[[Home|🏠]] <span style="color: LightSlateGray">></span> [[Interviews]] <span style="color: LightSlateGray">></span> September 26 2019 **Insider**: "[[Peter Beck]]" **Source**: [TechCrunch](https://www.youtube.com/watch?v=OZ-tEuGRffo) **Date**: September 26 2019 ![](https://www.youtube.com/watch?v=OZ-tEuGRffo) 🔗 Backup Link: https://www.youtube.com/watch?v=OZ-tEuGRffo ## 🎙️ Transcript >[!hint] Transcript may contain errors or inaccuracies. **Moderator:** Well thank you all for joining me here today. There's quite a lineup we've got here. Right off the bat though, I feel like I should - we should clarify something. People talk about the sort of "new space" industry, then you guys are all in it, but you all do very different things. You've got satellite components, high launch cadence rockets, and orbital imagery and analysis. What do these things have in common that makes them part of this "new space" industry as opposed to what we might have had 10 years ago, 20 years ago? Feel free to anybody jump on that grenade. ### The Space Revolution **Will:** Well I could start off. Look, I think that there's quite a revolution in space going on, there's a bit of a renaissance, and it's been driven by a few things. But the biggest thing in the satellite side of it is actually the miniaturization of electronics that you can stuff into a little tiny box - much more than was possible just 10-20 years ago. And that means, and cost of launch which has stayed relatively flat for decades, but you can stuff more capabilities in that launch vehicle like a hundred or a thousand times what was possible even 10-20 years ago. So we're leveraging the billions of dollars that being spent in consumer electronics to make satellites smaller. I think that's the core of most of the change that's happening. I think there's also a slightly different risk approach, which is just: let's not do it the old-school way where everything has to work with 99.99% probability of success, but rather you know, I could throw up more satellites than we need and if a few fail, no big deal. But that in turn enables us to put the latest components on it because we don't mind using something that hasn't been tested yet in space, that was something that wasn't ever done before, which is why sensors in spacecraft before were typically very old and obsolete by the time they were launched. So it's a different risk approach, miniaturization of sensors - I think that leads us to where we are today. **Peter:** Yeah, I agree completely with Will. I mean, you know, the miniaturization has been a key element. And if you break a spacecraft apart, what is it? You know, it's electronics, code, batteries and solar panels - all which have gone through through massive, massive change. And then, you know, for us on the launch side, we see ourselves as really the enablers to a lot of this. While this revolution is our job, it is basically to be a glorified freight company. Our job is to get these small spacecraft on orbit, and the biggest shift for us is really around frequency. It's, you know, there's a lot of technology and there's a lot of regulatory and a lot of stuff, but it really - the big needle mover here is actually frequency. **Natalya:** What they said. I hope, you know, new space really truly does embody a different risk posture, and I would love to see something closer to like 50 percent reliable being okay and building an industry around that. And I also think new space means a bit more consumer facing or new commercial applications than historically have been. **Moderator:** Absolutely. And to be clear, I think we're talking about a 50 percent once it's up in space, not a 15 percent hopefully success rate on the launch. **Peter:** I can make it cheap enough so that we can have 50% at the launch level, maybe someday. **Will:** I think the metric is probability of success times cost, right? And so, I mean, because the launch costs dominate us. So yeah, I mean definitely 80% probability success would be just fine with 100. So far you really are - so that's pretty cool, but we'll see. I actually have a mantra in our lab, but we're aiming for 20% failure. And actually we've had considerably less than that, only about 5% failure of our satellites in orbit. But that says to me that we're not pushing the envelope hard enough, right? Because otherwise we'd be taking a bit more risk. **Natalya:** And Accion actually has our first hardware on that rocket, so that one should be a hundred. **Will:** We can have separate rates here, it's okay. ### Cost Structure **Moderator:** So this not only - you mentioned the miniaturization, the technical side of things, and you guys all agreed with that. Has the cost structure changed as well? I mean, this plays into the risk component as well, but it used to be you had hundreds of millions going directly from governments to prime contractors like Boeing, Lockheed and those guys. But now you have VCs, even like angel investors going for smaller companies. They're not necessarily even building their own hardware, they may be piggybacking. How did that - how did this change in the financial strategies come about, and what does that enable? **Will:** Well, I think if you take the overall cost down to put a satellite into orbit because it's lower mass, because it's miniaturized technology, then suddenly the venture capitalists only has to put in a hundred million before business is real, rather than 10 billion, right? So it just became - because of the cost structure coming down, it became into the realm of what VCs could do, and before it just wasn't. **Peter:** I think there's been like a perfect collision of lots of things. There's a perfect collision of, you know, technology has reached the point where it's feasible, simulation and some of the engineering tools, also the dissemination of knowledge around the world, the ability to engage in other countries and other talents. And then also, as Will mentioned, you know, the lowering of the barrier. So it's no longer a billion dollar entry price, it's a hundred million dollar entry price, and that's much more feasible to enable the venture capital market to come in and play. And I think another part has been the, you know, the private billionaires that did come in