 Good afternoon Slush. I'm really excited to be here today to tell you a little bit about using satellites to map change on the Earth. See, some friends of mine and I left NASA about six years ago with a following singular mission. We wanted to image the entire landmass of the Earth once a day and make that data accessible and visible and actionable so that people can make smarter decisions around the planet. We thought that if we can image the entire Earth every day, people could use that for a wide variety of commercial and humanitarian applications, like helping stop deforestation and helping advance crop management and improving crop yields. We wanted to see and track wide-scale change rather than the satellite imagery to date, which has been every few years, and too slow to help to make smarter decisions on a day-to-day basis. This is how we imagined it would work. We would put just over 100 satellites into a polar orbit. Each one would take a strip of images as it goes around the Earth, and the Earth rotates underneath so that when the next satellite comes down it takes a strip of images just next to it. It ends up being like a line scanner for the planet. The Earth does the legwork of turning around once every 24 hours, so we scan the entire Earth every 24 hours. This is what it looked like when we began. They were literally just us in our garage, and we built our first satellite there, and our satellite is just the small thing on the bench there, the little telescope that you can see. Really what we needed to do to put over 100 satellites into orbit was to miniaturize the satellites. The satellites were typically larger than the size of a bus, and now we needed to make them this big. We spent a lot of time designing our satellites. We call them doves because they're on this humanitarian mission, so our satellites are called doves. They're about four to five kilograms in mass. They're about 10 by 10 by 30 centimeters, so about the size of a loaf of bread. We went through multiple iterations of these satellites to get the latest technology in there, 13 different iterations, which we launched on lots of different rockets. We've been launching them on rockets roughly every three months to keep iterating the technology, put the latest sensors, the latest hard drives, the latest processors into our satellites. We call it strapping space to Moore's law. The last generation of our satellites, which is now in orbit, have more than 10,000 times the image collecting capability of the first satellites, so we're really accelerating the capability per kilogram in orbit dramatically. We also figured out how to build satellites at scale because no one had ever built satellites in this sort of number. In fact, in the last year, we've built more satellites than the rest of the world combined, all other entities in the world. Now, our satellites are smaller, so it's easier in a way, but nevertheless, we've built more satellites than everyone else, which is a kind of interesting transition for the space industry. We've launched them on over 20 rockets from all around the planet, on SpaceX rockets, on Indian rockets, on European rockets, and have deployed more than 300 of these satellites. This is a video of a launch we did in February launching 88 satellites into orbit. And you can see it goes silent because it's in space, and you can see the satellites being deployed out of the back of the rocket here. Let me show you that again, because I think it's pretty cool. I've never seen anything quite like this, and then I realized no one had ever seen anything quite like this because no one had ever launched so many satellites. So, this is a record-breaking launch putting 88 satellites on one rocket. Before that, no one had ever put more than 14 satellites onto a rocket, and so this is a big change. This is the fleet we now have in orbit, over 180 medium-resolution imaging satellites. They take pictures around three to five meters per pixel, and then we have also just recently acquired Google satellite arm called Skybox, which have given us another 13 higher-resolution satellites so that we can zoom in on targets of interest. But I am here to tell you that after six years of hard work, we achieved our mission. We are now imaging the entire landmass of the Earth every single day. And we're making all of that data available on our website on planet.com, and I'll tell you a little bit more about that. Firstly, let me tell you some stats about this data. It's been a huge effort. We're now getting more than one and a half million images down per day of 29 megapixel images from our satellites, covering more than 300 million square kilometers of land, of the earth. So, the land area of the earth is about 150 million square kilometers. So, about double the landmass of the earth each day. Each one of our satellites can take about 10 times the land area of Finland per day. It's about six terabytes of imagery, which we collect with ground stations that we've erected all around the world, 26, 4 to 7 meter parabolic dish antennas that collect our data about 300 megabits a second when the satellites go over them. We can produce an image of the whole earth cloud-free once per week. And we built up this tremendous stack of imagery now, so now more than 400 images for every place on the Earth's surface. And what that really means is we've got a