Dave DeWalt, NightDragon | theCUBE + NYSE Wired: Defense Tech

Palo Alto studio connection Silicon Valley and Wall Street. I'm John F co here with Dave Volante my co-host. Hi, I'm John Furry, your host of the cube. We are here at the cub's NYC studio. Of course, we have our Peloto studio connecting Silicon Valley to Wall Street. This is our new defense tech series. We talk to the leaders who are exploring all the new business models, technology, innovation. Of course, the tide has shifted with AI as the battlefield and decisions are being made on the warfare front. Everything's changed. Dave Dewalt is here. He's the CEO and founder of Night Dragon, investor, uh, former entrepreneur, CEO, now driving and innovating. Dave, great to see you. And you're a leader in the field and doing a lot of great work. The market is on fire right now. Literally on fire with investment um, crossover private technology into the government realm for defense. We all know what's going on in the physical hardware world. We see AI factories. So everything's driving to the conversation where cyber security and AI infrastructure are all one thing. The music's the same, but the applications on the defense tech are booming. Explain the macro right now. What's the biggest change since we spoke last year? >> Well, thanks for having me, John. It's great pleasure to be with you, my friend. And uh good topic right now. It's very it's very gerine to what's happening around the world for for a lot of reasons. I call it almost like a perfect storm. And this perfect storm is a series of like conditions and confluences coming together and we can all read about them. When you really break them down, we've watched this incredible technology inertia, you know, malade with frontier models and AI's development, agentic autonomous systems, AI on the edge, really opening up new opportunities in the defense tech, but also new opportunities for the offense as well. And then when you combine geopolitical tensions with all the things we've seen, we start to create crisises that are happening with Iran and also with Israel, of course, in the United States, but Russia, Ukraine, and other parts of the South China Sea. >> And it's a very challenging market. And then you create almost a deregulated environment now where you can bring technology very quickly to market without a lot of guardrails which is great for capitalism but at times can also be bad for the bad guys or good for the bad guys. And we create this environment where we've got a boom in defense tech. And this is probably the golden age I've ever seen with what's going on in this market. 35 years doing this. >> Yeah. I don't know, maybe 1999.com and this is about the same era to me. >> You know, uh I know you just came back from black hat. Um and I want to get your thoughts on this because you mentioned the bad guys. You know, my two favorite football teams are as New England Patriots and the San Francisco 49ers. Um and the Niners, as you know, had the West Coast offense back in the 80s and they they they pioneer. >> Are you allowed to root for two teams? >> Well, they're different conferences. >> Okay. [laughter] Okay. So, I'm okay. I adjust the rationale. Um but, you know, look at Joe Montana. They had, you know, Bill Walsh, they actually pioneered score more points than the other guy. You win and you had the old thesis of defense wins championships. And obviously when Brady was at at the Patriots, they would throw up mega numbers. So it was really an offensive game and not a lot of defense. Now you're seeing that come back. Compare that to cyber because we're seeing now more offensive capabilities with AI. The bad guys get better. They're putting more points on the board. There's longtail opportunities. You're seeing it every day. Okay. So, there's a lot more offense going on and we've always been playing defense and talk about this dynamic. So, this kind of came out from my reading the tea leaves at black hat. Great. More offense is great to protect to beat on offense, but they have offense, >> right? >> How do you explain this phenomenon or do you believe it to be kind of a good maybe it's not the perfect analogy, but it's only I can think of. >> Yeah. No, it's a very good analogy. And by the way, I'm a Philadelphia Eagles fan. And of course, the Patriots are kind of our arch rival because we've lost to him a few times [laughter] and that guy Tom Brady, I don't know, but who's a great guy by the way. But, uh, they always say there's two things you can never leave, your mama and your football team. So, that's where I'm from, Philadelphia. But coming to your question, offense and defense, it's an amazing shift of what's just happened. I mean, I was CEO of Firey and Mandant and McAfee and I spent a lot of years defending, defend, defend, defend. You never cross a line on the defense towards what you now think of could be offense. But what you're now starting to see is platforms that can be developed, one in which you know called Horizon 3.AI who's built this incredible autonomous agentic platform who can now do dual use as well. Not only can it be one of the strongest defenders because now at machine speed you can defend and you can constantly monitor attack surface, pentest that attack surface, heal the attack surface, but you can all use also use that technology for offense. So not only you seeing the private sector producing better and better defenders, that same defender can be in offense. And then you throw the other side of the coin which is entropics and open AI's models. You know, initially starting with Opus and now and then Mythos and Fable 5 and Daybreak from OpenAI, you now have