Transcription
I'd like to now hand it off to Chris Kemp, the CEO and co-founder of Astra. From being the first CTO of NASA, he has had a remarkable journey within the space industry, and he would love to share a few words about his journey and the company. Thank you.
Cool. Hey everybody, good to be here. Um, what I really am uh excited about is just being here where you guys do so much stuff. You build hardware. I think um, you guys were founded, I think, back in the 1950s. Since then, I think you guys have done about two dozen rockets. Um, we've done, you guys have done like over a hundred satellites. Um, I think the coolest satellite ever, uh, just coming back from Burning Man. Um, in 1992, you guys did the ex, the uh, ultraviolet explorer satellite and determined that the universe actually glows in ultraviolet light. Um, I love things that glow in ultraviolet light. So, saw a lot of that last week. Um, and uh, you know, this is something that uh, really uh, makes Astra, Astra.
So, um, you know, we started out in a garage in San Francisco and we built hardware, um, just like you guys do. And the hardware, uh, one thing led to another thing. And the more you build, the more you learn, uh, because you can do a lot of analysis and design stuff, uh, but until you actually build hardware, uh, you can't, uh, calibrate the analysis and the designs and, and make real progress. A lot of, uh, what you do, uh, just doesn't, you know, doesn't map to what you thought it would do. And so until you actually build hardware and test hardware, you're just not going to move very quickly. Uh, so I think what we're starting to see in the space industry is two very different approaches to innovation. Uh, an approach where, uh, I call like the Blue Origin NASA approach, where you spend tens of billions of dollars and in 20 or 30 years, uh, you build a rocket and it works the first time. And this is super important if what you're trying to optimize for is it works the first time, and for a program run by a nation-state or a billionaire that doesn't want to have a rocket blow up. Uh, this is, uh, prudent. Uh, but it does take decades and cost tens of billions of dollars for you to do the the analysis and the testing leading up to the event where you get a success the first time. Very impressive, very intensive. Obviously, that wasn't the path we chose. It's not the path SpaceX chooses, uh, after 10 Starships are, uh, you know, very, uh, in a very exciting fashion dismantling themselves on their way to, uh, orbit, but they're learning very quickly, right? And, um, so anyway, I, I think that's that's kind of where I'll start this.
Um, you know, back, uh, in, uh, late 2016, early 2017, uh, we were in a garage in Soma that was about 1,200 square feet. And, uh, this is, uh, Dr. Adam London. Um, he is the brains of the operation. He did his undergrad, masters, and PhD at, not Berkeley, but at MIT. Uh, where, um, he was, you know, very focused on very theoretical things. Like he built, um, a rocket engine that was the size of your thumbnail that was literally photo-lithographically etched on a silicon wafer, and all the valves internally were MEMS. And if you took thousands of these little things and glued them together, you'd have like a Star Wars-like engine, um, where you, you have a hypothetically very efficient engine, but, and you built them in the lab, and they kept exploding because apparently silicon is a pretty, um, you know, dusty thing to work with when you start to apply pressure to it and things like that. So, um, you know, Adam and I teamed up, and the first thing I asked Adam was, well, like, how do we make this less theoretical? You know, like making really small rockets is one thing, but, you know, is there a business case, uh, that is different than say, a Rocket Lab? Because they were well underway. SpaceX, well underway. Is there something different that we can do? And so the big idea that we had in this garage was, well, what if you just make thousands of rockets? Um, it was actually another academic paper, you can find still on the web today on the Fermi Lab website, called "A Rocket a Day Makes High Cost Go Away." And it was written, um, back in, I think, 1992 as well, when you guys did that cool ultraviolet, uh, satellite. And, um, the thesis was that the reason why rockets are so expensive is because you don't make many of them, right? And so all of your fixed costs, uh, are amortized over a relatively small number of vehicles. And so this drives towards a thinking around, well, you know, if you're going to make a small number of something, you better make it reusable, because why throw away an airplane every time you fly it? Sounds pretty logical. So this kind of drove this, uh, this idea around the rockets must be reusable, and rockets must be, um, very large as well, because if it's, you know, if you're going to maximize the amount of payload you're trying to reach to orbit, the mass fraction of the rocket needs to be as large as possible. And so you've got this school of thought around, let's just make the largest rocket ever and make it reusable, which is the SpaceX approach, and I think the industry is going. So our approach was always the opposite of that. It's, you know, what if you take a soda can? Soda cans. Uh, I think we make about a billion of them every day out of aluminum. Aluminum is basically free. Uh, and if you can figure out a way to take aluminum and turn it into, I mean, a rocket's basically a big cylinder, uh, with some domes on it and an engine. And if you look at an engine, a rocket engine in some ways is simpler than an internal combustion engine. It doesn't have, you know, a bunch of pistons and moving parts and magnetos, and it's, it's just a combustion chamber. Uh, if you, uh, depending on the engine cycle, you can actually tap off the combustion gases into a turbine. You have a single turbine that can turn two pumps. It actually has fewer moving parts, um, by a significant margin, than an internal combustion engine. So, you get back to the, well, if you just build a lot of them, maybe the cost can come down. And so, the basic idea in that garage was, let's do the math. Let's just look at what the BOM cost of a rocket would be if you made 500 of them, as suggested by this paper, "A Rocket a Day Makes High Cost Go Away." Um, and so that's how things got started.
