Name: Defense Disrupted
Uploaded: 2025-12-10
Duration: 17 min 31 s

Autonomous drones and pilots are flying places humans cannot go, while national security increasingly depends on data as much as force. Two prominent founders of defense tech unicorns discuss how Silicon Valley and the Pentagon are rewriting rules in a rapidly evolving (and unstable) geopolitical landscape. Where will AI redefine deterrence, surveillance, and battlefield autonomy fastest? Where is tech making progress on the challenges that the U.S. defense establishment faces––and what remains elusive?

Defense Disrupted

Conferences

Hello again. Tara, Gary, thank you so much for being here. It's great to be here. Thanks for having us. So let's start really big picture, um, what are the biggest challenges of selling to the largest bureaucracy in the world? I don't think we have enough time for that question, Ali. Uh, I'm happy to start, although, Gary, I bet you have plenty of war. I do it, you can do it. I believe, um, it's speed and it is the restraints that are placed on federal government buyers and the ability to take great ideas, smart people who work in government who are trying to do good things and. Help them break through the bureaucracy and access funding. You know, we spend at least 6 months out of every year under a continuing resolution, meaning that year's budget hasn't even been granted yet, so it really compresses the ability of time to be able to drive this change and get that deal closed. Yeah, and I think about it internationally. I don't think about just the US government. I think about global governments and how we can better serve warfighters around the world. And the one thing that's really obvious and the thing that gives us passion every day is like how do we put better tools in the hands of warfighters and really make a difference. And the good thing is I think there's an understanding that it has to get better, and I think there's a lot of well-intended motivated people to try to make it better, but. As Tara described, like we're at the beginning of that journey to see radical improvements, and those radical improvements are absolutely required. And you think about everything you guys have heard here, the rate of change that's happening, which is super exciting, you need that rate of change to actually land in governments around the world, and they need to see the positive effects of that technology change. We're actually just coming off of one of the biggest defense conferences of the year where there was a lot of messaging around this. The Pentagon is saying things like it's time to move to wartime footing, for example, and really accelerating rates of adoption. How much of that is real and how much of it is theater? I'll start, um. I think there's very good aspiration to do something fundamentally different, but we're, it hasn't flowed all the way down into the bureaucracy. The one place that I would say it's happening every single day is in the Ukraine. So at SHIELD we have employees deployed in the Ukraine working alongside the Ukrainians leveraging our technology to get better outcomes against the Russians. And what we see there is this total agility where something works one day, you get feedback, engineers work overnight, we get them fixed, and it's back into the battlefield. And I think that agility, I think everyone's watching it around the globe. Everyone's watching really closely what's going on in Ukraine, and it's that agility and ability has constantly changed. It has really been one of the most important aspects of Ukrainian defense, and I think we as a nation, the United States, need to be better at that. Agility because that's the way the next conflict is going to get fought. I think building on those great points, a big part of what Secretary Hegseth's message is and what the US national security apparatus is trying to do from a move to wartime footing push is to really catalyze the American industrial base and drive production because much like we've seen in Ukraine, the reality of A future war that the United States is likely to enter is going to deplete stockpiles that are already anemic today because of what we are supplying to allies and partners around the world and rightly so because it protects US national security interests as well. Unfortunately those stockpiles are low enough and the United States has outsourced so much manufacturing capacity. At this point that the amount of time it will take to build the munitions and weapons systems that the United States needs is just much, much too long. That's why we have to get on wartime footing now, especially through the lens of the industrial. How long is too long? What are we talking about here? I mean, a lot of war games will show that in a notional war with China, the United States is out of munitions in 7 days. Whether you think 7 days is realistic or you think well maybe the beginning of the war will play out differently, you're still operating in things like days, weeks, or months, and the reality is when you look at Chinese war doctrine and the Chinese Communist Party's. Significant amount of time and attention spent studying, going to war with the United States. They are planning for a very protracted conflict and would be happy to draw that fight out to bleed American stockpiles dry because they aren't missing the economic piece of this puzzle. I'd be curious from the audience who here would say they're sort of familiar with the state of defense tech. Raise your hand. That's not that many actually, um, so I think it's maybe important to say, um, you know, I'll start with you Gary