Inside The Hard Tech Startups Turning Sci-Fi Into Reality

You actually can make some significant progress with like half a million dollars in 3 months. The best hardtech Founders do have very high clarity of vision around the future for hardtech companies. You have all this tactical risk; you don't know if you're going to be able to mine asteroids. But you have no market risk, right? Like, if you can mine asteroids, it's going to be freaking huge. I think this is a call to action for a lot of the very hardcore engineers that this is your chance to build something huge and really change the world, literally.

So, this is the kind of energy we need in society—solving some of the biggest problems that face humanity. Welcome back to another episode of the Light Cone. You probably know Y Combinator for a lot of our software investments, like Stripe, Airbnb, or Coinbase. But it turns out YC has a lot of wins, not just in the electrons world where you're slinging bits around. Some of our best companies have actually been incredible at slinging atoms.

So, I'm curious—what's some of the advice you give to the hardtech companies in a YC batch? Like, in particular, how is it different the milestones they need to hit during the three months and what metrics they might present on Demo Day? Well, one of the obvious things is they won't be able to build a rocket, let's say, in time, so they won't have actual revenue. One of the things we tell them is they still have to show some form of commercial attraction. It won't be necessarily actually selling and getting revenue in the bank, but it's a different kind of form.

I think we talked a lot about demonstrating that the customers that ultimately they will sell to want this. And typically, a form of an LOI that is actually significant with actual values—with actually legit logos. I've definitely found LOIs, like letters of intent, are something where everybody knows they're obviously not the same as actually signing up a customer. But for the software companies, if a software company says, "Oh, we got a customer on an LOI for a $1,000 MR contract," like, that's not going to be impressive. But if you can get an LOI for like a $100 million contract, that is actually quite impressive.

The way that people tend to get this wrong out there in the world is they're building some hardtech thing, and they're like, "Oh, well, I need like $20 million to build my thing." And then they look at YC, and they're like, "Well, YC invests half a million dollars. Well, that's not enough! I can't do YC! How would this make sense at all?"

One of the interesting things that I've learned from working here is that actually YC works extremely well for hardtech companies. For every hardtech company, no matter how crazy the thing you're working on—whether it's a supersonic jet or inventing fusion power—there's always, like, some small part that you can peel off that is like the very initial stage that you can actually make some significant progress with like half a million dollars in 3 months. Like, always.

Yeah, my experience and I feel this is familiar with hardtech companies is that they start YC and they always come in with this—the essential pitch is, "We need to raise like there's no way we can make this company work unless we raise like $50 million." And so, like, hard, can you tell me how to raise $50 million this week? Introduce me to some VCs who are going to give me my $50 million. I like to link is always no. That's your first office hour—the first off show me the money. Yes, basically it's like no, you're not going to be able to raise like $50 million off the bat. You need to think like a software company, actually. Like, how can we do this cheaply, quickly, and show something on Demo Day?

I just find it's always actually, at the start, it's a bit of a fight honestly to get them to think that way. In the second half of the batch is when they finally get some insight 'cause we've just been like pounding on them to think fast, cheap, less money. And then something clicks, and then they actually have a plan for the second half of the batch. The remarkable thing I found is actually how well it works that once they make that mental switch and they realize that they can do something in three months, then they go do that.

Then they have, it's not just an idea anymore. They've got something real and they can go to investors who maybe didn't want to fund them before YC. And now, all of a sudden, people just want to put money into this thing because they've actually, like, it's actually like going someplace.

The weird inversion for hardtech, I think, is it's not about the why. The why is often there's a problem, and it's sort of obvious. And then the why now even is often kind of obvious where there's some cost curve. There's some bomb costs that changed. The build of materials got cheaper; there's some breakthrough.

And so the actual question is not why now or why. It's actually why you. And I think that links up to what you're saying, which is if you can show that it's you—and it might be LOIs— in hardtech, especially the people who might spend a lot of money on atoms, they tend to be a little bit more diligent about only signing LOIs with people who are for real. But as much as the LOI, I think the other piece is they actually need to build a plan and demonstrate that the cost actually makes sense and have that whole financial viability aspect of it.

I think the other thing is actually, besides LOIs, is proving that they can actually make it. Because a lot of things, like the why—to answer the "you," can you show me a kernel of truth of the idea that you actually can build? So besides the LOIs, the second thing is a technical milestone. And of course, you can't build the whole plane, let's say a rocket, or build the giant decarbonization container thing. But maybe do it on a small scale. Maybe if you were aiming to capture at some point, I don't know, 700 metric tons of carbon, maybe just do it like one or a hundred grams, or seven grams, right? Demonstrate that you can actually do and know the science and can actually build it.