So about a year and a half after we founded the company, uh, that picture was taken. That was not generated by AI. Uh, what's super cool about this picture, among other things, there's a rocket firing with the Salesforce Tower behind it, but, um, gives you the time stamp. Um, this little area right here is where they filmed the Matrix. That little black thing, that's the highway scene where they were kind of jumping around between cars and things like that. And you can still see, I think they finally put dirt on top of this, but there was, you could see from satellite images, the the soot on the asphalt that we charred there. That was Rocket One. Actually, that's Rocket 2.0. Um, so, uh, anyway, so that, that whole system, we could transport out to the airfield and test, um, about, you know, a mile from our office. And this is all about five miles from here, just to give you a sense of what's going on in your backyard. Um, we time-stamped, uh, pretty quickly after that, uh, demonstrated that we could build an orbital vehicle, and that image, uh, was taken on December 15th of 2020. Again, the company was founded more or less, uh, January of 2017 is when we actually got going. So the business was incorporated on, uh, October 1st, 2016. So pretty quickly, um, putting that back in the kind of framework of other companies, uh, we, we basically did this in just over four years. Uh, Falcon 1 at SpaceX took about eight, seven and a half, um, and then other companies. So how do we do this this quickly? Um, well, there was a movie about it. Um, we were not afraid to build rockets and test them, just like SpaceX is doing the Starships. Like, if the more you're able to build hardware, collect data, and iterate on the design, uh, the better. And so this led to a mistake. Um, that's fine. So, yeah, so this was 2021. The next, you know, the next year, we, we got to orbit and we're like, "Wow, this is, this is great. Um, let's go public." Uh, Virgin Orbit had gone public. Uh, Rocket Lab was going public. Planet was going public. A huge amount of capital in capital markets had been moved into these things called SPACs. I didn't know what a SPAC was, but it sounded, you know, like we got to raise some money. You know, the thesis of the business was, we've got to build a rocket factory, right? We're not building a rocket. We're building a thousand rockets. So, I need to raise a bunch of money to build a rocket factory because that's the whole point of the company. Uh, so let's go public. And, uh, anyway, that was the day that the stock traded at its highest value. The next day was lower. We would launch a rocket. It would work. It would go lower. Rocket would blow up. It would go lower. Nothing we did, no no sale. We announced basically every piece of news that came out of this company, uh, you know, the stock went lower. Uh, and that got to the point where, uh, you know, it just didn't work out. But I'll get to that in a second.
