Running the Algorithm: SpaceX’s Approach to Exponential Growth with VP of Launch Kiko Dontchev

Welcome to the stage: investor, advisor, and former NFL player, Dhani Jones.

[applause]

What's up, Summit! I'm gonna need another one — look, it's 9:30 at night, I need you all to bring that energy. What's up, Summit!

All right, I've been coming to Summit for over a decade. It's a long time. But guess what — over that decade, you dreamers and doers have been changing the entire trajectory of the entire world in our daily life. Things that people have created have come from the Summit community, and it has changed the way that we look at societies, changed the way that we've done things in society. And it's your responsibility to continue doing so. So if I haven't met you yet, I will, because I'm excited about meeting you.

Now, our next guest — I'm so excited to hear about his story. Because I was sitting in my house on the couch and I was watching the TV screen and I saw a guy named Elon Musk come across my screen, and I was like, wow, that's Elon Musk. But the amazing thing about what I saw was that right next to him was one of the most important people that is changing the trajectory of our society. He is allowing us to see the distant and beyond. He is providing us visibility into the future of our society. He is a Michigan graduate. He manages over a thousand people. He takes care of more than a tenth of SpaceX. He runs that Falcon Heavy. He is the VP of Launch, and he is my close personal friend — Kiko Dontchev!

[applause]

Dhani, with the flair! Dude, what an event, man. Plus I feel like I'm doing this talk for my friends now, which makes it much easier. I went from a chat, to a book end, to a dance party, to dinner, to a dance party, to a talk about rockets. So we're gonna keep this light, we're gonna have fun. It's 9:30, it's Friday night — let's have a good time. Who wants to watch some rocket videos?

[applause]

[video plays]

SpaceX, com track, ground stations. We got to see ISS up the window, which was pretty neat — largest window ever flown in space.

So most people know us as SpaceX, but I actually think it's really important to talk about our full name, which is Space Exploration Technologies. The reason is, most people think we're a rocket company. We're actually an exploration company. I also like to say we're an infrastructure company — it just so happens that the place we want to go explore and the infrastructure we want to build is on another planet and on other celestial bodies.

Obviously our focus is on transportation, but that's really the core of our company — to be explorers and to push human boundaries and technologies farther than they've ever gone before. With that said, we gotta get there. So our current focus is on building rockets and vehicles.

I'm going to show you the family of vehicles we've got. First, a human for reference. Next, the Millennium Falcon — about 35 meters. Honestly, always felt bigger on Star Wars to me. Then Falcon 9 — it's our main vehicle, the one we're flying every day. And then we're going to talk a little bit about how it really accelerated our launch rate. The reason it's called the Falcon, by the way, is it's actually named after the Millennium Falcon. And the reason it's called Falcon 9 is because we have nine Merlin engines that power it.

On top of Falcon, you can either put a Dragon — that's a capsule that can take cargo or crew to the International Space Station or low Earth orbit — or you can put a fairing, which basically means you can fly any sort of satellite or payload that you want, depending on where you want to go. Falcon Heavy is basically three Falcon 9s put together — 27 engines. It's really meant for super heavy satellites, or maybe something interplanetary that you want to send to Jupiter or Mars.

And then last but not least, Starship. Starship for scale — Falcon 9 fits on just the top part of Starship. Starship has the vehicle on top plus the Super Heavy booster. Now, all of these vehicles — the place we put most of our innovation has been in reusability, making rockets reusable. Falcon 9 is an evolutionary technology when it comes to reusability. Starship will fundamentally change the game and change the way we live and will truly set us up to be a spacefaring civilization.

[applause]

So let's focus on Falcon for a bit. We're gonna try to keep this not technical and go through that part quickly. Basically, a rocket has three parts. The first stage — that's from where the engines are to the top of that black part. That's the part that gets you up out of the atmosphere and starts to accelerate you to orbit. Second stage — from the top of the black part to where those two clamshells are. That's the part that delivers the payload to its orbit. And then there's the fairings — just two halves that sit on top of the satellite, protecting it when you're on the pad and as you're going through the atmosphere, before they deploy and fall back to Earth.