Yeah, I have a trick for helping the hardtech founders present this to investors that I stole from one of our software companies. So, let's start with the software company—it was Brex. So Brex's Demo Day presentation was a lot about showing how they were able to move really quickly on getting a banking deal and sort of the regulatory aspects of starting like a fintech back then, right? And as I remember it, their presentation had like a timeline of here's how long it takes like a regular company to get the banking deal and go live with like a real credit card, which charge card. And then they were like, and we did it in like 3 months, whereas like the actual time was like 12 months.

So I tell hardtech founders to think about it as like what's like the timeline to build your product for like a regular company? Like maybe it's 12 months. And would it cost like a regular hardware company $50 million and 12 months to get to like this proof of concept stage? And prove that you were able to do it in like 3 months for $500k? And like that visual I find is always effective for answering the "you" question.

Actually, oh, okay, there must be something special about this founder if they're able to do this like with less money and much faster. The cool thing is that if people can build that habit during YC, it doesn't just help them raise money during Demo Day, but it actually creates a habit that then changes the operating cadence that they work at forever. A lot of their best hardtech companies, they still have that way of operating which is they're constantly pushing to do things faster and cheaper than anyone thinks is possible.

Should we talk about some examples of companies that have gone through YC and done this successfully? Yeah, and I think for each of these, the idea is we're going to talk about where they are now, and then we're going to talk about what they were able to do during the three months of YC. Yeah, that's cool, yeah, let's do it.

So I thought it'd be fun to start with Boom. Boom is a company that we funded in the Winter '16 batch, and they are building a supersonic jet—supersonic passenger jet to replace the Concord. Because this is a very strange area in which humans have gone backwards in terms of technological progress. Like, young people watching this might not realize, but like regular people used to regularly fly supersonic from New York City to London in like two hours at like Mach 2 point something. And then we stopped doing that because the Concord was discontinued.

Blake, the founder of Boom, he just thought that this was unacceptable. I remember when Boom was in the batch, they had this crazy idea to build a supersonic passenger jet which would cost literally billions of dollars to do. And so, like, the question would be how could you possibly make any progress on this insane idea in three months and a few hundred? He did two things during the batch: He tried to de-risk the technology and he tried to de-risk the commercial interest.

And you might think that it's inevitable that there would be commercial interest in this because, like, of course everybody wants a supersonic jet. But the thing is, you can't sell a super sonic directly to people; you have to sell it to airlines. Which means this is only viable if an airline actually wants to buy the thing. And there was a legitimate question as to whether airlines, which are capital-constrained and risk-averse, would actually want to take a bet on a crazy new technology like this that'd be very expensive.

And so the way Blake de-risked it during the batch is he basically spent the batch trying to get an airline to make an early bet on Boom. And at the very end, just a few days before Demo Day, he finally was able to get introduced to Richard Branson from Virgin Airlines. And Richard Branson gave him a $100 million LOI to buy the first Boom planes.

Yeah, and that was absolutely critical to his ability to raise money because he showed that the deep-pocketed customers that would ultimately have to place a bet on this were excited enough to sign an LOI like that. So, Jared, where's Boom today? So, the really crazy thing is eight years after they went through the YC batch with their like Styrofoam scale model, they've literally built a supersonic jet, and it just took its first flight last week, and it worked.

Wow! For those of us at YC who've been part of the Boom story since the very beginning, with like the Styrofoam scale model and everything, seeing them actually build and fly a supersonic jet was just like an unreal, like otherworldly experience that they actually pulled it off. That's so cool! So, you can just literally think things in your brain, yeah, and then manifest them in reality.

Gary, do you want to talk about one of YC's first earliest and most successful hard tech companies, this company called Cruise? Yeah, of course! So, Kyle from Cruise, of course, was one of the early co-founders at Justin.tv, which became Twitch famously. He was the guy who figured out the GPRS and Edge modems in a backpack, so Justin could wear the little camera on his head. Kyle worked on the hardest technical problems of Twitch, which is with a lot of the video streaming and encoding, where he made it so efficient that to this day, Twitch runs their own video servers, is what made them profitable too.

Yeah, so that was like the key thing that he unlocked. And just for a bit of background, he's like this technical genius, basically! Oh, definitely. It turns out I remember interviewing him for YC. The interview room was Sam Altman, myself, and Jeff Rolston—incidentally, the subsequent presidents of YC, which was kind of funny. And I remember it was a very short interview. He came in and said, "Well, I worked on the MIT autonomous car team. I really like that, and I think that it's possible for us to make a commercial version of this." He left the room; it was a great interview.