Um, what we were able to do is we were able to take the half a billion dollars we raised in that public offering to go build the factory that we needed to mass-produce the rocket. And that was key. So, uh, this is a building which the only thing that we painted since 1983 was the black little front of it. Uh, so every dollar, every wall socket, every Ethernet port, every workbench in this building was purchased. And literally, it was a shell of a building, uh, with lead paint, uh, just falling off the ceiling. We literally turned every penny into the means of production to build a rocket a day. And so that kind of looks like that now. Uh, it took a bit of time to build that out, like two years. But if you drive down the street, you can actually see a larger than SpaceX Hawthorne facility manufacturing plant that can actually turn out two rocket stages, uh, per day with one more of these little, I need two of these to do that, by a million bucks. Here's a little fly-through. It's kind of fun. [Music] And that's the rocket production line there. So we basically take a coil of aluminum that costs about $20,000 that can make three rockets. We uncoil it, we flatten it, we laser cut it, we recoil it, we friction stir weld it, and then there's a circumferential welder that builds the cylinders of the rocket. So the whole concept is, you know, you put in aluminum on one side of the building, and then rockets come out of the other side of the building. And most of the, there's Rocket 4, but most of the mechanisms, engines, valves, all the electronics, all the software, it's all made in that building. So this is kind of like a Gigafactory to Tesla is a rocket factory to Astra. And so you could put more rocket factories in other places, and you can make rockets in a lot of places. Why do you make rockets? You don't make rockets because they're fun. They are fun, but, um, you make rockets because they're delivery trucks to space. And so if you're going public and you're raising billions of dollars, then you need a business plan which is not just, we're going to make delivery trucks. Just like SpaceX is currently valued at $400 billion because of Starlink, like we wanted to have a satellite, uh, that would justify all the rockets we were planning on building so we could build a vertically integrated space platform. I'm not going to get too deep into that, but it's a whole interesting topic.
First thing we did was we bought a Hall effect thruster company because to build a satellite, you need like four things. You need propulsion, you need payload, uh, typically a phased array antenna, optical, um, interconnects, you need a solar array, um, and then you basically need the, um, what am I missing? The bus itself. So this was a key element of the efficiency of the constellation. So we went shopping and we found the most efficient engine because if we're going to have a small rocket, um, you need to get the most bang for the buck out of these satellites because the satellites are going to be smaller, uh, than what you might launch on a Starship. So, we found this company down in Mountain View, uh, led by a guy named Mike Cassidy. Super efficient engine, has like 1,400 seconds of ISP, perfect for putting about a 500-kilogram satellite in orbit and keeping it there for about five years, powered by Krypton or Argon. It had never worked yet, but we paid $150 million for the company because, hey, we just raised a bunch of money. Um, so, uh, this business, uh, actually kept Astra alive for the past few years. So when, uh, things got crazy, um, we were working on this rocket, and we had bought this spacecraft engine business, and then as we were trying to make that work, things went a little sideways. >> And, uh, I think we lost about a billion dollars that day. I looked at that and I'm like, it's annoying. So, um, at that point, um, things, you know, things got a little, uh, but we were working on this rocket, and the rocket was very important. Um, I decided, let's sell the satellite engines, you know, so let's, let's see if we could, because the engines had started to work, um, we had, uh, this big SDA constellation going up. We sold hundreds of these satellite engines for about a half a million dollars a piece. So we had about a hundred million dollars of orders for satellite engines. Um, and as a public company, uh, you know, this is revenue. This is good. Let's, let's see if that'll work. Didn't work. Uh, so, uh, basically, at the, there was a point, uh, in that journey where the market cap of the company was like $50 million or something like that. And so obviously, I'm the worst public company CEO ever. So I met with the board, called a little board meeting. I'm the chairman. So all my fault, obviously. I said, "Okay, board, um, you should fire me because this is just totally not working out, or you should sell the company. Um, and I want to help you do whatever you can do, uh, for our shareholders that have invested in this company. Like, whatever the best outcome is for them, we should totally do that. Um, and I'm going to fully empower you as the