Planet Labs PBC (PL) Earnings Call Transcript & Summary

Please welcome Senior Manager Investor Relations Cleo Palmer peroner.

Hello and Welcome to Planet's 2023 Investor Day. We are so glad you could join us today. Now before we jump in, I'd like to remind everyone today's presentation contains forward-looking statements and statements about our long term targets and goals. Forward-looking statements are subject to risks and uncertainties as detailed in our SEC filing. We encourage everyone to review our filings which are available on our investor relation website and the SEC's website. Additionally the slides from today will be made available on Planet's investor relations website following the event. We encourage everyone to review the disclaimers, including the company and the slide presentation as well as our filings. At this time, I'd now like to hand things off to Co-Founder and CEO of Planet Labs PBC, Will Marshall.

Hey, thanks, everyone, for coming. Welcome to Planet. It's really great to see you all here. We've got an action pack line up for you all today. Here's the agenda. So I'll kick things off. You'll hear about our products and our go-to-market focus from Kevin Weil, our President; you'll hear from Robert Cardillo, who is in from his native DC to talk about some of the insights in the Defense and Intelligence business; Andrew hop in -- Dito for sustainability. And he's just back from a week at -- the climate week in New York. So he will have some stories from that; and then Ashley will round out the day with our financials and our path to profitability. Prep in here, we've got a couple of demos, 3 products, including one on our new Sinergise platform, the company we acquired over the summer. And then there's 4 customer discussions that you'll hear firsthand from what customers are seeing value in our products and services. Perhaps most of all, you'll get a flavor for the team here at Planet and the quality of the team that I'm proud to be part of here. And also, for those of you that haven't seen the Lab or you're -- in person, you'll get to go downstairs and see our lab. It is really cool and one of the most distinguishing pieces about Planet. And I think it's hard to understand plan it without seeing that, so you really want to see that. Before I get going, I want to acknowledge two things. Firstly, the heartbreaking events happening in Israel and the Gaza strip. We're working with our crisis response to try to do our best to support customers and partners, our government, media and humanitarian organizations. But out of respect for how early interstate of those events are, we're not going to be talking further about that work here today. Secondly, I want to acknowledge that it's been a challenging year economically. We faced some headwinds here at Planet. Of course, we're not the only ones facing headwinds, but the environment has affected us. For planet, it really comes down to 2 things. Firstly, execution. We have a huge stack of opportunities for our business, and it's our job to go and execute upon them. Today, you'll be hearing about our focused execution. And then secondly, Planet's business is still really not fully understood. The market hasn't understood what kind of business planet is and you'll hear more from me shortly on that. I mean it's also not understood what products we're providing to customers, and you'll hear from some of our customers about the value that they get. The opportunity in all this is that we are becoming more efficient and focused as an organization. It's making us better. And the opportunity is also that there's huge value to create for our shareholders. So overall, there's 3 messages we would like to convey here today. Firstly, that we're focused on durable growth towards this massive market opportunity we have in front of us. The tailwinds driving planets business are tremendous. I'll talk about some of the sustainability piece on security, the digital transformation. These are huge opportunities to go after. Second, we're focused and getting more and more efficient in our path to execute upon them both on the product side and the go-to-market side. We've made real product improvements. We're making real go-to-market efficiency improvements. And finally, AI. AI is a real accelerant to our business. There's a lot of hype about AI. In our case, it's really real and it is providing critical value to our customers today and is a significant market opportunity for Planet in the future. So I'll talk a little bit more about that. So I'll talk in a few slides on each of those 3 points. First, before we get into that, I wanted to start with a little bit of a background on planet. For those of you that are newer to plant, but also hopefully a valuable recap for those of you that have followed us for a while. So Planet, when we founded set about this mission to image the entire earth every day, tracking changes, making changes visible, accessible and actionable. And when we went public in 2021, we did so as a public benefit corporation because we believe what we're doing is really critical, and we want to bring about in the right way. We also believe that our business and our impact are highly aligned, and that's important, and that's reflected in our PBC. Our daily scan fleet of satellites is what does the legwork for this. It's unique and remains in a class of its own, and that is what's opening up the new markets that the planet is entering into. And Planet at a glance, you can think of it in these ways. We have the 2 powerful data sets, the daily scan I just mentioned and then the high-resolution tasking system. We have a large addressable