We looked at each other and said, "Well, we have to do it because it's Kyle!" But it's probably a research project, and our laptops went to lunch. And of course, it became the biggest, fastest liquid exit of almost any Y Combinator company at the time. So, how much was it sold for? Almost a billion dollars, which was really wild. And just how long after you interviewed him?

I mean, it was a matter of a few years. A few years, like two years maybe from interview to a billion-dollar exit, pretty amazing. It was even more than any other company at YC at that time, even the software companies, because companies like Dropbox, Airbnb hadn't exited until later after their acquisition, right?

Yeah, and I think one of the things that's kind of lost to the sands of time now that you can literally walk down any Street in San Francisco and there are self-driving cars is that this was not something at the time that people thought was even commercially viable. It was sort of locked away; we knew about the Google X project for self-driving cars that would end up becoming Waymo. But it took, again, like someone with an idea in their head deciding, "You know what? I'm going to work on this thing." You can actually manifest things that, frankly, are on their path to remaking every city in the world, and we're still in the early innings of that at some level.

Do you remember what Kyle had at the start of YC? He had a control system for an Audi S4, which was fortuitous because I was driving an Audi S4 at the time too. And I think, think to the commercial point—we actually, this was a little bit during the era of the Kickstarter campaign. So, that was actually what helped him raise his seed round was getting commercial validation that, you know, we didn't need to build a full self-driving car on day one.

We’re going to build a highway ADAS, just assisted driving as a retrofit for Audi S4s. Okay. And that was sort of the initial thing that during YC he did. But, of course, once he got those resources, and he ended up raising a Series A, he could expand his vision significantly.

What did he pitch on Demo Day? Like, it was actually, I believe, the retrofit kit for Audis. It actually—that was smart because, like, if you had pitched full self-driving, it might have been hard to raise money because people would have been like, "Well, how are you going to get the $10 billion that you need to actually make this real?"

And so the fact that he had a shorter-term path to commercialization on like a normal amount of venture dollars was actually like a really good way to bootstrap the big long-term ambitious idea. And so, was the pitch that they'll take these like assisted driving kits and sell them to Audi? No, I think sell them to individual Audi owners.

Yeah, so I remember putting down my credit card for one of these things. I didn't end up getting one, but if you visit the Mountain View campus of Y Combinator, there is actually one of the original retrofit cars. Oh, interesting! And so, I think that that really illustrates exactly, you know, commercial validation with, you know, basically fantastic "why now."

And then they proved to the world, like, you know, the "you" and two different types of commercial validation. Interesting, like Boom was like, "We're going to convince a huge airline," and it probably seemed quite scary right up until Demo Day on whether they were going to get anything or not. And then Cruise is like, "We're not going to sell to airlines; we're just going to go directly to individual who want this," which was probably much more like a bite-sized chunk approach. That's what it takes to sort of create some—basically create a new industry from scratch.

Another way hardtech companies are viable is they're going after a known commercial space, but doing it a lot cheaper than the current incumbents. And for example, launching satellites to space costs like billions of dollars. Because there's like these giant satellites that take multi-years to assemble—tell us a bit about Astronis. They had a technical innovation as well.

Yeah, a lot of people don't realize but YC has actually funded some of the most successful space companies in the world. And one of the first ones we did was this company called Astronis. Astronis builds telecommunication satellites, and their core insight was that you could make satellites a lot cheaper if, instead of making a few big satellites that did everything, you made a lot of small satellites that just did a few things. It's kind of like when commodity servers replaced mainframes—it's basically the same concept for satellites.

And they were in the Winter '16 batch, the same batch as Boom, actually. Their goal for the batch was to build a fully functioning satellite that they could put into space in three months, which is like the hardest part about like typical timelines. Like the typical timeline to build even a demo satellite is years, and they were like, "We're going to build it in 3 months and then we're going to launch it into space." Because by that time, SpaceX was offering space flights for small cargo, and so you could actually book basically a one-way ticket on a SpaceX rocket for like not that much money for like a small CubeSat.

And so their Demo Day photo, you can actually see the satellite that they built during the three months of YC—there's Ryan holding it up. They actually launched it after the batch. They actually did? Yeah! The cool thing about Astronis is that telecommunication satellites are actually a really good business. Like, it's actually like super profitable.