deep stack of imagery, almost like a time machine for the planet, of all the activity on the planet. And it's a super useful data set to now apply machine learning to, and I'll get on to that in a minute. But firstly, let me show you what some of the images look like. So, the three to five meters, three to five meter imagery, what you can see is buildings and ships, roads, vehicles. In this case, this is an oil refinery. You can see the ag fields, agricultural fields in the bottom, the little forest patch. We can see an individual tree, but you can't see people with any of our satellites. It's not that high resolution. And in fact, we like that because it can't get into personal privacy stuff. Three meters per pixel, you can't see a person, but you can see a building, you can see a road. This is at full zoom, one to one. The previous image was 25 by 16 kilometers. This is zooming in one pixel to the image there. You can just about see vehicles on the road, on this image, the ships going into and out of ports. That's the sort of thing. Then the higher resolution satellites that we have, they take pictures at 0.7 meter resolution. And this is an image from one of those. So, that's a full frame image, 15 by 10 kilometers. And when we zoom in on the ship in the middle, you'll get to see what it looks like in full resolution. At this resolution, you can count how many crates are on that ship or how many carriages are on that train. Or in this case, which kind of plane is it? Is it a 767 or a 757 on the tarmac? Now, this data is useful for a wide variety of applications, commercial and humanitarian. On the commercial side, we've been selling to agricultural companies like Farmers Edge and others that use our data to improve crop yield by 20% or 40% because we can tell crop yield and type from orbit. Consumer mapping companies like Google use our data to try and improve their satellite layers to make them more up to date. Governments use it for things like border security or disaster response. So recently, we're working with the US agency FEMA on the disaster response after the hurricanes in Texas and Florida and the Caribbean. And in the future, the data is useful for lots of other things, so we're starting to get into areas like finance because we can literally tell the oil level of every oil drum around the world and therefore tell the oil supply or tell the output from all the world's soy fields and people in commodities in hedge funds want to trade on this sort of data. But perhaps most important to us is all the social impact, so tracking and stopping deforestation, tracking and stopping illegal fishing, helping people in developing countries to access clean water because we can look at all the reservoir levels and that sort of thing. So we are really proud to bring this forward, this data, but there's something more and I want to talk about what we're doing next, which is this. You see, we have deep stacks of imagery and six years ago we could not have imagined all the developments in computer vision that now what we can do with that imagery is pull out and say, do object recognition in each image and say, this is a ship, this is a car, this is a road, this is a house, this is a tree, this is a plane. If we can do that on all the millions of images we're getting down from space, then we basically make an index of what's on the planet as a function of time. And then people can query that index without having to look through all this imagery, they can just query what's on the planet. How many houses are there in Pakistan? Give me a plot of that versus time. Or how many trees are in the Amazon this week compared to last week? Where are the ones that went down? And they can answer basic questions. So in short, planet is indexing what's on the earth, a bit like Google is indexing what was on the internet and making that accessible to queries, so using machine learning. Let me give you a few examples. Tracking carbon as it changes through deforestation. Counting ships in port, so you can just add a polygon of your port of interest and it will count ships versus time. Counting containers, counting and tracking planes. In the Houston floods we saw and could help FEMA with not just images, but where are the roads that still will function so that the people could bring in the supplies of medicine and food? So it can help first responders with aid. And here's a few examples of how our satellite has been recently used in the press. So for example, when people do things in the world like explode a nuclear weapon, we can see mudslides. When people set fire to villages, we can track that and people can see that like the BBC here. Or if people cut off large areas of forests in the Amazon, we can help make people accountable for that. Or if there's just missed this one mudslide, we can see that. So we're trying to do a lot of good with this data and I wanted to take this opportunity here being a slush with all the entrepreneurs to talk a little about techno utopianism and technology and doing good. We started our planet with this mission of trying to help the world. And I think it's super important that entrepreneurs really take responsibility for the technology and how it plays in society. It's not always good. If you look back at technology, sometimes it does good, sometimes it does bad, and we have to, it matters how we introduce