this incredible model that can produce vulnerabilities and understand and ingest source code or binaries and produce vulnerabilities. And not only can they produce a vulnerability in a binary, they can create multiple chaining of vulnerabilities creating zeroday exploits at machine speed. Yeah. >> So now you have the attackers having models of AI that can essentially do things that humans took months and maybe even years to cultivate an offensive cyber weapon. They're doing them in minutes, hours. So you have this incredible offense and defensive problem. And last thing I'd say, I never thought in all my years, almost three decades in cyber security, would I see a scenario where an American company is doing a cyber exercise, in this case, open AI, that ended up breaching a French company, Hugging Face, that then defended itself with a Chinese company, >> an openweight open source company. Where would this ever happen? A Chinese open-source company coming to the rescue of an American company was attacking a French company. But this is all the age of aentic. That wasn't humans behind it. That was largely technology behind it. So, >> what an error we're in. >> Yeah. >> I mean, it opens up so many questions and I this is why I brought it up because if you look at go back 17 years ago when we started the cube in Silicon Angle, you know, the big conversation there was if you remember the perimeter's dead because back then you were perimeter based uh security uh and then you had the cloud. the surface area is all over the place and then now they rationalize that and they built platforms and tools best of breed but now the surface area if you look at what you just said it's not only bigger it's everywhere you can inject and change the surface area by injecting a binary so now you have a completely different situational uh awareness problem how do you think about that how what and how are successful entrepreneurs kind of putting that filter in and how are they thinking because you got to think differently it's It's a systems game. >> Mhm. >> Systems agents working across apps and platforms. >> It seems like it seems to be much more of a operating system thinking absolutely than just hey my platform here's an API and an MCP server, >> right? I mean, John, you're hitting it right on, right? We have this era of AI versus AI now, right? There's no doubt about it. You know, what can the good guys do to protect AI and create new opportunities in AI and what can the bad guys use using AI? And it's a battle and it's created almost an entirely new attack surface for companies to defend against to your point. So how do I protect against a rogue agent? I programmed the agent to do a certain thing, but the agent became smarter and smarter and escaped the sandbox and breached another company. So how do we start thinking about the attack surface of that? And what's really interesting in the age of AI, which I think you'll find interesting as well, we're watching the entire world flatten. It used to be United States and Israel, give Israel a lot of credit, especially in cyber and defense tech and America. You're watching this almost balkcanization occur of NATO countries, Asian countries who are now being forced to innovate because largely many of them were forced to contribute more NATO defense. So we went from 1% to 2 and a.5% to 5%. When you see a 500% increase in defense spending country by country, >> yeah, >> this is this is mammoth. So it creates innovation, creates new things that are occurring and now the world's innovating in a way that we've never seen. >> That's why I love the AI infrastructure converges with this because all the things you mentioned things are popping in my head. Sovereignty check that's a cloud AI problem. Um other factors that are more domain specific is cyber and this is where I think it crosses over. So the question for you is now is what's your view on the advances in some of these areas that were I won't say old school but previous categories were mechanisms that people would deploy could be hardware software electronics precisiong guided capabilities those were kind of things that were out there what changes have come down those the pike on these areas what's the what are some of the innovations I was talking with Hawkeye 360 your company I mean they're doing something very cool just around RF in space. But what that enables is a whole another intelligence level, >> right? >> That changes someone else on land, right? >> But they're not in the land business. They're in space and maritime right now. But they're collecting critical RF information. That's >> right. Like this changes. What are the changes in those areas? >> Well, let's talk about it. I mean, if you look at the history of mankind for a second, whenever it discovered a new domain, we fought wars over it, right? We discovered lands. Armies would fight. You discovered oceans. They would have naval battles. You discovered the air. We'd have battles in the air. We discovered space. We're having raced the space. We discovered cyber. It became a domain. The fifth domain. We had battles in cyber. We continually do. What's happened now is every domain now has cyber in it. So instead of cyber being its own digital domain with the internet, you now have cyber everywhere. maritime, air, ground, space and the wars are being fought in hybrid ways now. So when Venezuela occurred as an example, what happened? A lot of cyber activities occurred precursor to kinetic activities. So it's an integrated system. The government of the United States is forcing a multi-dommain cyber kind of initiative. So we're watching a lot of change across >> multi- domain. I