Now, if you build an airplane, an airplane's fully reusable. You fly from LAX to Miami, the plane lands, the people get off, the maintenance crew comes out, they refill the plane, people get back on, it takes off. Now imagine if on that plane from LAX to Miami, along the way the wings fell off, the tail fell off, the other half of the plane fell off, and the only thing that landed at the end of the runway was just a few slides with the people in it. You just spent $250 million on a one-way trip. That's kind of not very scalable and not very affordable.

Fundamentally, that's what rockets have been for effectively the last 50 years — until Falcon 9 came around. What we started doing is figuring out how to land that first stage. And by figuring out how to land that first stage and land those fairings and bring them back and start reusing that hardware, we were able to fundamentally start to accelerate what launch looked like.

Let's walk through the history of all the launches we've been doing. We started with this little itty-bitty rocket called Falcon 1 that we launched off an island in the South Pacific called Omelek, early on, 2008–2009. Then we built our pads at the Cape, just three hours north of here, at SLC-40 and LC-39A — historic launch pads — and also one out at Vandenberg, SLC-4E.

We had two launches in 2010, but we were fundamentally limited because we had to produce all of the hardware — we weren't able to catch the rockets, weren't able to reuse the vehicles yet, and that really slowed our pace. You kind of see, 2012 launched one, 2013 maybe we start to launch a couple more — still a huge gap. We go through some development challenges, a couple of failures in 2014–15. We start to pick up the pace a little bit. Then we have an accident that unfortunately slows us down for a bit, right around after 2015.

But then when we get back to flight in 2016, for the first time the rocket comes down — and it comes down in one piece. As soon as that happens, we start to pick our pace up, because you go from having to produce all of these machines to being able to reuse, refly, and just focus on maintaining the ones you have. All of a sudden, throwing a rocket away became an unacceptable thing to do. And our pace picked up faster and faster.

But even just going back to 2019, we'd really only ended up launching 15–16 times. Until we had this dire need, which we'll talk about shortly — we had a satellite internet constellation that we really wanted to launch. We needed to launch a lot of satellites very, very fast. So we put an immense focus on picking that pace up even more, focusing on reusability, going faster and faster — more launches. From a launch every two weeks, to a launch every week, to a launch every five days, to by the end of last year and the start of this year, a launch every four and three days.

[applause]

We're launching so fast that I made these charts last week and they're already out of date because we've added two flights. Over the history since we started launching rockets, we've flown 224 missions, recovered 184 boosters, reflown those boosters 155 times, and recovered 230 fairings. The reason we've recovered more fairings than missions is because there's two fairings per rocket launch.

What does that look like in a different way? We flew a record number of times in 2022 — 61 times. Effectively, we flew fewer new rockets than we did in 2014, but launched almost 10 times the amount. You can see the difference that being able to reuse your vehicles makes — that's the game changer, that's the evolutionary technology that Falcon is. No one's even coming close to this. But it's the key to making spaceflight like aircraft flight.

Record number last year. This year we're at 32 as of this morning. For those that were lucky enough to be looking outside at 2:15, you could actually see one, which was pretty cool. I didn't want to advertise it because I was nervous that you wouldn't be able to see it and it'd be a big disappointment. But it was actually pretty rad to see this streaking through the sky. And now we're kind of set up for a hundred.

Now I talked about reuse, but we do have a process at SpaceX for how we innovate, how we drive technology forward, and what is our approach to innovating with technology that's never been done before. We have what's called "the algorithm," and Elon really built this through his experience at Tesla and SpaceX — high-rate manufacturing, hardcore design problems.

One: make the requirements less dumb. When you're fundamentally innovating a new technology, you're wrong — it's just a question of how wrong. Because your ability to learn is changing constantly, and where you start is certainly not where you're going to end up. Another way to think about this is that your innovation is happening at such a rapid rate that there are constraints put on you maybe a year, two, three years ago from some department. Your first natural inclination is to go in and be like, okay, I've got this problem, what are the constraints? Someone told me it's this heavy or it's got to cost this much money. But fundamentally, a lot of times those constraints were driven by a department or a person three years ago that isn't even around anymore.

So if you're ever facing a problem and someone's giving you constraints, make sure you ask: who's the person that gave me this constraint? "Oh, well, it wasn't a person." Whose department? Okay, you go to that department — who in this department said this? "Actually, it was this intern six months ago that's not even here anymore." I'm like, I don't even know why this constraint exists. So be surprised what happens when you challenge the requirements. That's the first thing you gotta do.