independent members of the board to figure this out. You should hire new investment bankers. You should hire new advisors because obviously the ones that I've hired are crap. Um, and, um, here's an offer to buy the company." Which was a bit of an interesting board meeting. Um, I also told them that, uh, they could fire me, but I couldn't fire them. So they were totally independent and safe for me, even though I had a lot of power over the situation, but they should go and do that. So, I helped them attempt to find, um, a buyer for Astra. And what happened there, this is about 2022, 2023, Virgin Orbit went bankrupt, and no one bought them. No one even bought their assets. And so, we're out there trying to sell Astra to Rocket Lab, to everybody. Um, but no one wanted to buy it because, you know, you Richard Branson's thing where you have a 747 drop in a rocket and the, you just get knighted. They're not buying that. Why would you buy Astra, right? And so we literally could not find a buyer other than me and Adam for this company that was once valued at billions of dollars. And so we bought the company back from shareholders so we could finish what we started. Um, the lesson there is, is, you know, if you've got a team of people and you are working on something important and you have customers and you have a bunch of engineers and employees that are not giving up on you, well, don't give up on them, right? Don't give up on your customers. Don't give up on your employees. Whether you're public or private or what the thing is. These are all, name one satellite company that hasn't gone bankrupt. Name one airline that hasn't gone bankrupt about three times. You won't find one. So, uh, in this industry, this is really hard stuff. And, uh, if you're a leader or if you're involved in a company, just see it through, right? What the, the thing that matters is, uh, this rocket began its life five years ago with a tremendous amount of hard-fought lessons learned from rockets that flew sideways, rockets that did all sorts of things. If you can keep that together, that's the only way to succeed. Look at SpaceX. You know, it took them a failed flight, a failed flight rocket. The third Falcon 1 flight, Elon had literally, he was broke. Like he literally borrowed money from, uh, his friends to get that across the line. And if he had not done that, SpaceX wouldn't be. Um, and it was ultimately NASA, while I was there, that gave him a half a billion dollars that he loaned to Tesla that saved Tesla. So, uh, the lesson here is, is just don't give up, um, on this stuff because it's hard.
So, um, that's, you know, one, the other lesson is product-market fit. So, we thought in 2017 that satellites would largely be like Planet Labs' Dove satellites. Um, the, the dominant form factor for satellites moved from these CubeSats to 500-kilogram communication satellites. So if you're building a rocket and you're seeing thousands of Kuiper satellites and SDA satellites and Telesat satellites and Iris squares, you go look at what's out there. Um, that little rocket was too small, and so is Electron. And so the problem is, is if you can only throw a few hundred kilograms, your TAM, your addressable market is basically insignificant, right? So it costs the same-ish to build the small thing as the big thing. So just build the thing that addresses the market. The problem is, when you're talking about a rocket, that's a complete redesign. You literally have to start over. There is almost not a single part. Maybe the video camera, uh, that isn't, you know, every valve is sized differently. Every run line is sized differently. The entire AR, you know, you, you basically have to, from a systems engineering perspective, you have to literally start from scratch. It's a do-over. And so, uh, all you take with you is the lessons you've learned and the experience you have making all the mistakes you made, uh, to get, get you there. So, anyway, that's what we've been up to. Um, I want to take one, uh, one thing that we did from the very beginning that I really liked was, why do we need to exist, right? Um, there's plenty of rockets out there. Eric Schmidt, uh, just took over Relativity. That's cool. Um, rockets are hard. Uh, he'll figure that out. Um, but like, he's going to go build a rocket that is twice as big as Falcon 9. Okay. So you've got Blue Origin, you've got the Starship, you've got Relativity, you've got Ariane 6, you've got Vulcan, lots of big, you've got Neutron, lots of big rockets out there, lots of big reusable rockets out there. What we haven't seen is we have not seen a completely mobile tactical rocket. So, what if this, here's a, here's a big idea. What if you could put a dozen drones in that rocket? What if you could go put those rockets anywhere in hundreds of locations? I just saw yesterday, uh, the Lockheed Martin Missiles and Defense group got a $9.8 billion order for 1,780 Patriot missiles over the next two years. Okay, Patriot missiles, uh, are used when you have like Houthis lobbing things into Israel and you need to like intercept things. They're 10-inch diameter, 16-foot long solid rocket missiles, basically for $4.7 million a piece. They sold those $10 billion contract Patriot missiles. So, Patriot missiles, uh, it's a hundred billion program. There are about 1,400 Patriot missile launchers organized in batteries of 48 launchers, uh, in about 130 locations in 18 different countries. Why do I know this? I think it's interesting. Uh, because if there are Patriot missile systems in 18 different countries, how many space agencies exist in the world? 