love you brought that up because I was going to ask it, but I'll I'll come I'll get to it now. What what does that mean? is I like this idea that if cyber is native through all the domains makes total sense what other horizontal factors need to be in place and what's specific to that so cyber makes sense and how does that change cyber and and and whatnot so yeah so so you now think about to your question of hardware and software and capability now it has to be wrapped in almost every form factor for almost every domain so how do I protect a tank or a jeep or a truck or a car from a site cyber attack. And not only is that cyber attack potentially like a a spear fishing or an intrusion digitally, it could be an intrusion through radio frequencies, electronic emissions kind of capabilities. Airwaves, things like air drops and Bluetooth become vectors of attack for these physical systems. We now have to protect every way we go. So you mentioned Hawkeye 360 looking at radio frequencies in space to ground picking up RF and patterns of life of every RF is coming off Earth. What is the anomalous behavior of a GPS jammer over here, a hypersonic over here, a car driving on the freeway over here, a boat or a ship on the ocean? And you learn about patterns of life, but you also learn that there's risk to all of that. And we're now entering an era where there's cyber warfare in space, on ground, on air, and it's just all becoming. >> It's interesting. You know, we all watch movies or we have our favorite movies around some military backdrop. There's always a scene in whatever movie where they say, "Let's get a satellite image of what's going on in that camp or oh, they got guns and they're deploying stuff." So pictures were basically the the visual, but you can't see what's going on in the bunker. Right. >> So this is kind of where this multi-dommain tech multi-tech approach comes in, I think. Yeah. to me jumps out about about Hawkeye is that you can it's better than a picture, >> but you've got the picture and the other sensory sensors kind of this is kind of where it goes. Again, this is a mashup >> of technologies. Have you seen this before? Is this net new for us as thinking? Because it makes sense. Why why weren't we doing that before? This makes total sense. Why wouldn't you have the most complete intelligence picture possible? >> My favorite movie was War Games. Remember that movie? Like way back in the day. And suddenly they they are playing tic-tac-toe to end the game and like the computer had taken over and like we're in that era now. It's or Terminator when Skynet was going on and and I say that tongue and cheek but at the same time the power of these systems are getting so great that we have to put guard rails. We have to be more careful with what we're doing. We don't want to slow down commerce and capitalism. But the power of this tech is becoming so great and the weapons that are being able to create from that we really need to protect. The idea of a weapon of mass destruction was really thought about atomically or nuclear before. Now you now have AI models training on DNA and RNA to produce new viruses and pathogens. I mean that could create >> that's a vector >> a vector of attack. Yeah. Of course. >> Right. So we're watching all new ones emerging just like happened during co We now have the ability with AI to create new viruses, maybe even more lethal. What do we have to do to get that right, defend against that properly, and let defense get ahead of the offense? And this is a point in time for mankind in my opinion where we have to just take a pause, make sure we get defense ahead of offense, get the right regulation on it, and make sure that the gap doesn't get too wide too fast and in the hands of bad people too quickly. >> You know, I love the I love when I do these cube interviews when there's always a new technology, whether it's a disruptive enabler or an accelerant enabler. There's always the good side and the bad side. And you know in military it was always the generals had this like there was a room of people that were new military. They would have their caucuses. They would ga convene and gather and and do all that thinking. But it's kind of been democratized. So what is the current narrative and state of the the union if you will in the defense sector around government experts that do this for a living and uh commercial experts like yourself and entrepreneurs. What is the joint narrative? Because we see what SpaceX did. they they beat NASA at everything that they did and you know they helps everyone gets leveled up but what's the narrative of this new ecosystem of participants because it is a collective intelligence at this point as you mentioned things got to be horizontal they got to crossplatforms >> well you really hit on John because I'm I'm really proud of this administration for a number of things and there's a lot we could talk about there but one thing they've really been able to do is a infuse cyber to all multi-dommains B partner with private sector in a way that can create accelerance of innovation to the war fighter or to the mission that's needed. We used to have a very challenging and difficult procurement engine that was really an ogalopoly of defense primes. We've kind of broken that open now to allow new neos to come along, new technology to come along. >> They've also created a lot of innovation capital. Yeah. So that like the defense innovation unit or DARPA can really create innovation early in the life cycle. So America's getting its mojo in a way that we have never seen before and that's a really amazing thing in my opinion. >> I want to get to the