And so from pretty early on, they were able to generate a real revenue from customers and actually make money putting satellites into space. And they now actually have several satellites in space over our heads, and a really cool chapter of the story is that they actually are in like the Astronis office just across the street from the YC office.

And it's not just like some headquarters place where they like designed the plans for the satellites; they actually manufacture the satellites across the street from our office, and we can actually see it from our set right here, it's right there! And so I love to take the founders in the current batch on a tour of the Astronis office across the street and show them that you can go from your Demo Day CubeSat to like a satellite factory in a few years.

The cool thing about space is, you know, partially because of SpaceX and a lot because of the massive increase in launch capability, there's just this ecosystem now. Kind of like Astronis is a great example of it that now that you can get to space, there are all these things that you can actually do in space that are very valuable.

And there's a really crazy one that you funded, Gary, when you were at Initialized called Astro Forge. I love to hear about that one. Astro Forge is really interesting because they have a pretty scary huge ambition, which is literally to be able to fly a satellite to an asteroid. Some of these asteroids have something like 15,000 times more concentrated precious metals like platinum than on the earth. And then they're actually using robotics to refine the ore directly on the asteroid, and then take that asteroid and fly it back.

And the funniest thing about it is you don't actually have to land it. You could just fly the asteroids straight into the desert and then mine out of the crater and, you know, potentially mine hundreds of millions of dollars or billions of dollars worth of precious metals. So, are they going to try to bring the whole asteroid back to Earth or just like a little chunk of the asteroid? Just the precious metal itself? Just the precious metal.

Okay, they're early in their voyage, but I just really love extremely ambitious people. I always talk about, especially right now because we're prepping people for Demo Day, the best pitch is relatively small, simple ideas when you put them together and you zoom out, like that's actually a really big story that's totally achievable.

And so I feel like Astro Forge is one of those examples. As an investor, how do you think about the risk of investing in a company like that? They're a software company—it’s sort of, you know, right? Can you get customers? What's your sales plan going to be? But mining precious metals on an asteroid in space and bringing it back safely? How do you think about the risk?

I think some of it is all they have to do in the shorter term is actually fly to an asteroid and come back. And ideally, it would be great if they could prove that they had picked— they were able to pick an asteroid that actually had precious metals in high concentration. And there's a weird interesting regulatory aspect to it too in that I believe the United States has actually spoken to this, saying if you land on an asteroid, you actually confer ownership rights to it.

So there are many ways to monetize, and I think that speaks to what you were saying earlier. We want sort of achievable tranches that are not outrageous. You don't have to do self-driving cars on day one. You don't have to actually fly back a billion dollars worth of platinum on day one. You have to show that there is a clear path to build the tech to get there.

So, Relativity Space was also in the Winter '16 batch. Amazingly, we actually had three billion aerospace companies in the same batch. Is that wild? Yeah! Wow, Winter '16—that was a special time. And I think the reason is what Gary was saying, which is like this confluence of factors. Like, 2016 turned out to be an amazing time to be starting an aerospace company.

And Relativity Space, they make 3D printed rockets. And the founders were super young—they were like 23 years old when they started it. Basically, their Demo Day goal was to prove that they could 3D print a rocket engine—not one that you could actually take into space; that's impossible—but just like a scale model to prove that the thing was feasible technically.

And I remember Jordan him walking around Demo Day carrying this like rocket engine, and it was small but it was a real rocket engine. It had all the injectors; it had the right nozzle design. Like, it was actually a thing that you could like fire up and would produce thrust. It’s such a wacky idea! Who would have thought that it is possible to 3D print a rocket to space? Because there's so much of the tolerance with heat dissipation and energy, and to get everything at the right manufacturing is very like super high precision.

And the cool thing is they actually did it! In March of last year, they actually launched a full-scale rocket that was almost entirely 3D printed. It's the first time anyone's ever done that! Wow, that's so cool!

So, we talked a bunch about aerospace companies; another area we've invested a bunch in is climate and energy companies. I think you worked with Heart Aerospace, right? Oh no, that was Gustaf that worked with it. Ah, okay. So, he worked with Heart Aerospace, which is a Winter '19 batch company. What they're trying to build are fully electrical planes, and that sounds doable, but it has not been done.

Part of it is that the batteries in planes are too heavy and don't have enough range. And what they figured out is that a lot of regional flights are the sweet spot because actually today, with regular planes, fuel-based planes, regional flights are actually losing money, and they're subsidized by the government.

And what they figured out is they're going after the market, and during the batch, they signed a bunch of LOIs with a lot of airlines because this is actually a burning problem for airlines. They're actually losing money in all these flights. And the other interesting thing on the why for them—they're actually located in Sweden, and Sweden and Nordic countries are one of the most progressive countries in the world.