market. Each one of these areas being billions of dollars of market opportunity to go after. And we've got a highly scalable business opportunity. That is each image that we take can be sold multiple times to multiple customers. So the direct margins of doing so are very, very high. We're a software-like business in terms of the financials as a result of the one-to-many nature of our data business. I find a useful analogy here. I've spoken about this before, is the Bloomberg Terminal for Earth Data. We deliver customized information streams to enable our users to make smart decisions. And that's why the analogy to Bloomberg makes sense. It's not perfect. Different from Bloomberg, we have a proprietary data set. They mainly aggregate public source, stock data and so on, whereas Planet, of course, has its own proprietary data. We also -- they also primarily so finance, whereas we serve a wide variety of use cases. So in those sense they're different. But just like Bloomberg, data feeds going to make slated decisions. And our data feeds go deep into people's workflow into our user's workflow, making them very sticky. I want to talk a little bit about what kind of business plan it is. And really, it's a data platform business. So I think that that's the way to think about Planet. This is a point that is often misunderstood. Let me be clear about what we're not. We're not an aerospace company. Our satellites are incredible and they provide an infrastructure that generates a unique and proprietary data sets, but we don't sell satellites. What we sell to our customers is data subscription fees. Comparing a satellite company is like thinking of Google as a server company. Google are really excellent at making servers. But that's not that product. Their product is search and other things. What we're closer to is a software company, a SaaS company. But really, we're a data platform company. And there's some subtle distinctions between data companies and software companies, and that's the point of this slide. Data businesses are rare. So you'll hear less about them. There's thousands of software companies or SaaS companies, but there's only a handful of data companies. And that's why it's worth noting a little bit about. But when they're successful, they're very notable companies such as Bloomberg that I just mentioned, but also Google, MSCI and others, they are all really data companies. So that's the first feature about them. Few other features I would call out. Secondly, it's all about the data. It's about gathering and accumulating and necessary corpus of data that's distinguished and differentiated from others. And that's often very hard for companies to go build that, in our case, launching a whole fleet of satellites to get it, right? But once that point is reached, which comes to my third point, The [indiscernible] really huge barriers to competition and fourth point, those benefits accumulate over time. So as we add new data sets, for example, it's often 1 plus 1 equals 3. And as network benefits of users on the platform, there's data gravity benefits of people on our platform, so there's compounding benefits as we get going. So that's a distinction of data companies and software companies. And finally, they're very sticky. I mentioned this point. They're very sticky flows. I would add a final point that AI and data are really best buddies, but data is king, and I'll talk a little bit about that when I get to AI. So that's a little bit about who we are. Now let me talk a little bit about the scale of the opportunity. These are really -- the fundamental 3 tailwinds driving Planet as a business. The first is digital transformation. This is continuing to drive the adoption of new data sets and software solutions by various different markets, most large-scale industries. This -- it's basically about helping customers across industries make digital efficiencies, optimizing revenues, reducing costs. Canonical example of that is our work in agriculture, where we help farmers and agriculture companies improve crop yields and decreased use of fertilizers and things like that, to increase profitability, that digital transformation of the agriculture sector. Then there's sustainability transition of the economy, and it's an increasingly pressing matter. To just give one example, and I'll show some slides on this later, it's in disaster response. We just had the hottest summer in -- on the earth history or at least in modern era, and it's leading to [indiscernible] climate-induced extreme weather events. The U.S. alone spend hundreds of billions of dollars a year responding to those sorts of natural disasters. And this can be reduced by quick response which we can help with and also even better preventative work, which our data can also help with. So -- and then finally, the need for -- in the area of peace and security is growing too. We've got an increasingly polarized world. We've seen the war in the Ukraine increased awareness about the need for global situation awareness, transparency and accountability of events around the planet. Let me make the opportunity a little bit more specific by sharing this third-party assessment that was done of just the several use cases, several government use cases within in the EU context. Basically, the analysis of what Planet data could do in domains like disaster response, agriculture and so forth for the EU. And just in their context, they found that they expected it to generate between EUR 5 billion to EUR 14 billion of benefit to their users in the years 2022 to 2030. A significant ROI, a huge number of jobs just on the back of planes data. Now let me turn to some