and take what was a duopoly serving the government and show that you can introduce competition. So I do think Moore's law was happening and all these costs could be coming down, but without a little bit of that spark, it might have just stayed the status quo for a while. **Moderator:** And although we have seen a lot of technology enabling things, where do you think the main obstacles are that can still be addressed by research, R&D? You know, somebody in a lab, somebody in a garage - where do you think we're gonna see that innovation that, you know, makes a more efficient thrust nozzle or something like that? Where do you think we can still improve in the technological standpoint? ### The Future of Space **Peter:** I think the most exciting thing to be done in space is you haven't thought of yet. We have no concept of what's capable. I liken to where we are in space right now is kind of go back 10-15 years when, you know, to the early Internet era, and we've really just sent our first email. That's where we are with space right now. I have no idea what technologies and services are going to be, you know, as a result of this revolution, but the world's gonna be a totally different place. **Moderator:** I know you mentioned a time when we were talking earlier, I think that certainly the ground, the terrestrial side of things actually needs to be improved. And Will, you've done a lot of work in this side on the data relays, like the handling on the ground. Like it's not just, you know, better satellites, but it's better space infrastructure on earth and by governments, by private companies. All that's like - what kind of improvements can we look forward to there? **Natalya:** So we'll probably spend more time thinking about this part than I do, but I believe, you know, some business cases still don't quite close because the ground stations are still so expensive, and that connection from the down link to the actual customers on the ground - there's a lot of room to improve there. But like I said, Will probably knows more on that. ### Ground Station Passes **Will:** Ground station passes were something like 300 dollars a pass. So for like a few minutes to connect your satellite, you'd have to pay a few hundred dollars for commercial services. And for us, we need so many passes with hundreds of satellites with hundreds of passes every day - it's just obscene to do it that way. So we had to build our own ground stations all around the world. We have 34 four-to-seven meter parabolic dish antennas to collect the data from our satellites. That was an infrastructure we had to build. Now we could rent it out to other people, but other people could provide a service like that. That's a good thing, you know, building that sort of infrastructure so that other people don't have to do that would be a sensible strategy. But I want to touch back on Peter's point because I think it's really excellent. That really is here - it's good. There's gonna be a green field of opportunity coming out, and we don't know what is gonna come from it. From launch, what are people going to pop up there that makes sense in terms of a commercial business? What are people gonna use our satellite data for now that we're producing this daily image of the earth? What are people gonna do when they have better thrusters to enable different kinds of operational concepts of missions? I think this is - there really are so many different things. Now, it's not for the faint of heart. These are really complicated endeavors. I mean, we're both a startup, all three of our companies. Not - we literally started building ourselves in our garage. Sounds like a startup. We've got VC funding, sounds like a startup. But if you come and examine the systems engineering complexity in our organization, it does not look like a startup. This complexity of our satellites is hardcore, right? And then they all have to work together. We had to have these ground stations that we had to build and mission control systems to operate all the satellites. And we had to process millions of images every day and automatically stitch them to the earth servers and do all that. And then build a business on top of that, you know, to sell a new dataset to the world. And that is not for the faint of heart. That is a complicated systems engineering challenge. I liken it a little bit to our Apollo project. So at the present time, undertaking these kind of startups isn't like a normal startup in that sense. It's really complicated. ### Ion Engines **Moderator:** Natalya, I wanted to ask you about this as well. You created the technology that your company's equipping satellites with right now. At the time, you may have - you must have had the option or at least the idea, "Well, I could patent this and sell it to, you know, one of these big companies." But you decide to pursue it as a startup with you at the head, being a technical founder. What made you do that, and would you recommend it to other people getting into this business? **Natalya:** So, you know, when you have an idea like a new type of ion engine, and you really believe in it, and feel that there's a need for it, you want to get it out into the market as fast as possible. And you want it to be available to people. And at the time, you know, there was a little bit of maybe delusion, and there still is, and I think you need that. But because we were building our engine so differently than what had been done in the past, it wasn't that clear that we could have just handed it off to, you know, a Lockheed Martin and said, "Build these ion engines in this totally new way." We use a silicon MEMS foundry for a lot of our components, which is not common to space propulsion. They would have had to team up with, you know, an Intel or some other foundry to make the chips. So we believed that we could actually get these out into the market. You know, we had as good of a chance of anyone of getting them out quickly. And would I recommend it? I mean, you have to really believe in your idea and have some - you know, Accion's really excited about the commercial new space market, but also beyond that, interplanetary missions, exploration. And so if you have something like that, but that gets you out of bed in the morning, I