There's a goal for them to fully electrify all flights by regulation by 2040, which is like a huge thing. Talking to the why now and making the pain even bigger, what they achieved is actually, it's been four years later—they actually built a version of it. They even have a test pilot. Wow, that's a big plane!

Yeah, I think it's a 19-seater. Wow! Their first plane is going to be a 19-seater, so they actually have built models for it already. That's pretty cool to see, right?

And the other cool thing about hardware companies as they progress is you need a lot of funding to get this going, and VC funding is not going to be enough. You have to get customers to pay you for it. So, they actually got purchase orders from United, Air Canada, and other regional airlines locally, as well as government grants.

Remora is a very interesting company. It is another of these wacky ideas that is trying to retrofit semi-trucks to be carbon neutral—to sequester all the carbon while they're moving, which sounds like wild. So, how do you do that with a truck? Because truck emissions in the US is about 3% of the total emissions. The US emits close to six million tons per year, and they could capture and make a huge dent.

So, the way they went about it was first principle. Paul was very interested in working in climate, and he studied and looked at all the emissions in the US. And the top one was transportation. Then he found this thesis from Christina, who is the world expert and published this PhD in 2019 on mobile carbon capture for heavy-duty vehicles.

It was like the only thesis in the world, and he basically got in touch with her and convinced her to be his co-founder. Wow, and it is a wacky; there's nothing built like this in the world. But you see now where they are; they actually built it! You see these giant tanks that are actually capturing right now 80% of what gets exhausted in a truck.

Now, Cound is this other company, sort of like Remora. So, Cound is basically retrofitting cargo ships to reduce CO2 emissions, and they're the only solution today that's possible. And the other thing to why now—there's actually regulations that came up that are forcing cargo ships to meet carbon goals.

The cool thing about them during the batch—I worked with them in Winter '22—is they were able to close a bunch of LOIs with ship owners, which is a really hard industry to break into. And now, after the batch, they actually just had their first pilot here. As you can see, they did their first… this is them on an actual cargo ship! Yeah, just this year! An actual containership! Well, last year actually, yeah.

So, this is the kind of energy we need in society. You know, if you're a top 150 IQ or above person who wins math Olympiads and things like that, instead of going and optimizing ads at Facebook, why don't you go work on something like this? Yeah, solving some of the biggest problems that face humanity—which is pretty important for people watching to know.

To whom much is given, much is expected. So, Jared, we talked a lot about hardware type of companies. What about chemistry? So, I see, funded this company called Solugen that I worked with in Summer '16. And I think Solugen is one of the coolest companies I've ever had the experience of working with.

And what they do is they make industrial chemicals. They started with the smallest possible scale that you could imagine making these chemicals at, and then they just scaled up in like successively larger and larger scales until they ended up where they are now with this massive chemicals plant in Houston.

And I remember when they applied to YC, what they literally had was a beaker like this big, and they had made like one beaker full of hydrogen peroxide just to prove that they could do it, just to prove that the thing worked at all. And then their Demo Day goal was basically to go from like a tabletop scale to a garage scale.

And they literally took over their garage—I think it was in like Mountain View—and they built this garage-scale production platform where they were able to make like gallons of hydrogen peroxide. And one of the coolest things about Solugen is, unlike a lot of the earlier chemicals companies that were venture-funded, that tried to raise a huge amount of money before selling anything, the Solugen founders started selling their product during the YC batch.

So when they were making like a few gallons of hydrogen peroxide, they would literally go out and they would sell it! And so they were revenue-generating literally from day one, and they never sold the product at a loss. They always figured out some way to sell like small quantities where they were at least not losing money on the actual thing that they were making, and as a result of that, they've actually been really capital efficient for such a company like this.

And today, they have a really healthy, revenue-generating business that makes sense, as if you don't look at it as some like sci-fi startup, it's just like, it's a thing that actually makes a lot of money. That's awesome!

I actually have a couple of companies in this current YC batch that I wanted to talk about because they build off some of the lessons that we've just been discussing about all these companies, right? So the first one is K Scale Labs. And so K Scale, the vision is to build consumer humanoid robots. Very cool!

And Optimus for the rest of us! Yeah, basically. And you can like all kinds of ideas around this, like maybe in the future we'll all have a robot that's like cleaning our house or it's… I want one that does all of your chores for you, right? What's interesting about it is it follows the pattern we discussed earlier of where Ben, the founder, came in and said, "I have to raise like a huge amount of money."