of the execution. And we're really proud of some of the developments in the last 12 months. But I'm most excited to start with an announcement about this guy, which is our first Pelican space graph which I'm pleased to say has arrived at the Vandenberg Air Force Base for launch next month. So it's a major milestone for this program. We announced this program 2 years ago, the higher as continuity and increased capabilities of Pelican to be really exciting as we announced 2 years ago. Incredibly fast. We have our first demo on the launch pad today. I'm incredibly proud of the dedicated team that has got us to this point using our agile aerospace space approach. This is really a tech demo mission. So it will be testing out the integrated design of our platform, and we'll learn a lot from the mission. As you may recall, the Pelican is our replacement for the high red system, the SkySat system, but it's also going to be more capable, higher resolution, higher revisit rates, higher, lower latency and less expensive and getting to this milestone is a huge step. That's the first thing. But followed closely behind that is our first tenant of spacecraft using the same spacecraft bus that's going up. And we've got an exciting update for that, too. The NASA JPL developed payload, the hyperspectral instrument is not just in our lab, it's integrated into the first Tanager spacecraft downstairs in the lab today. And that's scheduled for launch next year. For those of you that are not familiar, Tanagers are hyperspectral constellation, and it has over 400 spectral bands capturing many phenomena, the human eye can't see. The Tanager to be funded by a partner at Carbon Mapper. And the main mission is mapping the world's CO2 and methane point-source submitters around the earth towards our sustainability mission. It also has many applications in defense and intelligence, agriculture, biodiversity and other things. It's really a new area and a new field because really, there hasn't been commercial hyperspectral instruments before. And I'll be remiss to talk about spacecraft without talking about our flagship fleet, the silos. It seems run of the mill, but we continue to improve and update that fleet. We just built 36 more satellite. They are also at Bloomberg on the launch pad. And this is run of the mill for us to build these things. But just to remind everyone, this is the largest fleet of Earth imaging satellites in human history. It produces our proprietary daily scan of the earth that is the backbone of our imagery archive and the backbone of our commercial opportunity. And these spacecrafts are truly in a league of their own in terms of cost performance, and speed to manufacture. And again, for those of you that are in person, you'll see that downstairs in the lab today, just how special that is, it's completely unique capability in the world. And here are some of the product milestones. Let me talk through some of the major things happening -- happened over the last 12 months. I've talked about the first 2 already. I'm not going to go through all of these. Kevin is going to talk through many of them, but I do want to touch upon the Planetary Variable. It's really exciting that this is enabling all sorts of new customer opportunities. I'll be speaking at the end of my section with someone from Swiss Re, who will -- is using one of our planetary variables, and you'll get to understand that a little bit more. I want to touch a little bit more on the forest carbon Planetary Variables that we announced this summer, they'll be coming available later this year. This is a groundbreaking data set that will provide information about forest change and carbon capture at the nearly individual tree level on a quarterly basis for the whole world. That is going to enable it -- at this stage is what will underpin carbon markets, conservation and regulators in this space. Frequent board area yet granular data are crucial for making this transition. I want to underscore the significance of this. The sustainability transition of the planet of our economy because of the planetary boundaries is a multitrillion dollar transition. And it's not just that our data set is relevant to it. It is foundational to it. In this case, I firmly believe that companies and countries across the world will have to balance their carbon books just as they have to balance their financial books. And in fact, regulations are already moving in this direction, both here in the U.S. and the EU. And the first step towards balancing your carbon books is going to be measurement. You can't do it without measurement. So this is a critical capability to underpin the carbon markets to make them real and it's towards this multitrillion-dollar transition of the global economy that's necessarily coming to us. And we also talk a little bit about recent customer wins. Over the last 12 months, we've got -- we've had a load of new customers here, some of the highlights. I'll just touch on 3. The NRO this is one of our strongest partners as you're probably all aware, through our ECL contract to do with the hyperspectral data from our Tanager program. So they're exploring our data there and synthetic data, and it's really exciting to have them on board with that new mission. Secondly, I want to touch on the U.K. rural Payments Agency. So that's a civil government agency looking at use of our data for sustainable agriculture monitoring in the U.K. So that's a really cool use case, and it's very rent and repeatable to other governments around the world. And then finally, I want to mention PG&E. So PG&E are using one of our Planetary Variables that enable them to monitor