and some trust.
24:24 Speaker 01 Yeah, I think it's really true of these new waves of technology. As you think about physical AI and leveraging physical AI around the globe, I think it's the partnership that allows us to drive that agility and speed. And by ourselves, while we will go and tackle it, we can go so much faster with an amazing partner like Booz.
24:47 Speaker 00 So expand that to the entire ecosystem. And who else can the two of us bring into this conversation that would add a new effect, a new behavior, a new platform, a new way of operating that would accelerate both of us and give somebody else the ability to play? I think these complex partnerships, I hate to use the word ecosystem because everybody uses it, but it sort of is that.
25:12 Speaker 01 It is that, yeah.
25:13 Speaker 00 If we can get that done correctly, we bring the best of the US economy into the national security mission, which is ultimately how we want the Second World War, how we want the Cold War, and how we're going to need to win this one.
25:26 Speaker 01 Talk more about, you mentioned AI, talk more about what it takes to certify trust and safety in, as it relates to AI and lethal systems. Is the bottleneck the technology? Is it policy? Is it something else? How should we think about that?
25:40 Speaker 00 I think it's an all of the above. I mean, I think that the challenge with AI is that it is both similar and different from what preceded it. Right now, one level is an algorithm. And we've been using algorithms since, you know, think about the bow and arrow in the forest with the string and somebody walked through it and... And the arrow went that that's an algorithm, right? A landmine operates on an algorithm of a certain amount of pressure and so forth. The difference is these algorithms are dynamic and they're learning, they drift, and they're incredibly powerful. But at the end of the day, we can't create an entire new framework to regulate them,
26:24 Speaker 00 which is part of what some countries are trying to do because it's unnecessary and it'll slow us down and it'll sit on top of the existing frameworks. And at the other end, you can't just say, well, whatever, we'll figure it out later because all of these technologies do have a downside. So there is a regulatory component to all of that. But I really think the biggest challenge is to learn how to use and evolve these algorithms, again, like the Ukrainians have done, real time, in the field, in the fight, to be willing to have some level of experimentation upstream and upfront and then to prove the thing enough that you can actually launch it much faster than it's been launched.
27:08 Speaker 00 Right now, it takes too many iterations to get that done. I think we can accelerate the heck out of it. Frankly, technology itself will be part of the solution. You know, we're doing amazing things with digital twins that allow you to both train the algorithm a lot faster. This is what we're doing with Nvidia, but also allows you to demonstrate the validity and the power of the algorithm a lot faster. We can get there. It's not rocket science, but it's going to take a concerted effort.
27:40 Speaker 01 Yeah, I think as a supplier in this world, the one thing that we think a lot about that we think is really important is ensuring that we're doing everything we can from a cybersecurity point of view to give our customers the confidence that there will not be tampering, there will not be some form of intrusion. And we think Very holistically about that and we own and are responsible for ensuring that what we deliver has the right security controls that are necessary to give our end customers the confidence that we're tamper-proof.
28:19 Speaker 00 That's an excellent point.
28:20 Speaker 01 Super important.
28:21 Speaker 00 One of the areas where we're collaborating is Booz Allen is the leader because of our cyber business in adversarial AI, which is both how would you attack an algorithm and how would you defend it? And we're bringing that knowledge into this set of missions because obviously the most powerful thing you could do is mess with somebody's algorithm and have the opportunity to affect their autonomy at the right time. And SHIELD is at the leading edge of understanding that and the leading edge of actually implementing the right practices. I think you didn't ask this question, but the biggest challenge, to me, the biggest risk that we have to this whole thing is an event that pushes everything back, either
29:05 Speaker 00 a serious safety event, a serious cyber event, or something like that. And so this is the balance between it's gotta work.
29:13 Speaker 01 Right.
29:14 Speaker 00 But you can't wait until everything works perfect.
29:16 Speaker 01 No, and I think the way to think about this, different than the military has thought about it in the past, is you need to think about it like traditional commercial software. So think about updating a fleet of autonomous jets with a new software update and how can you deliver confidently from a security and resilience point of view that that's going to work flawlessly? And you think historically, you've built a very big, complex, custom stack of software that's different than everybody else's. No, you want to be able to leverage the practices of commercial software to drive updates, but to do that, you gotta ensure that there's no cyber issues.
30:00 Speaker 00 Right.
30:01 Erik Torenberg? You know, we spent time talking about China. We've talked about Ukraine and Russia. You briefly mentioned sort of Iran doesn't seem like as a big of a threat for now. Why don't you talk about what's happening right now in the Middle East in terms of its significance? Or how we should think about sort of that region going forward as we sort of evaluate, you know, potential short, medium term threats?
30:23 Speaker 00 I can start. I mean, I think what's happening right now has the potential for being historic in terms of the expansion of the Abraham Accords and creating a different model for cooperation in the Middle East. And I think that I'm hugely optimistic about that. I think none of us are also assuming that this will be perfectly peaceful forever. You have a number of additional dynamics like the Shiite Sunni dynamic and, you know, some of the tribal issues in and between some of those nations. So the Middle East is going to continue to be an area with that is important to participate
31:05 Speaker 00 in heavily, to look out for and to build alliances and partnerships, many of which are based again on the deployment of the right technologies. And I won't speak for Gary, but I think one of the realities is that you can invest more if the customer base is bigger, if your TAM is bigger. And having the ability to seamlessly trust a set of allies and partners that you can sell technology to at the speed at which is being developed, which again, we talk about the acquisition processes, but the export control processes are just as slow. I think that is, that is part of the future.