100. How many space a, how many sovereign nations can actually get to orbit? Six. So there is a total addressable market of like 75 space agencies that would love to be able to get to space. How are they going to do it? China, India, or Astra. This is the way you do it because I can pack these rockets up and the spaceport into shipping containers and I can send it. We're getting, we have a DoD contract already to do the first launch out of Australia, the UK. You know, every European attempt to build a small rocket has up until this point failed. There is a global market to, just like in the early part of the last century, use American aerospace technology to to transform the, uh, aviation industry. We can transform the space industry, but you're not going to do it with the Starship, right? Because the international airport starts out as a dirt airstrip. You know, um, when you land at the Oakland Airport down the street, the, this is where Amelia Earhart landed. There's an old building there. You know, you, you take a small airplane and you land it on a dirt airfield. We can launch this rocket from a dirt strip. The last launch that we did out of Cape Canaveral, which they told us it would take two years to do, it took us six months. Um, first part 450 license was done on a launchpad that did not have electricity. So, we showed up with our containers, we set up, and we did a launch in six days from arriving at the site without electricity, and we had a network connection. Now, we don't need it because they're Starlink. So, we could literally go and airlift this system onto an island and launch it. I like to joke we could do it from a Walmart parking lot. Walmart doesn't like that. I mean, you don't want to crash on a Walmart. But, you know, the idea is, if you could put these portable mobile spaceports in a dozen countries, if you could have millions or thousands or hundreds of thousands of Anduril drones, uh, in thousands of rockets, in shipping containers, uh, in hundreds of locations in dozens of countries, why the hell do you need 850 military bases all around the world that cost the better part of a trillion dollars to operate? If you look at the trillion-dollar defense budget, about $800 billion is O&M, which is code for that's what we pay to run the machine. It's not hardware. It's just we are spending $800 billion of your money every year to operate hundreds of military bases all around the world. So, a bunch of Navy SEALs can put a bunch of drones next to a border that can fly to do a mission. Rockets, AI, and drones are the future of defense. Full stop. And so if you can put drones in rockets and you can put rockets anywhere, why would you? Right now, we have a defense posture of, if space is so important, if GPS is so important, if communications and Earth observation is so important, um, who cares if you have thousands of satellites if you only have two spaceports? You know, the adversary just has to take out two locations and they shut America out of space. That is a vulnerability which will actually create a war. But if the adversary knows that you can replace any satellite from anywhere, it's actually the most powerful deterrent for World War III ever. So if you go back to the thesis around, uh, the, the whole arms race with the with the Soviet Union, there is a new arms race, and it's about rockets, AI, and drones. And whoever can have a more resilient infrastructure in space, uh, if you can achieve that, you can achieve peace because you basically become impossible to attack. You can't disable what you can't see. So our rockets will always fit in shipping containers. Our spaceports will always fit in shipping containers, and we'll put little Walmart logos on them and Amazon logos. And so you don't know whether it's the Amazon delivery truck. You don't know whether it's an Amazon distribution center or a rocket factory because what's going, what's going on is, you know, in like in space, what do you see? So, there's a future here where, um, this kind of stuff is becoming increasingly important. There's our rocket being loaded into a C-130. We've only ever launched this way. This isn't hypothetical. Every time we've ever launched, we've done a hot fire test here in California and then we've shipped the system to wherever it's launched from. [Music] Cool. [Music] Okay. Um, other things going on. So yesterday, uh, we just, Al Weston's here, one of my heroes. I have the opportunity to work with him now. Um, uh, can't say enough about Al. Al and I worked together at NASA as, uh, he ran all the programs. He discovered water ice on the