power line encroachment of vegetation in an automated way across thousands of kilometers of power lines. And this is written repeatable to other utility companies around the U.S. and across the world. I gave a more general sort of business opportunity before, but I want to give you a sense of a specific customer ROI. And I've called out this case because I found out about it recently when I met the Indian Prime Minister. And we discussed the state of Odisha who has been using and implementing a program of subsidies involving more than 1.5 million farmers just in their state. And the initial program rollout was stimming because there was all these middlemen that cause mass amounts of fraud claiming subsidies on behalf of farmers and various other things, and they needed a verification system basically. And they used our data, and it's going to be an incredible example of -- and really the results speak for themselves, which you can see there. Already in the first year, they saved $200 million worth of savings on their subsidy program. They basically eliminated the fraud and the middlemen from this program across these millions of farmers. It's pretty cool. I want to end by talking about AI. There's a lot of hype about AI, but for us, it's real. It's an accelerant. That's the bottom line. That's the way to think about it. It's a huge moment, obviously. So let me just step and talk a little bit about that. The pace is something we've never seen anything like. And the value of the planet is it's unlocking the value of our data, its ability to pull out and extract our information from our data set. Our data set is particularly suited to AI because our daily scan is a consistent database going back many years to train the models on. And I want to be clear that AI without data is useless and data without AI is very useful. So data and AI are best buddies because AI can help extract the value out of the data. But it's clear who's king. I'll note that all the tech titans are releasing code in AI but none of them are releasing their data. We -- they all know where the key asset is at. Everyone can train the AI on data of the text of the Interne. But proprietary differentaited data sets is where it's at when it comes to AI. That's where it's at in terms of solving important problems. And Planet has one of the most compelling data sets for solving important challenges around the world with which to train AI upon. Now -- so this -- basically, the AI is enabled by, as I said, and it comes from this huge moat. And here are some of the stats about the tremendous amount of data that we collect. I'm most excited about the last one. This 2,400 layers of imagery for every point on average around the earth land mass. And this is a moat that grows with every turn of the Earth. And that's what really differentiates us. Not only do we have many times more satellites than any other company or earth imaging satellites in the other company. But we also have orders of magnitude, more area coverage than any of the incumbents in our field. And this is what unlocks the tremendous value in various markets from agriculture to others. I also want to be clear that the AI opportunity is not new to plan. It's something we've been building out for years. We knew out of the gate that it was -- we had too much imagery for the human eye. We were going to need analytics. And it's just the reason that we would need Analytics. So this is something we've been building out for the whole history of Planet. And here are some of the major milestones. On the top, the AI milestones and the bottom and the planet milestones. On the AI side, the first -- the major development there, one of the early success of CNNs it was in the field of computer vision, enabling the identification of objects in the imagery. First it was cat and dogs and pictures online, if you could recall that sort of phase, but that core technology, of course, could be applied to the satellite imagery to find objects in there, and that's exactly what we did. And we set out in 2018, the vision of what we call Queryable Earth, which is basically taking every image and identifying all the objects in it, the roads, the buildings, the ships planes. And in effect, indexing the earth, so just as Google index, what's on the Internet to make it searchable, we're indexing what's on the earth to make it searchable. And suddenly, with CNN, that was becoming possible. But in preparation for that, we actually did some really important things. One is we redesigned our satellite, so the scan fleet to be what we call AI-first satellites. All the extra spectral bands that we added were for machines, not for humans. The first fleet of AI satellites. And we really were pioneers in another thing, which has standards to do with analytics-ready data for the industry to adopt all the pixels are self similar and calibrated and so on. And then on the back of these AI development, the next thing that came along was the transformer model, which then led to all the large language models and what we've heard with chat GPT and so on. And what they are doing in short is a massive accelerant even to the CNN in terms of getting to our Global Earth vision. And this year 2023, we really was the first time that all of this stuff came together. The use of generative AI models developed by a partner, Synthetic in particular, on top of our data set to locate this high altitude Spy Balloon across the U.S. That was the first time it all came together. But now we're already seeing some really fascinating use cases beyond that. Generative AI is technology that barely existed 6 months ago or a year ago and is