battlefield. In fact, um I was sharing to Sol at Horizon earlier today that around 2017 or I forget what year 2017 or 2018, I participated in a Chattam House rules thing that General Mattis put on >> and this was right around the Jedi contract that Amazon was working on, you know, Tactile Edge and it was all a lot of >> DoD stuff. And the number one thing he said was friendly fire is one of the highest forms of casualties. not having the right systems, who are you shooting at? And again, at that time, but when you factor in drone swarms, machines, there's kind of a friendly fire problem, meaning we could we we could hurt ourselves if we don't know what we're doing because if there's drones and there's agents out there, does that come up or am I a little bit reading it too early on this? But like I'd imagine there'd have to be some sort of software intelligence around the congestion problem, the airspace. I mean, drones alone could take down a helicopter. I mean, I could imagine there's a friendly fire dynamic, not with humans involved, but machines. And of course, the side effect impacts could be systems could come down. What's your thoughts on that? >> Like we talked about multi-dommain integration of digital and cyber and all of the domains beginning to have to work together and coordinate. Space tips and cues aerial, aerial tips and cues to maritime, maritime tips and cues to ground or to cyber. We need multi-dommain intelligence and communications integrated to all of them and securely so we don't get friendly fire problems. We don't have issues like we have sometimes kinetically on the battlefield. And that technology is here now. I'm sure you know Palunteer uh the Maven smart system has created a hell of an integration platform. There's other startups doing similar things to enable us to ingest. We invested in a cool company called Armada.ai AI who essentially takes compute and moves it to the point of need for the war fighter for the mission and it's like a suitcase size data center all the way to a truck size container data center and you can move it out in the desert on an oil rig or wherever you might need it and now you have massive compute for data ingestion and communications that you have to go to satellite or something else to do to really create more effect. So they light up essentially the theater where they might be constrained on bandwidth. I mean you can have a Tesla running around in a Whimo in San Francisco but you can't be doing that in Ukraine, >> right? >> Because you need power and you need compute. So if you can create small modular reactors, SMRs to the point of need and you can put ruggedized compute at the point of need now you can start to use things like electronic warfare, electronic magnetic pulsing. You can start to use lasers. you can start to use non-kinetic weaponry that's safer as well to bring down kinetic threats that might be occurring. >> Let's let's talk about the battlefield and decision-m because you mentioned the Terminator um and Skynet which everyone kind of uses a metaphor in a bad way but you know it was technically a Marvel if you think about the sci-fi of it decision making at the commander level and at the at the edge to the to the war fighter. How does AI impact that? How do you see that playing out? because that's an area you're going to start to see a lot more. We hear about decision loops. You that's an AI term, looping, you know, recursing through uh you know prefill, decode, all that tokenization stuff that we we talk about with Nvidia. That's going to be a key thing on the battlefield where you have to have essentially rack scale clusters available decision-m the humans in the loop, the agents, the machine speed. How does that play out in your mind? Well, we're starting to see scenarios, John, where autonomy is beginning to be a multiple of humans, right? So, think about this. At every domain that we had talked about on ground, we're starting to see soldiers and other types of ground equipment be completely autonomous. So, if we want to be able to move a truck or a tank or a jeep or something into the battlefield, we can move autonomous vehicles into the field, having humans in the background, saving lives. We're already seeing that in Ukraine battlefield today with aerial drones where you're using FPVs or fiber optic cabling on drones that are essentially able to fly and be able to eradicate a target and there's no human involved necessarily with that defense. That's happened in maritime now where we see drones. We have a company called Seronic doing that where they build ships that are completely autonomous to essentially rescue Apache helicopter pilot that was in the straight of Hormuz or now more and more monitor the coast guard or deliver goods across the ocean. So so many things are moving autonomous and in that you have to be careful because the autonomy is programmed. >> Yeah. And that can be hacked. And if that gets hacked, what happens? So we have so many new levels of defense and offense. >> Again, back to my earlier question around the surface area. It's not only a surface area, but it's below the surface. There's code, right? >> So you have all kinds of code involved. All right. Well, this is first of all, thank you so much for sharing that insight. It's really primary data on what's happening. And >> are they more scared or less scared? I'm not sure. >> Well, no, it's I mean, you know, it's they're hard problems and engineers love to solve hard problems. So to me, it's like the more you're transparent about it, the more it'll work. And there's a sovereignty angle too which I love. Um what is your focus now? Obviously you're doing a