moon. Uh, crashed the spacecraft in bombed the moon. Uh, worked on a bunch of missile defense stuff, made a lot of things that went boom. Uh, generally, the, the, the Aegis interceptors that are protecting us from things that are attacking our bases was his program at the DoD. Kind of cool. But, uh, this is his engine. So this is, we have a new rocket engine. Uh, there was a company called Firefly. They went public. Um, and, uh, we bought the engine from them, and it was garbage. So, um, we literally couldn't get the same engine twice from them, and none of them matched the CAD. And if you're in engineering, you know that that just doesn't work. So we had to basically start from scratch with this engine. So we built a completely new combustion chamber out of a new copper alloy called GROP. And, uh, yesterday, that engine fired for the first time, and, uh, we are planning on trying to fire it again here in the next few days. Super cool engine. It's the, I believe, the largest 3D printed, uh, engine. Um, it's printed in two parts. Uh, uses some of the heritage we developed around welding the engine. So, it's, it's basically printed in, inside of itself in a printer bed, and then it's welded together. So, it stands about that tall. Um, and it's a super cool engine. And we're going to have two of these powering Rocket 4, producing just under 50,000 pounds of thrust. So, we'll be able to throw about a ton to orbit with the new rocket. So, a little bit bigger, a lot more payload capacity. 100 times the payload volume, at least 10 times the payload capacity. Um, here is a video because rocket engines are cool to watch. Wait, would you turn down the brightness? Anyway, they're cool. This is, Oh, this is actually kind of cool. That's the, So, this is what's happening literally right down the street from you guys. We have two indoor rocket engine test cells where we did thousands of tests. This is actually being, uh, this test is occurring at our Castle test site, which is about an hour and a half from here, because this engine would disturb the neighbors, especially if it didn't work. But, so this, this is using a tap-off cycle based on Ukrainian IP. We take the gases from the combustion chamber, we duct them into the turbine, and we have a single turbine that turns the fuel pump and the oxidizer pump. So basically, once this engine is running, it stays running, uh, because the combustion gases are what's powering the turbine, and it's a very simple, it's a very simple, inexpensive engine to build, um, versus other engines that effectively have multiple rocket engines that power the turbines. A lot more valves, a lot more complexity. So we've really tried to figure out what is the simplest engine on the simplest rocket that you can mass-produce. Um, there's a lot of cool software stuff happening, but I'm gonna run out of time. Um, and this is a cool video. It's like, oh, you didn't. This is the dark music. This is the Trump music. [Music] [Music] [Music] [Music] Heat up here. Cool. All right.
So, with that, um, the reason why I'm here is because you should all come work at Astra. Um, we're, we're literally like 10 minutes, uh, from here. Uh, we have an incredible team. It's a startup again. And, uh, you know, one of the things that I really enjoyed working on with Al is, uh, just ripping out all of the overhead that we took on as a company, as a public company, uh, all the financial systems and compliance stuff, and just getting it down to this like really beautiful core. Um, and so we're focused on how do we make engineers productive? How do we increase the velocity of the team? How do we increase the rate of learning on these teams? And so, um, it's great to see all the different teams working on avionics, working on the stages, working on the engines. Um, and, uh, we had about 25 interns this last summer. So, if you're enrolled here and you want to like come do some part-time stuff, uh, just go to this website. There's an application for, um, a new graduate or a, um, graduate position, just apply to that, and, um, we, we'd love to have you. And, uh, it's, uh, it's one of the, one of the most fun. It's, it's, it's more fun than SpaceX because we're not, we're not on the border of Mexico where they'll chop your head off if you accidentally took a left turn. And, um, you don't have to live in a trailer, and we don't make you work six and a half days a week, 12 hours a day. It's appreciated if you do, but, um, it is not required. So with that, I will, um, I will wrap. Thank you all.
On behalf of Mspace, we would like to thank Chris Kemp for our being our keynote speech. And we are going to open the floor for a quick Q&A. So if you guys have any questions, please raise your hand.
Questions. Yes. Go ahead. Hi. Hi, Chris. Uh, thank you so much for that presentation. My name is Rama Fulu. I'm the founder and CEO of a company called Satellite SATLYT. Uh, you spoke about software and you said that there wasn't enough time to go into that. Like I'd love to hear a little bit more about what the software angle is from from, uh, uh, the new, the new Astra.