transforming whole industries. And as you can see, it's important for Planet, but we've been preparing for it for some time. So now let me show you a little bit about the difference in practice. Here's a couple of real case examples. This is some of our work with more what we call classical AI in a real case of the table fires that swept through Lahaina in Hawaii. We had images before and after. And we had trained that building damage assessment tool for Ukraine context with our colleagues at Microsoft. And we've quickly replied it here in the context of Lahaina to do building-by-building damage assessment that was available within hours to first responders on the ground in Lahaina. But this was based on a model that actually had taken months to develop this building detection. The difference is large language models, and I'm just going to flash up this one slide , with one of the first use cases of it is that you train once and then you can search for anything. And it's really hard to grapple with that out of the gate, but it's incredibly powerful. So you train a model, then you can look across large regions for anything that you've decided in that moment in minutes. Here in China, there's a result from a few of our quick searches for ships, for solar farms, bridges, you name it. just across that entire region, we did it as just a test case. And you can imagine how important and powerful it is to have individual analysts able to look at large regions for new events and so on, whether it's in the area of disaster response or peace in security and so on. And yet large language models are really not only able to do this, but to do it with low compute. So we're able to do this with a small model. So once you train it, it's quite low compute, which means you can do it many times with low cost. And this is really just the beginning in a way because large language models are now moving to multimodal versions where they can be expressed in text, video, audio, et cetera, which enables them to get meaning between them. And then -- so let me put it in the context of one of my earlier slides, I mean one way to think about it is that the AI is enabling us to easily query that data stack, right? And it is enabling us to get SAM closer to our TAM, if you like, in terms of who we can get value to because people that were traditionally had to be experts in satellite these now longer have to be. And then because of the point I was just making, it enables more queries faster and a bit like our one-to-many imaging model. This enables fundamentally sort of course analogy of one-to-many query model. So there's a lot of margin in that potential there. So that's enough about AI. I'm going to just end by coming back to where we started today, we can't talk about all of the exciting and incredible progress we've made without talking about some of the -- and addressing some of the recent headwinds we have and what we're doing about them more importantly. So we've -- I think, really, the way to think about it is that we have this massive opportunity, a set of opportunities in our pipeline. But it's our job to execute it. Pipeline is not revenue. We have to cut them into contract. And so we did, over the summer, a lot of focused work to focus our teams on the large growth opportunities. We reduced our cost infrastructure. And the last slide is, it's going to be on where we're focused going forward. because we're very clear and focused across our executive team, and you'll hear from some of them today. On the product, we're developing our next generation of high-resolution fleets. You've heard about the Pelican. We're scaling our platform were accelerated by the synergized acquisition, and we're unleashing AI onto the Planet Scope data. On our go-to-market, we've focused our direct to sales teams on large customers in our core vertical markets only. We're serving all the other vertical markets via our global partners, and we're shifting the smaller capital opportunities to automated self-service platform via the synergized platform, which both automate that for those small deals, but also frees up our sales around the world to focus on the big deal opportunities. And finally, on the financial side, we're focused on durable and efficient growth. We made a commitment to making -- getting to adjusted EBITDA profitable by the end of next year. And we're focused on building a high-margin business, delivering sustainable cash flow. And you'll hear more about that from Ashley later today. In all, I feel we're very honed on scaling the opportunity and challenging -- though the economic headwinds have been is making us a better business. So in closing, let me come back to what I said at the beginning, which is the 3 things I'd like you all to take away from today. First, we deliver -- we're focused on delivering towards our high-growth opportunities towards our huge market. Secondly, we're focused on getting more efficient and focused, like I just mentioned. And thirdly, AI is an excellent to our business. So with that, thank you very, very much. So with that, it's my pleasure to turn it over now to our first customer conversation, which is going to be with Marcel, the Head of Agriculture Product Center at Swiss Re. Swiss Re's first few use case were really cool. They've been doing automated Parametric drought insurance, protecting farmers around the world, leveraging our soil moisture content Planetary Variables. Marcel is helping to pioneer this use case. And I thought it would be really fun for people to hear from this use case. So he has a few slides to share, and then we'll have a brief chat. Marcel, are you there? Over to you.