it. And I think it's possible to do good and do well. That is, do a lot of humanitarian good and do well financially. Those two things aren't incompatible. We are certainly doing that at Planet. We're doing well financially, but we're also doing a lot of good. And I encourage you as entrepreneurs to think about that. It's not always the case that technology does good. We've seen, as a physicist, we're very cognizant of the example of nuclear weapons. But now with social media and it's interplay with democracy, it can have some negative effects. And of course AI is coming up and we need to think carefully about that. As entrepreneurs, what I'd like you guys to all think about is what can I do to help the world? The way we approach this problem was we took the Sustainable Development Goals of the UN, which is a good summary of all the world's challenges, and said how can we as space geeks help this set of challenges? How can we feed the world? How can we help access to water? How can we help with stopping deforestation? How can you use our technology to help the world? I really think this is possible to do. And in fact, if your startup is not doing something to help one of these goals, I think you should just stop. Like do something else. Do something to help this stuff. This is what we should be doing. It doesn't have to be one or the other. You can have commercial and humanitarian impact. So coming back to plan out. So we're indexing the earth a bit like Google was indexing the internet and enabling users to query it. I'm going to show just a short video capturing some of this analytics steps. In early 2017, we launched enough satellites to have the capacity in orbit to image the entire landmass of the planet every single day. Our next phase is to add the analytical features that enable customers to extract the information they need from this imagery. We think of planets platform as a way of querying the earth, sort of a macro search engine for the planet. So we are indexing where all the objects are around the earth, a bit like Google indexed what's on the internet. For anything that you can see on the ground, we can drive information from it. So we can count the number of cars in a parking lot in a shopping mall in Iowa, or we can count the number of oil pads that are active in Russia right now. What we've done so far is actually create an end to end imaging chain pipeline that can actually get you about one and a half million images per day. What that pipeline does is it takes the images from the satellite and aligns them to the ground, cleans them up a little bit and then allows us to apply some algorithms and some machine learning so we can understand what is in the scene. For example, what we can do is look at a given port and we can understand the number of container ships going in and out of that port and we can count the number of containers coming in and out of every ship. Lots of activity means it's going to be a good year for the economy. Less activity may mean it's going to be a bad year for retailers. You can even combine our platform with supplementary data sets that you have on your side to get even more powerful insights. An example of this is natural disasters. Let's say you're doing flood analysis. You're trying to compare what was there before a flood to what was there during the flood and after the flood. You take our imagery as a base layer, on top of that use government provided flood maps that give you an historical understanding of where the floods have occurred and where the risk is. On top of that add where flooding actually occurred and even on top of that which roads are still usable and which aren't. This confluence of imagery can actually provide you meaningful actionable insights. We're extremely excited about bringing this new capability to the world. We haven't even dug through some of the data we have yet. There's so much data that can tell us so many interesting things. What's exciting is that Planet has an empirical data set that is covering the entire globe and it updates on a daily basis. This has never been possible in the history of humanity before. The upshot for businesses is the information to make smarter, more informed decisions ultimately to get a competitive edge. So we are bringing this data set to the world and helping put analytics on it. I think of it in terms of human consciousness like when the Apollo astronauts first took an image of the earth when they were zooming around the moon. That sort of led to a birth of the green movement. Even though we knew we were on a planet, this is the first time people saw it in the vastness of space and the thin atmosphere and people went oh my god we got to take care of that and it led to a phase change in our understanding that we need to take care of the planet. Well just like that we hope that daily imaging of the planet will help people not just to understand that planet is changing but to take action and make smarter decisions to help us to take care of the planet. And so if you have challenges and ideas for how to use that data set you would be absolutely welcome to come to our website and planet.com and build apps on our API on the top of this data. You can just go there and play around. We'd be delighted to see what you can do to build capabilities on board our our platform with our data. Thank you very very much.