Shield AI's Brandon Tseng on 10 years building AI pilots for fighter jets and the new developer platform launching in May

doing we're doing great how are you doing great Fant fantastic Thursday afternoon that's great yeah thanks for taking the time to be with us uh for those who don't know could you just give like a little intro on yourself and the company uh yeah sure quick intro on myself uh mechanical engineering background I went to the US Naval Academy became an engineering officer on board of ship uh deployed to the Arabian Gulf on that ship um then laterally transferred became a US Navy SEAL uh and did uh three deployments two to Afghanistan one to the Pacific got out uh in 2015 while concurrently going to business school started Shield AI uh with my brother and have been at it for the past 10 years that's background on myself uh quick background on Shield AI uh mission is to protect service members and ions with artificially intelligent systems in pursuit of this Mission we've been building the world's best AI pilot easiest way to think about an AI pilot is self-driving technology for unmanned systems that enables them to U maneuver without GPS without Communications it enables them to maneuver without a remote pilot enables the concept of swarming or teaming um notable Milestones along this journey we're the first company in the world to put an AI pilot on the battlefield in 2018 we put it on a quadcopter go inside buildings um ahead of Special Operations forces uh that was deployed to Iraq Afghanistan Syria uh most recently used in Israel to rescue hostages which was really cool have the text messages back from uh Israeli counterterrorism forces uh during uh from October 2023 um and then uh also uh we were the first company in the world to put an AI pilot on an F16 and fly an F-16 autonomously we won the D alha dog fight in 2020 in 2022 um did the first F-16 flights in 2023 we did the first ever AI piloted F-16 versus human piloted F-16 uh dog fight and then in 2024 the secretary of the Air Force flew uh our AI piloted F16 we are finalist for the Coler trophy which is given to the greatest achievement in Aeronautics or astronautics R Brothers won it Chuck jger won it for breaking the sound barrier we were finalists and lost to some NASA asteroid mining thing cool whatever yeah but uh yeah he's paying attention yeah I mean it sounds like an overnight success the type of overnight successes we like to highlight on the show overnight success that's always the way it goes all the all the greatest overnight successes take decades in my experience yeah I mean it's a 17-year overnight success if you count the the naval experience yeah yeah exactly uh uh what's it like going uh having your your first you know real startup you know become a multi-billion dollar company uh most Founders have the benefit of you know starting a couple things that don't maybe you know little ideas or things like that is it you know going back 10 years ago at this point I guess nine years and 11 months uh did you feel like you knew what to do from the beginning or you had good people around you I'm curious you know um one is you don't know what to do in the beginning um you want to surround yourself with great people just to be uh directly responsive to those aspects of the question um it's uh I I don't think too much about like you know that that aspect of um like the aspect of oh it's now a multi-billion dollar company principally because it's a Night Bite every day um it's a lot of pain and suffering I I tell everybody that like when Jensen Hong from Nvidia talks about pain or suffering when Elon talks about the pain and suffering that goes into uh building these uh these types of businesses it's absolutely true um and so uh like there's a a massive amount of stress there's a massive like number of times that you fail there's a massive amount of embarrassment like when you fail in front of customers or when you disappoint customers along that 10year Journey um and so uh you know I think the re you know one of the reasons I think we've been successful is um the ab the resiliency the ability to keep moving forward the ability to pick yourself up once you get brought to your knees like when you get gut punched uh and it certainly happens and so you know in in that sense you know I think we had that when we started I I experienced that in the Seal team so I was you know grateful to have those experiences and um and basically used to the life of suffering so uh you know that's that's paid off well over the past several years years can can you talk a little bit about the path to full autonomy for fighter planes uh we saw this with chess where you know deep blue beat Gary Casper off but for a long time a combination of AI and human would outperform they call it like Centaur chess uh some people say we're in the Centaur era of AI image generation or a lot of people use chpt but then they're editing it themselves it's not pure AI creation it's not pure human uh are we in the Centaur era of Air Supremacy right now or are we getting there and how long do you think it'll last I I I think um at this stage in the game uh autonomy is at superhuman levels of performance and so I don't yes people will always augment machines and make those machines better and more useful to the people but the level of involvement in doing that um is becoming minuscule in terms of like what's required and so um look I you know that Journey right for us uh to flop starting to fly in F-16 unfortunately in 2015 um when you're a you know threers founding team with zero money zero track record uh they don't hand you to the keys of the keys to the F-16 right away and so we had to earn our right you know our path you know our right to fly that aircraft um and it took a long time and we St stepped our way to get there uh right it took seven years before first life T 16 proving it on a quadcopter proving it in simulation uh with the DARPA alpha dog fights uh proving it on our vbat aircraft and eventually you know they you know more and more people now are giving us quote unquote the keys to uh their jet aircraft yeah um but it's um I yeah I'll just say like I autonomy is a really really hard problem it's also really you know value thing to create and we're at the stage now where you know I yeah I'm a big believer I think the entire future is going to be an autonomous future I think I agree with Elon Musk when he says there's going to be you know millions of autonomous systems in the world his goal is to put a billion humanoid robots on the planet you know we'll have