Good morning.

Good morning. or evening for you maybe.

Yes, Evening here. Now good to see you. Let me tell with something about drought insurance around the world. I'm representing the company Swiss Re, and I am part of the Agriculture Department. Next slide. So with Swiss Re, where we are, we are the #2 reinsurer in the world. we write around EUR 44 billion of premium and reinsure almost every insurance company that there is in the world. We are based in Zurich, we also have a lot of offices around the world. To be honest, I represent a small part of that with only 30 people. We are focusing on agriculture. And here, we have around 400 clients all over the world. Next slide. I have a bit of echo. What I'd like to show you today is about drought insurance. So drought Insurance is a super important topic for farmers, but of course, also for the world's population overall. So what we see statistically is that the number of drought, heat waves is increasing around the world. And that every year, we see a lot of news headlines about different parts of the world where there is problems with drought. This is increasing and the -- honestly, there is no end inside of this trend. And it's our task to make this insurable or to keep that insurable for our clients. Next slide. So what's the difficulty for us with drought insurance. So how classically drought insurance works, in a nondigital ways that you send a loss adjustor from the insurance company, he goes to the field. He looks at 1 by 1 meter of the field. He looks at the core, you look at the wheat and then he estimates what is the loss for that farmer. This works relatively well for around 20% of the global fields. But now there is a huge demand from different organizations, governments, farmers around the world to also ensure the other 80% of crops around the world. To do that, we need an accurate objective, transparent way to estimate what is the loss of the fields. The easiest way to think about this, about this what we call index insurances, for example, with the weather station, where you will do a payout in case, for example, there is within the summer less than a certain amount of range, which you see here at the bottom. So if it rains less than 100 millimeters, you will pay out. Now this is, in a way, is a bit too simple. So what we have been looking for in the last 10, 15 years is a solution that can help us to do this more accurate in a more global way. And with that, we landed with planet. Next slide. So with which data set, we are working with Planet. We were having -- receiving a data set called Soil moisture. This is one of these Planetary Variables we talked about before. So we get from Planet globally, real time, every 100 by 100-meter every day, we get a data point for how much soil moisture that there is in the first 10 centimeters of the soil. Now what's so great about the solution? It's global, it's high resolution, but the best thing for us, it works with radar satellites. So even if it's clouded, we can see through the cloud, and we can still see what the soil moisture is and what the payouts for the farmers should be. Next slide. So maybe to go a bit more in detail, so how -- what that product typically look like? So we have a crop cycle, it could be corn, for example, in the U.S. We know that by May, farmers has planted the crop start to grow. And then we divide that crop period in different phases. We define moderate drought, severe drought, extreme drought we define payouts. We look at the historical data. So the great thing of this Planet's data that it comes with 20 years of history, and this allows our actuarial team to accurately create insurance products. Next slide. So this slide really brings it all together. So when we created an app that directly connects through the API with Planet servers, the satellite slides over in the afternoon and in the evening, the farmers, the insurers, the Ministry of Agriculture, everybody can see what's the current soil moisture in the country, in the different municipalities. And then we see here in the graph. So if there is a deficit of soil moisture, then the automatically payouts created. So this is an end-to-end digital insurance system based on the planetary variables of planet. We are running this now. We started some 5, 6 years ago with the first pilot. We are running this now in 15 different countries on every continent around the world. Maybe to finish with the last slide here. So the last project we have been working on the last one and a half year is based on PlanetScope. So this is not the someday, this is the optical data where for farmers in the whole of Africa, we measure the greenest of the crops or of the grass land. This is a big program sponsored by the World Bank to help the cattle farmers in these 3 countries. It works very nicely and it's planned also to be extended to other countries in the world. Really, what I said, transparent and objective with a historical view and a real-time view being able to pay to the farmers the money that they need. When they need it most. With that I'd like to hand it over to you, Will.