So, yeah, very early on, we realized that data, uh, was important. And so, uh, we have, we have a platform that I'm super excited to leverage where whenever we do a test, uh, we can actually take every sensor that's recorded, every bit of data from any test in the entire history of the company. We have all that data. Um, we know we don't use LabVIEW because that's that's like a toy. Um, we take the the data from the data acquisition devices, which sometimes are made by National Instruments, but don't have to be, and it gets queued up onto a controller running a platform that we've built. It gets, uh, streamed up into AWS GovCloud. Um, and then there's a massive data lake. So, I can take all the data from, for example, all these launches, and I can take every sensor from every vehicle, and I can overlay them and I can see patterns. This was actually super cool five years ago. But what's even cooler now is that you can use AI to tag the data and you can find patterns and anomalies. Um, and so we're in the process of building a GPU cluster right now. Um, most everything we've done has been published in a wiki. Um, if it's knowledge and in, if it's an action, it's in Jira. Uh, so this Confluence Jira suite of tools from Atlassian literally contains the sum knowledge of every employee that's ever worked at Astra. And so we're going to train a model on all this information to effectively create Jarvis. So like imagine if every requirement, if every document from every supplier, everything that we've ever learned, every employee that's ever worked at Astra were available to every employee that does work at Astra, and then you can ask it questions and you can have it effectively cross-correlate these things on a model. So I think that's that's the short answer.
Yep. Next question up here in front. Hi. Um, you mentioned earlier that you're working with Anduril. Anduril is a defense, uh, company. They're kind of like a tech company. So I think of them as like the Silicon Valley of like defense. Um, in terms of like loading Astra's rockets with Anduril drones and having them fly anywhere around, you're saying that it's a huge possibility that this is the new deterrent for World War III, and having that deployability and having that action. So in terms of like deployability and reliability, where does Astra stand on that?
Well, I think the defense department understands how important this capability is. So they've been embedded at Astra throughout the development of Rocket 4, uh, to, uh, certify this as what's called a category 2 capable vehicle. And so there's an enormous amount of, there's, there's, there's a real shift. Rocket 3 was the kind of cowboy rocket. Um, Rocket 4 is the, okay, um, we know we can do this. Let's do it, let's do it, right? And right is it's not the $20 billion, 20-year journey, but it's also not the, you know, let's just throw it together and see if it works and not document it and not have. So there was, there was a huge effort while we've been quiet, um, to do a fairly, uh, comprehensive, uh, suite of failure mode effects analysis, systems engineering, require, you know, test coverage analysis. So, uh, what's been happening at Astra is every single weld, you know, has been pulled, you know, a hundred times so that we know exactly what the manufacturing processes for everything that goes into the vehicle because that'll pay dividends when you can knock these things out on a production line because you can automate the stuff. So we literally deployed hundreds of millions of dollars in a factory and hundreds of millions of dollars in an effort over many years to do the systems engineering and the work to make Rocket 4, uh, something that can be mass-produced at scale. The big idea was, if you are going to launch a rocket every day, 1% 99% reliability, you know, like Space Shuttle reliability, not good enough, right? Because if you lose a rocket two or three times a year, are you going to really be able to fly every day? So the big, the big idea was actually, you can't get to scale without reliability.
Next question over there. Hi Chris. Uh, right over here in the fourth row. There you are. My name is Abi Kumar. I teach the new space economy course here at UC Berkeley. Cool. Uh, in a previous life, I was also leading the partnership between Planet and Astra for for a few months when I was at Planet. Uh, two-part question. Uh, question number one, and this might be five years before I should be asking this, but since I have you in the room, if and when you choose to go public again, what would you do differently this time? That's that's that's question number one. And question number two, what you just described about Rocket 4's, um, you know, capability being deployed in different geographically disparate locations sounds like almost like tactically responsive launch capabilities, right? What bottlenecks or what roadblocks do you see, uh, especially from a geopolitical standpoint, uh, for that vision to actualize?