And then there's a competitor in the space that has just closed this huge round of funding because the founder is one of like—he started successful companies before and he's probably just like a natural fundraiser and very well connected. Right? And the first few office hours were all about, you cannot—like, you have to think like a software company. Like, how can we build this company and prove stuff out without raising like a $100 million Series A?

It took a while to figure out what that would actually be, and eventually, like about halfway through the batch, we found it. And the answer is like what K Scale really wants to do, what they're really excited about is building a new foundation model for perception in robots.

And this is what the founder, Ben, had worked on at Tesla, actually! So he was in Optimus. Yeah, he put the first perception model into the Optimus at Tesla. But in order to build the best model, you also need it running in real hardware, right? Like, you need actual robots out there that you're gathering data back from to improve the model.

And so the idea he came up with is they're going to build the first 10 of these human robots themselves, so it's pretty low scale. And they'll run their models in those, but they're going to open source the hardware designs. And what they found is that there's a lot of excitement amongst the engineering community to build their own humanoid robots. People just need the design, and so you open source the designs, you build a community of people who already want the how-to kit on building a humanoid robot, and all of those robots will run K Scale's foundation model on them.

And this is sort of like the plan for how do you get to—it's like crowdsourcing the costs to build them. That's a very clever hack! Yeah, I think so, right? And it is a work in progress, and we'll see how it plays out. But what we do have really strong early signs of, and people want this.

Like, Hacker News is a pretty common thing. People talk about, "Hey, I really wish I knew how to build like my own prototype humanoid robot." And soon, they'll be able to. If you look at early interviews with Steve Jobs and Steve Wozniak, they talk about how they never wanted to start a company; they just were selling breadboards and plans and the Apple One was a bag of chips hobbyist.

It's actually very similar to this. Yeah, and explicitly, Steve has gone on the record saying, "We never wanted to start a company. We were just sick of building these things for our friends, so we thought we needed to start a company." And so it sounds like Homebrew Computer Club again, but for robotics. This is exactly that. It is exactly the same pattern.

And Diana Canest, it's the same sort of founder profile—somebody who is just very enthusiastic about this technology and what it can do for the world, and like I feel like the help we've been able to provide is how do you funnel that excitement into a path that shows commercial viability within 3 months.

Another example would be Astromechanica. It's another company in the current YC batch, and so they've created actual real tech breakthroughs. They built an electric engine that is efficient—electric jet engines that are efficient at every speed, which is incredibly hard to build, right?

Because if you think of current jet engines, for every speed, it's a totally different set of optimizations. Like you have different amounts of air coming in and out; you have like different compressors. You need to build—you just can—you make a fundamental trade-off to be efficient at any specific speed. So it's an electric jet?

So is the idea that it's going to power fully electric jets? Yes and no. Like, the actual—like, the thing that's exciting about Astromechanica is if you have an engine that's efficient at every speed, you can use it to power different types of aircraft. So, you could have like a subsonic aircraft that's just much faster than a traditional jet engine aircraft, or you could also use it to power potentially like a supersonic aircraft—all with the same engine, right?

So, like the actual plan, the big picture vision for Astromechanica is to replace Boeing and to be like the next Boeing with like the engine as the core of it. And it seems like we might need a next Boeing, so timing is good!

Yeah, it's time. It's definitely timely with the news, right? But they did a lot of simulations, right, and software to be able to really prove that they can run this engine at multiple speeds, which is like a huge technical accomplishment. I asked Ian, the founder, on that—like, how have they been able to think like software companies and just move very quickly on the technical front?

And he said one of his key insights was that you want to innovate on as few things as possible in the hardware. You want to try and buy as much off-the-shelf hardware as possible and just pick the one or two things that you actually really want to innovate on and, like, put all of your energy into those, but have as much like off-the-shelf components as you possibly can.

And then on the commercial side, what they've done is like, you have this big vision, right? Like, we want to compete with Boeing; we've got this brand new engine that could be used for like a thousand different potential things. But instead of trying to build, for example, a commercial aircraft right now, they're focusing on one very specific use case, which is just launching payloads into orbit.

So, if you have this engine that's efficient at every speed, you can launch things into space far more efficiently and cheap. And so they're going to focus on just that. They've already signed up LOIs for quite a lot of money and used that to like fund their future plans, right?

Which is sort of—it’s like the Tesla strategy where it’s like you start with the Roadster and you use that to fund the development of like the Model S and use that to fund the development of the Model 3. I think the cool thing about a lot of the founders we highlighted is that I think they've been building a lot of the tech on these products in their head for a long time.