since you brought up Ukraine. I know SHIELD has been deployed in Ukraine had technology has been deployed in Ukraine. Um, what have you learned on the ground there and how much did the war in Ukraine change the state of defense tech? I think that the conflict there informed the way conflicts can be fought. There's a lot to be learned, but there's, I think, a conflict with the US would be different. But one, it's dominated by autonomous systems, whether it's one-way attack drones or more sophisticated autonomous systems. You're just not going to put. Warfighters at risk if you don't have to, and both the Ukrainians have learned that the Russians have learned that the speed of change is incredibly high because certain tactics work and then you have to regroup quickly. So one of the very simple examples is GPS is basically denied 100% of the time and comms are denied, meaning that if whatever weapons system or aircraft you had that was dependent upon comms or GPS. It's just not going to work in that conflict and so you learn and that could be life or death and that could be life or death, and it could be life or death for sure. And so being able to respond to the actual conditions on the battlefield have been absolutely critical and I think through those experiences for people like Shield, if you think about it from a product development perspective, you get a whole lot smarter about how do you drive. The innovation cycle and the iteration cycle that's absolutely necessary to ensure that you're delivering something to deliver mission outcomes and so I think Ukraine has really. Shown that you can be agile, that you can use lower cost kinds of lower cost kinds of systems, and that you can iterate quickly and get real outcomes, and I think that's that's so important in the world that we're living in today that is under so much change. And Tara, how has Ukraine affected defense tech from your perspective and particularly how you think about readiness? Well, I think that it has shown the urgency that we all feel in sort of an abstract way in terms of just how quickly those learning cycles have to happen. So today the. The United States is in a readiness crisis. Our weapons systems and military platforms have historically low operational availability, which basically means if we need to go to war, half the fleets sitting in depot or at dock. In what we've seen in Ukraine is that adaptation is probably the most important skill and so what we are doing as a company at Gvini is using data and software and AI in order to try and bend that curve on readiness. And I think to Gary's point, it's both making sure that you have the right set of systems, many attritable lower cost systems, and addressing that cost curve. Today in the United States we'll use a $20 million missile to shoot down a $2 million drone overseas. That doesn't work and it shouldn't work for US taxpayers. So how do we increase military readiness and also start to think about what the conditions of the future of warfare will look like and ideally this we sort of take for granted in the United States, but ideally do it in a way that this country can afford for the long term. And when you think about AI since we're at Brainstorm AI, um, how fast is AI being adopted in the military right now? Over the past, I would say 12 months, much, much, much faster than we've ever seen before. What's spurring that on? I think a lot of it is a combination of a maturity of technology and frankly a presidential administration that wants to work with tech and once you open the doors to tech industry and of course credit to the tech companies. For saying we want to support the defense mission and US national security, that's what happened first is the the company showed up, then the investors moved in and said, OK, we're in this too, we'll give you the capital to be able to develop your products with a national security or military application. And for a while the Pentagon was still like, yeah, we're not sure what to do with you, and we don't know if we can trust you. Aren't half of you gonna go out of business and, uh, and what we've seen over the real side eye there, yeah, exactly, and the reality is actually 95% of companies will go out of business, but the 5% are gonna change the world, and I think that's. Exactly the appetite that this administration has, which is let's find the 5% that are gonna drive this significant change and uh try and reduce risk in acquisition. I'm sorry, accept more risk in acquisition in order to reduce risk operationally. What about you? And, and I agree with Tara in that. It feels a lot better, but we are still, we have so much more to be done because there is so much opportunity, and I, I look, one of the things I'm most excited about is. The fact that AI will fundamentally transform the aerospace and defense industry, it, it's going to look different in 20 years than just radically different than it does today. You're going to have lower cost systems, AI led, software-led, um, not these super expensive. Incredibly elaborate systems that just get shut down and so I, I think there's a revolution happening and we're at the very beginning of that journey, um, as Sarah said I think the administration is really open to this, but we're, we're 9 months into that so a lot more to, a lot more to come on that front and we've got to move faster. We're going to open the floor to questions, um, but I am curious, uh, you know, the Pentagon is famous for moving slow. Silicon Valley is famous for the phrase move fast and break things. What is the biggest cultural clash between tech. And the Defense Department. We