I want to take the second question first because it's just more fun. I think with this administration, it's just a question of like, can we send the right 24-karat gold-plated trophy situations to the right people? Um, because I think this administration really wants to deregulate everything, right? I think this administration, uh, might be the administration that makes, uh, America less protectionist in space, right? That abolishes, uh, aspects of it, CPHAS that are impediments, uh, to, uh, seeing American space technology proliferate globally. Um, and so say what you will, the rest of the policy stuff, but that's, you know, that is not a bad thing because otherwise, we're just going to see China and India just dominate, um, global space, and that's probably not good for America. Um, so that's the second answer. Uh, the first answer is, you know, I think being a private company is great. You know, having, you know, all of my time focused on the business is great. You know, as a public CEO, you have to focus a lot of energy on what's happening this quarter and, you know, what the analysts think of you and a lot of compliance stuff. Uh, you've got a, a board that's entirely independent from you that you're having to manage. Uh, being, you know, being actually running a business is a different thing. Running a public than running a public business. Um, if you have to be public, uh, the one observation I would have, the reason why American manufacturing is just, you know, screwed is because American public companies don't manufacture anything. They integrate. If you look at a Boeing or a Lockheed or Northrop Grumman, they don't have machine shops there, right? They're buying parts from a bunch of private mom-and-pop companies that have no ability to scale. And so, uh, what, what that was never our playbook. Our playbook was, you go to Astra, and there are 20-foot bars of aluminum on shelves and racks, and there's a saw house, and we saw them into pieces, and we bring them in, and we have CNC machines, and we machine the things. That's important for development because if you're trying to go fast, you don't want to wait for a part. You know, if you say, if you go to any machine shop and you say, "Hey, I need this part. I need one of them." You know, good luck. Okay. So, if, but, but if it's, they work for you and you say, "I need this." Our entire machine shop is working all weekend this weekend. Why? Because I asked them to, right? You can't do that if you don't control the company, right? Or you pay a lot for it, and then the costs go up. So, you know, if the real problem is that if you're a public company and you have Price WaterhouseCoopers or whatever come in to audit you, and they start asking questions like, you know, where's the oil that goes in the saw blade accounted for? You know, that's annoying, but it's also expensive, and it means that you have a large finance organization, and you have a large, um, you know, you're paying five to six million dollars a year for PWC to dig through all this stuff. So, running a manufacturing organization with all of the detail and complexity of cost accounting and the, and the accounting that goes on of all of these things, like, is expensive. And so, what I would say is, it's not impossible, but it's a huge hit to margin. And so I would say we'd need a billion dollars of revenue at 30 to 40% margin before I'd even consider it, right? Because otherwise, it's just too much of a tax on margin. It's like, it's like a rock. I like to equate companies to rockets. You know, if a rocket has too much mass, it just doesn't reach orbit, or you have no payload. So I would say a company is the same thing. Like, if, if your company does not have enough thrust, like enough revenue, enough growth, uh, then it will literally never reach critical mass, and it will not be able to stay a public company. It will not reach orbit, like it will never get there. Um, and that's, that's really fundamentally what a lot of companies, that's the mistake a lot of companies made. They, they underappreciated the cost and the overhead and the friction associated with being public. And unless you print money, uh, you know, like a lot of tech companies that are just very high-margin software businesses can grow that quickly because you can grow revenue that quickly and margin that quickly. But a manufacturing business where you're having to scale capex and hardware, it's extremely hard to be, it's hard to be a small public manufacturing business.
Thank you so. I would be a large public manufacturing. Um, last question. If you could come up. Hello. Um, I had a quick question, um, sort of a follow-up, um, about your new approach in terms of launching from different geographical areas. How is Astra, uh, kind of going to find the balance between, um, giving your potential clients who may be, you know, nation-states, giving them as much flexibility in terms of launching how they want to, and kind of also finding the balance between, hey, you're not giving up too much, and then they're not able to back-engineer it. So how does that balance between knowledge transfer and export restrictions, kind of like, what's your ideal view on that?
I mean, even a spaceport is considered ITAR, right? You know, it's because it's where the QD interfaces with the rocket, that's IP. So if you really want to respect ITAR, just put the entire thing in shipping containers, show up on a dirt pad, and take it with you when you're done. There isn't another rocket system that can be deployed that way. Um, it's how the Patriot missiles are protected. Literally, you can just have someone that works for the DoD with the thing and don't have them leave and put cameras with AI on it and just make sure that your IP is protected. Um, that's, that's the answer. Like no other system can, uh, be deployed this way, which is why the launches are being, uh, done in the, in the Five Eyes, which is our, our most strategic intelligence partners first. But beyond that, I think this is the ticket, like to see, to see American space technology really proliferate around the world. Good with that. Such a pleasure. Thank you. A lot. Thank you. Wow. This is great.
Amazing. Thank you all so much. Such an honor. Appreciate it.