Two: delete the part, process step, or person. Try really, really hard to delete it. Elon always says this, and I very much believe it because I've seen it happen time and time again — the most common mistake he sees a smart engineer make is optimizing a problem that shouldn't exist in the first place. The best way to think about this: our education system is built around being given a test with questions. You have to sit down and answer the questions and turn your test in. You're not allowed to say, "This is the wrong question — this is not what I was supposed to be asked. You're asking me fundamentally the wrong problem." As engineers, we're trained: I'm given a problem, I've got to solve it. You really gotta think hard — am I solving the right problem? Because you might be optimizing something that you shouldn't be putting effort into.

Three: optimize or simplify. Really important — you run steps one and two as many times as possible, and then you jump to step three. You don't go to step three until you've iterated a couple times. A common misconception in engineering is that something complex is more reliable or better. It's false. Oftentimes the most simple thing is the best design and the most reliable. Complexity, we like to say, is the devil — it oftentimes leads to a much less reliable product, let alone something that you can make many times. It's easy to build one rocket, very hard to build many rockets. Easy to launch one rocket, very hard to launch a lot of rockets. Doing something at scale with technology is much more difficult. So spending a lot of time optimizing, simplifying, and deleting parts from your product is critical.

Four: accelerate. You can always go faster than you think you can. There's always time to gain, there's always efficiencies to be found. You think you went fast enough? You can go faster. I've seen it every time — my team's like, "Dude, we can't go faster." Yeah, you can. Move the goalpost. Be surprised what happens when you challenge people.

Five, last but certainly not least: automate. Automation is an extremely powerful tool, but the reason it's last is that if you don't run through the algorithm, you can end up automating something that's fundamentally more complex than it should be, and that's going to slow you down in the long run. So many times people make the mistake of going to step five before running through step one, and that's when you end up putting a ton of effort into automation before realizing, "Holy smokes, I don't even need this thing in the first place."

So let's do an example of how we apply the algorithm at SpaceX.

[video plays]

Here's a story behind these fairings. Elon challenged us one day — because building a fairing is a fairly complex task, costs about six million bucks. He's like, "Look, guys, imagine you had six million dollars falling from the sky. Would you try and go get it?" And we're like, "Yeah, dude, obviously we're gonna try and go get this money." He's like, "All right, then go get the fairings."

[laughter]

At the time, we thought we had to catch them with a net. We can't take this really complex piece of technology and let it fall in the ocean, get super corroded with salt water, and then have the parts not work, right? And it worked — we did it. You basically have this awesome algorithm: the fairing would fly with a parafoil, steer itself, and the boat would have automated control that would turn and follow, and the two would close together — that's how you'd capture them.

But what do you guys notice about this picture? This is like a lake. This is the Atlantic at the most calm it could ever be. The reality is, most of the time it's a choppy hot mess with seven-to-nine-foot waves with super short period and a ton of wind. So even though we caught it once, our actual success rate for bringing fairings home was quite low — under 50%, around 40%. And our ability to get fairings ready to fly again was choking our launch rate.

Now, in that process, what we found out was fairings actually float pretty darn well. It's composite structure — sailboats are effectively composite, it's really just a big boat. And in the process of discovering that, we're like, well, do we really need to catch them? We challenged our requirement. You don't need to catch the fairings — they float well. If we just move some of the parts to the higher part of the fairing, even if a little water gets in, it's gonna survive, and it's ultimately going to make it much easier.

So we basically said, we don't need to catch them. We can completely delete the boat, the big net, the automation — all of that stuff. We allow the fairing to just fall into the water, we use a very regular standard crane that comes on one of our recovery vessels to pick the fairing up, put it on the boat. And what happens? We go from less than 50% to a 99% success rate on fairing recovery. We have more fairings than we have space. We built an entire facility such that a fairing can come in, spend a minimal amount of time being turned around, go back out, get on top of the rocket — and fairings are a thing we don't even come close to talking about when it's time for launch. They're always ready, no problem.

Now that's a good example. I didn't talk about automation much, but there's one really cool piece of automated equipment. The whole rocket flight is automated, but when you want to land the vehicle — a lot of people think going to space is just going up. But that's just getting to the Karman line. If you want to go to orbit, it's actually about going up and going really fast, because you've got to get going 17,500 miles per hour and be up out of the atmosphere so the drag doesn't slow you down.