I think the thing that stands out to me a lot of times is reading a lot of these applications for the founders. They're actually super well thought out, and a lot of the science and engineering is exactly as it plays out as the company year one, year two, year three. Even for Astronis, just reading their application, it is what they have built right now. They do the six-panel design, and it is what they have!

And I was thinking SeaBound—the design of doing carbon capture with calcium looping—it is what they've done at a large scale. And that's the cool thing is I think a lot of the risk here is actually more how do you take that first step? And I think you said something cool, was that secretly, a lot of hardware companies actually think like software companies.

Yeah, I do think that the best hardtech Founders do have a very high clarity of vision around the future. And I think where we help them a lot is on the sort of the compressing the timelines and like think about things in a more commercial—like how can you prove that people will pay for this? But the actual—like, they live in the future already somewhat, like they already know where they want to get to.

And I think that is probably different from many software companies, where you can have a fuzzier vision of the future and iterate and figure it out as you go along. Whereas with hard tech companies, like you kind of—you often have the vision of the future in your head, and it's just proving that you can get there before you run out of money. I think that's the thing—it's like, can you build the road? It's like, what is the first step, the second step?

And I think the first one is also very daunting because you have this super clear image of what the future looks like with how your company changes the world. Because a lot of these companies we talked about, when they succeed, they're going to be huge, and they're probably going to be a lot bigger than our software companies too. They're going after massive industries, like energy, one of the top expenditures in economic GDP index, right?

Yep. As straightforward, I mean, we were talking about Astro Forge. The thought that runs through my mind is, like, as investors and to some extent as founders, you're thinking of the expected value of your company, right?

Like, you know, there's like a very small chance of it succeeding, but like hopefully, the outcome of it succeeds is quite big. But when you think of a company like Astro Forge, like, okay, like the odds of being able to successfully like bring back mined precious metal from space seems pretty low. But if you do that, you own the whole asteroid.

You know, that company could be absolutely gigantic! So the expected value ends up actually being really high. Like, that could be one of the most valuable companies on the planet if it actually works. And then zooming out, like, that's sort of the job of every hardtech founder, period, is that in aggregate, the amount of risk might be this much.

And then we're saying that day one, you don't have to take on this much risk. You need to take on this much to show that you could take on this much, to show that you could take on this much. And so it really is about how do you break down the problem, which is itself—great engineers are very good at decomposing problems to begin with.

Yeah, if you are skilled, like, if you have the skills to build like real things, to really go out at a startup, you want to think as big as you possibly can. Like, don't make the tea-making robot; build the thing that goes into space and brings back precious metals, right? Like that, then you can actually have a shot at making the expected value work out and building one of the most valuable companies. And it makes the risk worth it!

I think the interesting thing that I've noticed looking at the YC portfolio data is that even though there's this reputation that hardtech companies are really risky, when we look at it in terms of our success rates funding these companies, it's actually about the same. Our success rate is actually about the same! And I think for space companies, it's actually higher than other segments of our portfolio.

Like, space companies is like one of the highest performing segments of YC, and I think the reason is that it's just two different kinds of risk. For hard tech companies, you have all this technical risk; you don't know if you're going to be able to mine asteroids. But you have no market risk, right?

Like, if you can mine asteroids, it's going to be freaking huge! The asteroids are there; they've got like platinum. The only question is whether you can get it—it's like a machine that turns lead into gold! Exactly! Yes, that will be valuable! Our software companies, they may have no technical risk.

Like, of course, you can build a website, but they have all this market risk. Nobody knows if people want this website at all, and there's often a lot more competition. And so it turns out when you like add those two things or like when you compare those two things, it actually turns out to be kind of a wash because these two variables—our saying, "Make something that people want?" The want is already clear.

The want is already clear, yeah! And all of the risk is stacked up on the make. Can you make it? And I think this is a call to action for a lot of the very hardcore engineers that this is your chance to build something huge and to really change the world, literally.

Because these are atoms and huge industries that you can go after—solving climate, creating a lot more efficient chemicals, going after being interplanetary species into this whole vision of space. It's like so many cool things that are also very inspiring.

I think the other cool thing about founders like this, when you talk to them, they generally have this level of excitement. They really believe in it, and they're very good storytellers to really convince not just like investors at the beginning, but later on—they have to have a very hardcore technical expertise.

Like Blake, right? He didn't come from—Blake from Boom didn't come from aerospace, but he has this level of enthusiasm and clarity that he was able to convince really legit experts to join him. And the ability to recruit, tell the story, is a huge superpower that founders here need to have.