absolutely would. **Will:** If I could just add, I think Lockheed Martin would more likely have shut it down. **Natalya:** That's all right. So you're giving it to them and then everyone... **Will:** Yeah, it's a nice, nice little electrospray engine you've got there. Be a shame if somebody were to... ### Transitionary Phase **Moderator:** So it feels like all of your companies are actually in kind of a transitionary phase right now, going from this R&D and design and launch phase to building a business, making that profitable, and that actually something that actually makes sense that you can offer as a commercial service to people. I know, for instance, like Peter, I feel like you told me that, you know, the R&D is basically done. Like, you've got the rocket, now it's a matter of stamping them out, getting a million of them going, because you want to get to a launch a week, which is the idea that, you know, many people a couple of years ago they'd be like, "Launch a week? That's utterly insane." But you're doing it, and you're well on your way. **Peter:** Yeah, I mean, following on from Will's point as well - a space startup is like, if you talk about barriers to entry to the market, I don't think you can think of anything that has more. I mean, generally people think medical and stuff, but, you know, from our perspective, one percent of the total rocket mass is the payload. So if you're a fraction of percent out on anything, it's just a firework show. And, you know, we had to get up, you know, basically broke or a bilateral treaty between New Zealand and America to be able to launch out of the U.S. We had to upgrade entire internet backbones to towns in New Zealand. Like, it just goes on and on and on. Like, the barriers are just enormous. But, you know, we're at right now is, yep, we've got through those and we've lived there, but we're, you know, you go from milestone to funding round and milestone to funding round, and, you know, on the cliff of death all the time, to a point where, you know, we have a viable product. And for us, it's just all about scale. The launch site in New Zealand, the reason why we have a launch site in New Zealand is it was the only place that we could achieve the launch frequency we needed to achieve. So we're, you know, building giant factories everywhere, and, you know, we're hiring about four or five people a week, you know, consistently and just trying to scale as fast as we can. ### How is this offering different **Moderator:** Speaking of extraordinary scale, Will, your entire operation is a mind-boggling sort of thing. But right now you're in this transition where you're going from the challenge of actually getting the birds in the air to doing something with all the data they're sending down. And of course, there's - we've had earth imagery for years and years and years, but how is your - how is this network different? How is this offering different? Is it just in terms of, you know, is it quantitative, qualitative, all of the above? **Will:** Well, so roughly speaking, satellite imagery of the earth was taken every few months to a few years at this kind of resolution, more like every few years. And we thought it would be useful for humanity to have more regular imagery of the earth for lots of things, lots of commercial and humanitarian reasons - to have a daily snapshot of the planet. You can see changes as they happen, and humans are changing the earth over time. And to make smart decisions, you need data on a timescale faster than you're affecting the planet. You know, I liken it to the spaceship earth concept the Buckminster Fuller promoted, where seven billion astronauts on a spaceship that's hurtling around the Sun. And just like if you're on a spaceship, in a little spaceship, you have to take care of the closed-loop life-support systems. Or if the spacecraft goes into a spin, you need to - and you need in any of those cases to take data on the atmosphere or the spinning faster than the timescale you need to fix it, otherwise you can't do anything about it, right? So it stands to reason that we need to have more up-to-date information about the planet. So what we needed for that were lots of little satellites or cameras and telescope systems imaging the whole planet every day. We've now erected that system. It's a lot of data, so we get a few million twenty-nine megapixel images down each day from these ground stations we have around the planet. So now we have about 800 images for each point on the landmass of the earth. So we've got this deep stack of data indexing change. What we're trying to do now is - satellite imagery was all about sending images to individuals that would look at it and analyze it. Well, you can't do that anymore because we're producing millions of images every day, and it broke everyone's system. And so now what we're trying to do is apply machine learning to it. So, just like Google has figured out how to look for a cat or a dog or whatever in a picture, we can apply the same core machine learning technology to our imagery and go, "This is a train, this is a bus, this is a ship, this is a house, this is a road, this is Burning Man." I was just showing these guys a nice picture of Burning Man. So we index what is on the earth. And I liken it to Google's sort of indexing what's on the Internet and making it searchable. We're indexing what's on the earth and making it searchable. So soon you should be able to go, "Hey, how many houses are there in Pakistan? Give me a plot of that versus time. What's the area of the flooding in Indonesia this year compared to the last year? Or, you know, tell me the latitude and longitude of the trees that were felled in the Amazon between last week and this week - can you just tell me where they are?" These sort of things should be answerable based on this queryable database of all the objects on the planet. But that's a kind of ambitious vision as well. And meanwhile, we're selling that data to big enterprise directly that can consume satellite data, but the AI piece will help make it more useful and democratize it to more people. ### The current state of the space industry **Moderator:** Peter, you talked a little bit earlier about how you had broken an agreement between New