The way you catch a rocket — if it's a pretty light payload, you can have enough fuel to basically turn the rocket around, boost back, and land. But if you want to maximize the amount of stuff you're putting in space, you're going to use as much fuel as you can to get it where it needs to go, and then let the rocket fall. Because you launched over water to ensure you don't overfly people, you need a way to catch the rocket.

So we created the world's largest autonomous vessel. This is basically a river barge that we took, put a bunch of thrusters on, put our own software and systems on, and created "A Shortfall of Gravitas" — a completely unmanned vessel. There's no one crewing this at all, and it's able to hold its position. The Coast Guard won't let us do this yet, but it could actually just drive itself out to the landing location, let the rocket land — we have this cool little robot that comes out, grabs the rocket — and then drive it all the way back to port. We end up using it semi-autonomously, but this is one of the coolest pieces of tech that I think we've done.

You think rockets, you're like, oh man, you fundamentally changed the rocket game. But it's cool — this is why I like to say we're an infrastructure company. The drone ship in itself was a piece of innovation that without, we wouldn't be nearly as successful.

So why does this matter? Increasing launch rate enabled two primary things. The first was the ability for us to build a broadband satellite internet constellation that could give you internet anywhere on Earth — internet from space for humans anywhere.

The Antarctica picture is especially special to me. I actually got to go to Antarctica, super lucky, in February. I have two little kids — a one and three-year-old — and the ship was like, "Do not plan on using any internet." I was like, "Nah, I got you." I actually brought a Starlink terminal with us, set it up on the back of the boat, and everyone was able to use it. We were able to FaceTime our kids, which was really special.

I've worked on a lot of cool things at SpaceX — human spaceflight, I've had the opportunity to really be on the forefront of a lot of the technology. Building Starlink was the first time I felt like I was really changing the lives of everyday people — giving access to education, to healthcare, allowing people to visit their loved ones if they're sick and they can't go to the hospital. The stories you hear just make your heart so warm and make you so proud that, man, I didn't just build something for the government or something like that — I built something that's changing people's lives. We have a pretty cool video that shows some of those examples.

[video plays]

A lot of people don't realize this, but this little dish sits outside — snow, mountain, small frost, high winds, you name it. "Myself, that roaming — got my pole-mounted Starlink here." "Starlink unboxing — satellite Wi-Fi right here. We're going to be using it in emergency situations when people need to communicate, hopefully in the middle of nowhere." "I'm playing on my computer off of a battery, online on the Starlink, dude!"

People are wacky when it comes to getting their internet. It's amazing.

The other thing that was really awesome is our flight rate has really helped us pick up human spaceflight in this country. We're just a little under three years from when we were able to return humans to low Earth orbit. Most people don't know — in 2011, the Shuttle retired. Between 2011 and 2020, America was paying the Russians a little over $90 million a seat just to get us to the space station.

SpaceX won a contract, put a lot of focus into developing a 21st-century spaceship — Dragon — and on May 30th, 2020, we launched Bob and Doug, which was the start of our human spaceflight program.

[video plays]

"Welcome aboard Dragon. SpaceX, com track, ground stations. SpaceX Dragon, go for launch. Three, two, one. On behalf of the entire launch team, thanks for flying with Falcon 9 today. We hope you enjoyed the ride and wish you a great mission."

"We are proceeding toward the primary landing site and your timeline is current. Visual confirmation — Dragon for splashdown. Endeavor, on behalf of the SpaceX and NASA teams, welcome back to Planet Earth. Thanks for flying SpaceX."

[applause]

So awesome. Bob and Doug were really — like, flying those two — we always like to say, "dads in space." They're both like two dads. It was a really special and awesome moment to get them to space.

But the thing that's really neat right now about Dragon is that it's not just trained professionals anymore. We've really gotten to the point where private spaceflight is taking off at a level that we've never seen before. I think we're really close to being exponential on that curve.

This mission I think is still one of the coolest things we've done. This was the Inspiration4 mission. It launched September 15, 2021. It was a project funded by a guy named Jared Isaacman — he started the company Shift4. He generally has a passion for doing cool things, and one might think most wealthy people that want to go to space — kind of like a bunch of white dudes on a capsule going for a joyride. Jared had a much bigger mission. He wanted to raise money and awareness for children's cancer research. He basically wanted to partner with St. Jude's and see if this mission could generate funding to help change the lives of children dealing with illness.