31:46 Speaker 00 And again, from our perspective, our program, Thunderdome, which is essentially the essential zero trust program for, for all military networks and going more broadly is a big part of it because one of the challenges in lots of these places in terms of sharing technology. Is you have to trust that their networks are now so compromised. That the moment you give something to an ally, you simultaneously gave it to all of your adversaries. And that's a big part of the work that we're doing.
32:21 Speaker 01 And I think if you look at what's happening in the Middle East from a positive point of view, there's very heavy adoption of AI across the Middle East and a willingness and desire to take a lot of that advanced technology and apply it to their defense. And so I think there is an opportunity as things begin to stabilize there to engage more directly to get the right aligned outcomes across the Middle East. And I think AI can be frankly a conduit from an innovation standpoint to deliver different outcomes and joint funding for how these outcomes can come about. So I'm I'm actually reasonably optimistic. Things change all the time and I think we need to be thoughtful about that. It's also very difficult for the US government to figure out
33:06 Speaker 01 one day you're going to get an export approval or not. That's always a challenge. But I think the interest in AI, the willingness to fund it and step up from a dollars perspective, I think they can be very interesting from a partnership perspective with the US in funding some of the things that are critical for us.
33:25 Speaker 00 For sure, and we need them as allies and partners on what's really a global geopolitical situation where the Middle East will play a role.
33:33 Speaker 01 That makes sense. I want to go back to procurement reform. You mentioned it's been getting better, you know, the last five years, the last 10 years. What is the mechanism by which it gets better? And I asked to say, you know, what needs to happen for it to to get better again? And we said it needs to get better faster. Or we have a long way to go. For people who are not in the weeds of how this works, why don't you give a, you know, take us behind the curtain a bit?
33:58 Speaker 00 You know, there's two aspects to this. It's one, as Gary is saying, is cultural and the practices and the way people think about. What they want to do and whether they have the authority to do it in the way they want to do it. And then the second part of it is, is the regulatory framework that constrains those authorities. And what's been getting better is I think people are exercising their authorities with a little more alacrity and with more vision and more foresight. So even without major regulatory change, we have seen over the last few years greater desire to buy, not just buy commercial products,
34:42 Speaker 00 but buy commercial products the way a commercial buyer buys them. That's right. And that's been the big breakthrough because you can take a commercial product and layer on top of it all of the traditional government requirements and it stops being a commercial product and it slows down just as much as something custom would slow down. And that has really changed over the last five years, last 10 years for the better. The piece that is left is the regulatory framework is still 75 to 100 years old. There's an overreliance, for example, on the idea of cost plus. And cost plus does have a role probably. I don't know that you can buy a nuclear submarine any other way,
35:27 Speaker 00 but it's a much more limited role than the way it's been driven now. I think the understanding that outcome based is the way of the future that you, you know, the government should pay, whether it's in military or in health or in any area for an outcome as opposed to for a set of inputs is both fundamental and transformative. But then the second thing is this is regulation on top of regulation on top of regulation. And more things keep getting added and nothing is taken away. And so there's an effort right now, which is highly promising at moving away from these regulatory frameworks and almost started with a blank sheet. Because you almost can't fix it from where it is.
36:12 Speaker 00 It's not about can you reduce this thing by 10% or 15%. It's the only way you can reduce it by 80% while keeping the stuff that is really important is to start fresh. And there's a couple of efforts, one in the executive branch and one in Congress as we speak, that aim to do just that. And if any of that comes to fruition, I think that would be the key to unlocking the next level of speed.
36:36 Speaker 01 And you're starting to see some efforts within the current administration, under the Department of War, where it's really rapid adoption of technology. And it is a path through a faster path of procurement, but then putting pressure on the supplier like a shield to deliver in a particular timeframe. And so there is this positive movement in the thinking within Department of War about how can you drive a faster process for new technologies that are going to fundamentally move the needle? And while that's not a broad reform of all government procurement, it's green shoots and changing the way defense technology has been procured in the past. I think it's a really good thing.
37:18 Speaker 00 Yeah, ultimately, companies like ours, whether alone or in partnership, want to get paid for an outcome. Give me a tool, whether it's an autonomous drone, or a communication system, or a cyber platform, or whatever it is, that accomplishes this in this way at this cost and this timeframe. And you figure out how to do it. As opposed to, we're going to tell you exactly how to do it, we're going to monitor every single aspect of it, and we're going to test you every five minutes. And if you can't deliver the outcome, you're going to go out of business. And if you can, you're going to grow. And I think that to people that live in the commercial world, where I spent a lot of my career, that's obvious.
38:03 Speaker 00 If you operate in this context, again, for good and valid historical reasons, it just operates differently. And so while people are being held accountable, they're not being held accountable to the outcome, they're being held accountable to a set of interim steps.
38:18 Speaker 01 That's a good description. I think that's a good place to wrap. Gary Horacio, thank you so much for coming to the podcast. Thank you so much. Appreciate it. Nice meeting you.
38:25 Speaker 00 Take care.
38:28 Speaker 02 Thanks for listening to this episode of the A16Z Podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts, and Spotify. Follow us on X at A16Z and subscribe to our Substack at A16Z.substack.com. Thanks again for listening, and I'll see you in the next episode. As a reminder, the content here is for informational purposes only, should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast.
39:12 Speaker 02 For more details, including a link to our investments, please see A16Z.com/disclosures.
Previous Marc Andreessen and Ben Horowitz on the State of AI Next Sacks, Andreessen & Horowitz: How America Wins the AI Race Against China Page updated October 26, 2026 · transcript and summary are machine-generated