Zealand and the U.S. As far as the red tape situation goes, I feel like space is just the limit, like this has to be the most regulated - to coin a phrase. So how do you - how much of your companies has dedicated certain navigators for that regulatory landscape? I know you mentioned import-export restrictions, you talked about just like all - I mean, with launching, of course, that you are just walking through a minefield. So I'm curious how much time and resources you have to dedicate to it. **Peter:** Try not to depress us all. Well, I mean, the reality is that when everybody sat down sort of 20, 30, 40 years ago and drafted all the outer space treaties and all the regulations, nobody in their right mind thought there'd be a startup company would be launching or try to launch once a week, right? And the spacecraft wouldn't be governments - there would be also commercial companies trying to do crazy stuff. So it is just horrendously broken. And, you know, a third of our company, if it money and time has gone into developing the electron launch vehicle, a third of our effort and time has gone into infrastructure, building launch sites and tracking stations and all that kind of stuff, and a third has gone into regulatory. Because that is still a massive barrier. You know, our goal is: someone turns up with a spacecraft, no different to Phoenix, we load it on the vehicle, and we launch the same day. That'd be great. The reality is that it's a six-month flow by the time you get all your FCC license, export control licenses, and FAA launch licenses and all that. It's a long flow. So we are really challenging that and pioneering that. I mean, a lot of people don't realize the space industry is big and it's exciting, but America went to orbit 28 times last year as a whole country. So we aim to do that like in a couple of months, not as in a whole year as a whole country. So, you know, the regulatory systems need a lot of massage. The good news is, though, that they know that - all the regulatory bodies know that, and they want the same future that we want. So they're investing heavily into it and into finding ways around it. But nevertheless, it's kind of the unsexy bit of the - at least the rocket industry, you know, the regulatory pieces. ### Privacy and national security **Moderator:** Will, I imagine you've got three-letter agencies calling you up all the time asking for snapshots of this or that secret fort. You know, do you just have to pick up the phone and be like, "Oh, it's the CIA again," or, "Okay, GRU wants more photos of Washington." How do you - I mean, that just seems like a fundamental capability of the system. So how do you navigate that as far as ideas of privacy, national security, things like that? **Will:** That's a great question. I mean, firstly, it's something we care about a lot because I mean, we set up Planet when we left NASA. We said to do good in the world, and so we're very cognizant of thinking through who we work with and ensuring it aligns with the values of our company. Now, we do work with governments, including the US government, and we work with an agency called the National Geospatial-Intelligence Agency, which basically provides satellite imagery data across the US government. So it could provide it to the CIA, or it could provide it to the State Department, or it could provide it to USAID, and they use it for a wide variety of purposes. Of course, the US government does have its own satellites, but a bit like Peter's thing, we've launched a lot more now and a lot more than any country. Last year, we launched 146 satellites. I think that was about the same as the entire rest of the world put together. I mean, so typically over the history of the Space Age, there's been about 100 satellites launched per year on about 60 rockets per year. So one a week roughly for the whole world, and a hundred satellites per year. And we alone as a company launched 146 satellites. So it also breaks the regulatory systems. But yeah, we do get a lot of calls from different agencies. Our operative is to get the data to as many people as possible to ensure that they can help them make smarter decisions. I am a firm believer the more people knowing about what's going on in the planet will tend to help us. And if I didn't think that was the case, I would immediately shut down the company, frankly. And so I think - and I think it's technologists' responsibility to try and think very carefully about how they bring technology about and its societal implications. And so this is not a point lost on us, and that there's geopolitical relevance to our data. But the data set has much greater value in stopping deforestation, protecting the ice caps and the coral reefs, and other things, and helping farmers improve crop yields than it does for national security purposes, in my opinion. **Moderator:** And what do you hope your tech will be used for? I mean, you've got this extraordinary new type of thruster, and it could enable so many new applications, but it's up to - ultimately, it's up to the people making the satellites or whatever. But what do you think it's going to be used for? **Natalya:** Well, quickly on the regulation side, we have mostly export control regulations that are a pain for us. We're classified as ammunition, so, you know, same as a ballistic missile or like a launch vehicle. But one of our thruster chips provides about the amount of thrust that could hold a mosquito up, and so it's like a very antiquated set of regulations from the Cold War. But as far as Accion and our applications and what I hope it's used for, I'm really excited about the commercial new space sector, probably mostly to connect everyone on the globe. I'm excited to bring everyone online and can't wait until everybody around the world has access to doctors and schools and banks and things like that. And then beyond that, I feel like to start an ion engine company, you have to be a little out there. So we're, you know, mostly excited about exploring the solar system and beyond and scaling our technology up to be able to do that. **Moderator:** Well, I hope so too. And with that, we are out of time. Thank you very much all for joining us and for your extraordinary contributions to space. [Applause]