So he put on this Inspiration4 mission and picked a crew of just regular humans. Jared himself; Hayley Arceneaux, the woman in the middle — she is the first cancer survivor to ever go to orbit, also the youngest female to go to orbit, a really badass girl from Tennessee; Sian Proctor, Dr. Proctor, a scientist and artist from Arizona who once wanted to be an astronaut and tried to apply to the candidacy but never made it — she won her seat through a competition with Shift4; and then Chris Sembroski, just your down-the-middle average American guy who didn't actually win the sweepstakes. His buddy won — it was like a sweepstakes where you pay $100, you get a number of entries. His buddy won and was like, "Dude, I don't want to do this, but do you want to go to space?"

[laughter]

Imagine getting that phone call from your friend — "Yo, look, I just won a trip to space but I can't go. You gotta show up to this thing. I think it's gonna be cool." And Chris was like, "Yeah, yeah, okay, I'll do it."

They put this mission on — these are not trained professionals, these are everyday humans — and they raised $250 million for St. Jude's and children's research, which is awesome.

[applause]

And the next video is really awesome. You saw Bob and Doug in action — here's what happens when you take these four civilian astronauts, put them in the capsule with the largest window we've ever built in space, and they open it up for the first time to look at Earth.

[video plays]

"Holy — oh my gosh!"

[applause]

So that was really special. Jared has a follow-on mission called Polaris Dawn, planned where we're going to do the first planned spacewalk with Dragon. I'm not gonna lie, this picture makes me nervous. It's like — who's been skydiving? You know, they open the door and you're like, "I don't know if I want to jump out of this perfectly good airplane." Same kind of vibes. But I'm confident we're gonna get it done safely, and it's going to be really special. There are some really exciting things going to come around this mission.

One of the last points I want to make about this vehicle: private spaceflight is here now. This is not something that's far off in the future. And I know this is generally a bunch of successful people, so if it's something you're interested in, we should talk about it. I think you're gonna see private spaceflight really take off over the next five to ten years. If you're inspired by it, you really should consider an opportunity to go to orbit, look down on Earth, and just realize how precious the place is that we have.

I don't know if you guys have heard of the overview effect, but it's a feeling that most astronauts get when they leave and go to orbit or beyond low Earth orbit and they look back at Earth. You realize just how fragile this one little planet that we're all living on is. Mothership Earth is actually a spaceship hurtling through space at millions of miles per hour through the universe.

But if you want to go to Mars, you're not going to want to take the minivan — you're going to want to take the really sweet tour bus, which is Starship. Dragon, Falcon — evolutionary technologies. We throw that second stage away. We're designing Starship to be fully and rapidly reusable. The vehicle launches: both sections — ship on top, Super Heavy on the bottom. You launch, booster comes back, ship comes back, you refuel, and you go again. It's all about making spaceflight like aircraft flight. Without that step function, we're never going to be a spacefaring civilization. So it's fundamental that we get that piece of technology right.

On the way there, though, there will be some excitement.

[video plays]

[applause]

It was exciting. As most people know, for us, the way we approach technology — we want to learn and we want to fail fast. We want to go as fast as possible. People don't understand the scale of this vehicle. It's effectively almost the size of the stage just in diameter. It feels uncomfortable standing next to it. It's the world's biggest rocket by far — bigger than Ariane, bigger than the Artemis rocket that just launched, bigger than Saturn V. A lot more thrust. But it's what's necessary to drive the economies of scale low enough such that we can truly push exploration farther.

We gotta break that barrier. We gotta go back to the Moon. We did that in '69 — it's about time we put humans back on the Moon and help deliver that future where we can go a little bit farther.

A lot of people ask me, "What about Earth? We obviously have a lot of problems." There's plenty of problems and there's plenty of talent to go around for both these things. I truly believe, and we truly believe, that a future in which we're a spacefaring civilization is an exciting one. And I actually think it's one in which we've taken care of our own planet. We're in the small window of time where we're actually going to be able to do this, and I think if we accomplish this goal, it actually means we've also solved our problems on Earth.

So it's late, it's Friday night. We're gonna make life multi-planetary. Q&A tonight — come find me later if you want to talk. Thank you, guys!

[applause]