get to wear jeans. No, I, I think, I think, um, risk tolerance is the biggest clash and um. I mean, I'll use a company that isn't here on stage, but you know I certainly think highly of as an example. Look at the recent coverage in the Wall Street Journal of ARR and highlighting the failures of that company as sort of a way to put a wet blanket, what I would characterize as a way to put a wet blanket on what's happening in defense tech and what I think that story actually. Highlights is the reality that in technology when you are starting to really do innovative things that are brand new you're gonna see a lot of failures and that's totally fine so long as you're learning from them rapidly and adapting your own businesses and making products better and in the national security sphere especially around the US military. Risk tolerance is exceptionally low and frankly for quite good reasons you mentioned it earlier, Ali. This, this is all life and death and so I think both sides are genuinely trying to figure out what's the right balance of that when the money we're throwing at this innovation and experimentation is taxpayer money, so we have to be responsible about it and these are. Serious systems with very serious outcomes, we have to be able to rely on their lethality on the battlefield. And yet if you only ever play in the safe space and you don't take some risks and you don't embrace some of that failure, we're going to continue falling farther behind. And so that's the biggest clash that I see. Yeah, and I can give you a slightly different perspective, but building on that. So one of the things I'm most excited about, so at SHIELD we're building a next generation autonomous warfighter. So what is that? That's what does that mean? That's a plane that takes up vertically, lands vertically, flies long distance, 2000 nautical miles. OK, carry's munitions, like they would be the same. Same that you would see roughly on an F-35, 1/10 the cost, lifetime cost of an F-35. But what's exciting about it is if you think about the people that had to build the F-35, their number one thing is about not killing the pilot, right? Making sure that in all circumstances you're not going to kill the pilot. In our world, we're building an autonomous aircraft. We can go hard and crash. And try things that you wouldn't be able to try if you're engineering with a human pilot because you're going to be safety is going to be your number one priority in this world where it is autonomous, you can cycle fast. So think about the agility mindset in engineering and being able to try things, experiment, and, and iterate fast. What's also different about this next generation of young companies. Is we're raising private capital so it's not on the, it's not on the taxpayers' back that we crashed one day. It's, we have to be, we have to be prudent to our investors we have to be prudent and thoughtful to our investors, but that speed of iteration is what ultimately creates value and shareholder value for investors and you can be aligned on that, but it's very different if you're using government money and poor tax papers watching some auger into the ground. But I love this. Fast agility, test fast, go fast, and drive innovation that looks fundamentally different than what we've seen because you're not focused on, oh my God, we've got to protect the pilot. Do we have a question from the floor? We have time for exactly one. If not, I have one final question. Yes, please, name and company, and you're bringing us home. Hi, my name is Margaret Tong, and I'm uh building an AI agent to automate planning for aviation programs. So as someone from aviation background, I'm just curious about, um, like you mentioned, safety and risk is a very big um concern in our industries. So like, what kind of regulation, um, restrictions or challenges are you facing right now? And do you see that, um, As something that is like overwhelming you, like you said, like it's pulling us back from like letting AI to do more things for us, or actually it's something that we do need those safety guardrails to make sure that, for example, in terms of like cybersecurities, when we get hacked or electricity just gets stopped, like hacked by foreign countries, like um how do we make sure the safety is still in place when we have AI to do a lot for us? Sure, I think. So as we think about the broad profile of delivering final autonomous aircraft, we have to think through all aspects of safety, um, including cyber so what is that? Software development cycle, how do we ensure that we haven't, that we're tamperproof, that there hasn't been something that has, um, taken over the code, and one of the things that people think about is, oh, somebody's going to press a button and the drones will fall from the sky. It doesn't really work like that, um, and so, but that is a concern, and I think for us the burden is on us to demonstrate that we have safety throughout the process, whether it be cyber safety or flight safety. But we use a very strong path of iteration to test and ensure that we can deliver on those qualities and then we certify against a set of um kind of classic certifications that are absolutely required to give our customers comfort that we can deliver on those values so it's a very high priority it registers. Um, as probably the most important thing we do, but we don't operate with the mindset of killing human pilot, which has held back a lot of the speed that has traditionally been needed in this industry. We're out of time, but one word for what the future of war looks like, Gary. Autonomous Chaotic. Thank you guys so much.