millions of self-driving cars um I believe there will be millions of self-driving aircraft and millions of other autonomous systems and shield AI our objective over the next 10 20 years is really to enable that future can you talk about the developer platform that you guys were building I know that's part of the that that was part of the announcement that you did with the new round at the beginning of last month uh we asked ourselves the question um you know how do we put a million AI Pilots into customers hands in the next 10 years we spent 10 years building the world's best AI pilot for all the reasons I mentioned I feel like we've checked that box um and then it became how do we how do we put a million AI Pilots into customers hands and the answer was we have to we have to supercharge um the Aerospace and defense Market we have to supercharge uh the autonomy industry and we took a page out of Amazon web services Playbook right uh if you're unfamiliar Amazon web services um they had basically built a great internal product uh Team said Hey what if we you know brought this to Market what if we commercialized it would it be valuable for other people all this infrastructure these data pipelines these developer tools uh and obviously the rest is history massively valuable for Amazon web services most profitable uh business unit they have um and so uh we basically said let's package up all of our developer tools our infrastructure our pipelines and enable the rest of the defense industrial base and again the rest of the the autonomy industry and those interested in autonomy to actually uh develop evaluate test and deploy autonomy um and so spent you know the better half of last year well more than last year really focused on building that product um and rolling out to Beta customers and we're going to General availability here in May for that product and the response has been gamechanging it essentially enables you um to reduce the number of Engineers reduce the time uh to actually build a ton of and take it to Market by 10 to 50x so it's just massive improvements we're we're flying first flights on a jet aircraft two engineers in six weeks right wow um if we had these tools back in 2015 they may have let us fly uh the F-16 right out of the game so it's it's incredibly powerful yeah uh speaking of the F16 what what is actually the hardest part of flying an F-16 in a military scenario I can imagine the dog fighting feels like the most intense the most high stakes but then uh it also sounds really hard just to fly an X F16 like into the theater for 6 hours straight you get tired and then landing on an aircraft carrier in like stormy seas that seems really hard are you trying to do it all is there is this the idea like self driving you kind of turn it on when you're on the highway uh what's the most difficult so uh just to be clear the work on the F-16 uh was to burn down the technical risk for this next generation of uncrewed fighter jets but in terms of like what the you know the most technically hard like the techn technically hardest part about it is actually around safety and certifi ability right um the US Air Force the US Navy they're not putting up 20,000 pound aircraft uh in the air right that are just going like not perform right especially in a you know in the dog fighting scenario when we have relative closure rates between two 15,000 pound aircraft going uh at 1,00 miles per hour closure rates within like you know call it 50 meters of each other you have to be able to say I can guarantee the safety aspect of what's going on here yes there's risk but this is how it's been burnt down this is why it's safe this is why we believe it's going to do uh exact or why we know it's going to do exactly what it say it's going to do so as people think about um you know and I think that's where Shield AI has uh really uh shown is our understanding of not just like the autonomy aspect but that intersection of autonomy uh jet aircraft safety reliability certifi ability um that's where Shield AI stands out do we need spent a long time working on uh do we need a new Top Gun movie where the protagonist is a software developer um it probably wouldn't be as entertaining as as Maverick too but we're you know or as Maverick but we're I'm open to yeah I'm open to it I'd love how do you how do you think about uh m&a going forward I think you guys made a couple or at least a couple uh Acquisitions to date there's now a ton of new investment but you guys have spent 10 years winning customer trusts you know winning uh you know contracts Etc and I imagine there's going to be some companies that develop you know cool Technologies or have great talent but don't quite uh reach the where you guys are you know how how how do you how do you look at that opportunity going forward certainly we we've built in acquisition muscle and acquisition arm in our in our company um and we are constantly looking at opportunities um and there's just a very very um it has to go through a lot of different wickets for us to say let's let's get serious about something um right ideally it's strategic um because it takes up a lot of executive time if it's truly going to move the needle for the business at the same time you'll see us do smaller Acquisitions um we'll announce like a really you know a smaller uh IP type acquisition here um in the next several weeks um and so there's that range but in the ideal World um you know strategic needle moving for the business um enh you know gives us better Market access gives us uh enables us to build a better product uh one that's aligned with our strategy road map with our product road map with our vision for the business and so we're always on the lookout for uh new companies it's fantastic last question or we good no that was it I mean come back on when you have UHS this was fantastic I mean such a fascinating business and like such a entertaining and just dramatic in so in so many ways uh I mean we could go way deeper uh but thanks so much for taking the time to talk to us uh really fun conversation thank youw thanks guys I'll see you we'll talk to you soon bye jordy's uh hit in the bathroom but uh in the meantime we have Peter from Rune Technologies here let's bring him in and I'll have him do his intro and then Jordy will hop on in just a second uh so Peter are you there can you hear me hey Peter how you doing he I got you cool uh yeah thanks so much for joining I've heard a lot of great things about the company I actually haven't gotten a chance to really dig in too much so could you just give me a brief intro on