I think Blake is actually a really powerful example of this because, Boom is actually Blake's second company, and he started both kinds of companies. So, his second company is Boom, the supersonic jet company, and his first company was basically a Groupon clone. It was this like social buying site when like Groupon clones and social buying sites were all the rage.

So it's like a quintessential typical startup kind of startup, and so he's had both experiences. And he came, and he spoke at a YC dinner, and he sort of contrasted what it was like to run both a super typical startup and a super out-there, insanely ambitious hard tech company.

And I'll never forget what he said at the dinner; it really stuck with me. What he said is, "You might think that running Boom was much harder than running the social buying site, and actually, it wasn't." ‘Cause like with the social buying site, it was really easy to like get the thing live! Like, you can build a version of like a—like a V1 of that in like a week, and you can launch it.

So it's really easy to like build a thing and launch. But then what? Now you need to hire great employees and convince them to join. And like nobody wants—for like great—it's really hard to convince great engineers to work at like the seventh social buying site.

It's really hard to convince investors or users or anyone to care about your site because it's just like not that interesting. And so everything after launching is actually like really hard. And with Boom, it's the opposite—like, it's super hard to actually build a supersonic jet, but if you just tell people that that's what you want to try doing, like the world will like rally to the cause.

And like investors and employees and partners and the press and like everybody wants to talk about it, and everybody wants to help you. And for all those examples that we highlighted, that is the case. Like for SeaBound, likewise, like the two founders are like young women out of college, and they were able to really assemble a team of hardcore industrial engineers to really pull it off because what they're going after is a very mission-oriented critical—it's like really solving the climate crisis and going after the shipping industry.

It's like no other companies that are solving climate are trying to touch that industry because it's so hard to sell to. But they've been able to do it! And not just the employees are taking a chance on them, but also the shipping owners. These owners are signing the pilots, the LOIs for them because they really take this bet on the founders.

I zooming out in general, like another tailwind for the whole hardtech startups, it's just like the cost of prototyping things is coming down over time. There's a good chance that I feel like with AI being able to run simulations and just abstracting away some of like maybe even like the middleware involved in building robots, for example, all trending in favor of it becoming easier to do things like with less money and in like faster time scales.

Yeah, the platforms really seem to build on themselves, which is what actually like space—if you think about space before, that's probably what's happening there, right? I think if you zoom back like five, six, seven years, we wouldn't have predicted that space would be the area that the big hard tech wins would be in.

And like you can speculate that what's going on is like SpaceX set the scene for hey, you can build a company in the space; there's a bunch of engineering talent that works there for a bit and leaves, as is the case with like our portfolio companies. That talent starts companies, builds on just like everything that's come before.

Like there's probably another vertical out there in hardtech that's going to be like the next space—it's just possibly robotics. Probably robotics is like would be my guess. And Nvidia at its market cap with virtually unlimited access to capital is probably one of the biggest accelerants to compute. And then, as a result, like that sort of robotic future?

Yeah, what a time to be alive! This is a piece of advice Paul Graham would give way back in YC, where it's sort of when you're fundraising, some people are natural fundraisers and some people aren't. And if you're not sure which bucket you're in, you're not a natural fundraiser, right?

So I got to feel like one of these things is there is clearly an, like, Elon Musk path to building one of these huge companies, which is being a fantastic fundraiser who is able to attract like, you know, hundreds of millions of dollars and go out and build Tesla and SpaceX. But like there is if you're not Elon Musk, there is also a path to doing that.

It just requires what we've been talking about here, it's like thinking like a software company. And I actually think a consequence of that is like the people you want to surround yourself with may actually be software people and like people or at least like investors who will push you in the right directions.

Because I think it's another thing I see with founders is, hey, like the advice they want to hear is—yes, of course, you need to raise like $50 million, and I can teach you how to do that. Like, the advice we give them is, "No, you have to go through the pain of figuring out how to do it with like $3 million, not $50 million." Right?

And I think that can be—that's a right playbook if you're not Elon, which is most people. And then at the same time, that teaches a discipline that sets people up to actually run businesses that are real businesses and not money-raising exercises.

Yep, well, we're almost out of time, but I hope that all of the examples we talked about today give you, the great builder out there, the sense that you don't have to be a multi-time billionaire. You don't have to have all these exits already. You don't have to be an Elon Musk; you just have to be smart.

And you can surround yourself with really smart people who will help you figure out the rest. So with that, we'll see you next time!