lot of work and really appreciate what you're doing and certainly at the government level bringing you know the ideas and the leadership there but investments um you know you mentioned a few of them. Is there a thesis you got right now? Is there is there a certain um pattern that you're looking at? What jumps out at you? What's your focus right now? >> So Night Dragon's entirely focused in on what we call securing the world for tomorrow. So my entire focus is on security and safety across every domain from space to ground to cyber to AI to quantum whenever we see a new sort of domain coming and now we see quantum's domain coming how do we secure the quantum domain and what does quantum do to other domains where we got to protect that that's called the postquantum compute era PQC because it's going to infect encryption and decryption so we're focused on that what does AI do to all those markets We're already working on a lot of those areas. We also see next generation drones in space now. So you're going to see low earth orbit satellites, but you're going to see drones flying all over doing inspection, training, maintenance, things like that. Also, they could do cyber attacks and electronic warfare in drones in space. So wherever there's a threat or a risk, we want to have an investment model to secure the world. I I would just tell people that it's very difficult what you do because what you just laid out is great mission 100% love the mission however the tide shifted it reminds me of the days when I broke into the industry when I was my late teens um the computer industry was a term the computer industry and that was just computers >> and they put them into business government bought and people did stuff on them they bought software you were involved in many of those companies now if you want to power save secure the world for tomorrow you have to build it on the right stuff >> which means you have to know everything about the stack. So you kind of have a compatibility matrix I'm sure that it must match >> the ability to surround yourself with smarter people than you are and that's what I do too. So >> but you have but that's that's central to your thesis right you have to say hey if I'm going to have intelligence >> across domains and it has to be architected properly >> and something we tried to do you mentioned this earlier was we know at night we're not going to know it all but what we want to do is create a platform of people partners government who can help us with these thesis so working closely with the government and the war fighter you know what the threat is >> you know what the risk could be working with partners like the primes and global prime system integrators, you kind of know if they could sell it or deliver it. Working with your adviserss like generals and admirals and CISOs, you know, if they'd buy it. So, if you get all that, you can come up with what the next threat risk problem is. And that's what we hope to >> I have to ask you this question because you're um historian as well as a leader in what you're doing now. In the old days when you were running companies, you had a department, business development, you had salespeople, enterprise say oh go get the get the customer and there's a well-known process for you know that go to market bisdev would build relationships with um partners um some Barney deals as they call them and the old school um vernacular but partnerships were pretty much let's do a business deal maybe some technology transfer uh it seems that with cloud that's changed and now with the AI infrastructure and the and the role that it's impacting accelerating the role goal of being an ecosystem partner has changed. Could you share your vision on what's changed and or what's not changed? Because it seems to me that there's a lot more co-designing going on, a lot more deeper conversations than simply papering. Hey, we signed a deal. We're doing go to market together. We do a little >> license technology. We'll integrate it on our side. Maybe it's a little slower. What's the the new form of partnering in this era? Ironically, there's almost a back to the future happened here. Like a a full life cycle that's occurred because now in this AI era, you have a lot of Fortune 100, Fortune 500 type companies that are getting disrupted. So you end up with like primes. You take a Lockheed or a Northrup or a large prime suddenly all this autonomy comes along and access to procurement comes along and they have to change. So partnering became more important to them than ever as well. Or you take a Visa or a Mastercard who didn't worry so much about cyber security. They worried about it, but now they have more. >> They built their own stuff. >> They built their own stuff. Now they're buying companies and partnering in ways we've never seen. You look at the blockchain and all the Coinbases of the world like you start to see every era require partnering to be better because you don't want a catastrophic breach or threat to your environment. come all the way around. >> Yeah, I mean it's a whole another world and again it's game is kind of still the same but it's just different fashion I guess. >> Thanks for coming on the cube. Thanks for having me. >> Dave Waltz he's innovator but also leading the charge on defense tech. Again this is the kind of stuff we want to unpack with this program. Figure out where's it going? What does it mean to us? And again where the innovations because safety and security are paramount in the AI era but it's affecting up and down the stack from the AI factories all the way to uh people on the streets and of course protecting our society. Thanks for watching.