Thanks, Marcel. Thanks. So we've got a few minutes for a couple of questions. Why do you start working with Tim in the first place?

The main reason was really the data. So without unique data sets, cooperation have not started. So you look at different providers of data, you look at weather stations, different satellites, then you look for what's the best data around and then we ended up with team that's maybe the first thing. And then as a second step, you need to feel a real partnership to feel that you do things together that you are willing to try really to capture that market jointly that was really a good. so far, we're happy with the corporation.

And can you just tell us a touch time a bit on the ROI for you as a customer? And how are you doing it before. And how does this help you save money or improve efficiency in whichever way? What is the value proposition express that way?

So the main thing is to go into markets where we were not before. So to make drought insurable, you need a lot of experts. And those experts are not available. You have a lot of countries like Brazil or Catapano Australia with million kilometers of roads, millions of hectares and it's just physically impossible to serve those. Also in Africa, we cannot without the satellite data to cover those. So -- in the past, maybe 15% or 20% of the crops were insured that we're really trying to push that to 40% or 50% of the global number of hectares.

Great. Thank you. That's very helpful. So -- and where do you see our collaboration going over time? I mean it seems like there's a lot of scaling potential in both the use cases that you mentioned. How do you see it scaling over time? And what are the products and services are you interested in?

Yes. So 2 things, of course, first, on the agriculture side. So I think we can still grow much further ensure more clients, maybe get more precise in a way. We also see a big potential in a hybrid model to kind of support the markets like France or Germany where -- or U.S. where we have a lot of these experts that still can help them using the satellite data to become more efficient, to be more accurate. And then as an insurance industry, I mean a lot of the topics you said, digital transformation of the insurance industry, AI, sustainability, all these topics are key rights for a company like Swiss Re, for our clients. This will happen right -- it's just a question of time. Honestly, insurance industry may be not the fastest will we get.

Shocking. Well, thank you very, very much, Marcel. Thanks for joining us here today and taking your evening to join us live. With that, I am now going to welcome to the stage President of Product and Business Kevin Weil, who came to us from small organizations like Twitter and Instagram, but now he's the real deal.

All right. Thanks, Will. Hi, everybody. At last year's Analyst Day, we talked about our strategy to make daily change, visible, accessible and actionable and to build a great business as we do. I'm going to show you how we're delivering on that and the positive impact that it's making with our customers. As our product line has expanded, it's become increasingly clear that we have an opportunity to move upstream in the value chain towards delivering solutions, not just delivering data. So I'll talk about how we're reorienting our go-to-market and how our acquisition of Sinergise and its geospatial cloud platform, Sentinel Hub, is helping us focus our sales team on larger customers. I'll show you a demo of Sentinel Hub and how it speeds time to value for us and our customers. I'm excited also to show you our partner Synthetics generative AI product, which is highlighting the value of our daily scan of the whole Earth to DNI customers around the world. And I'll end with a glimpse at where we're going next. But I want to start here. I have this poster that hangs in my office at home, and it says in big red letters on a white background, good work, consistently over a long period of time. It's something I learned working with Zuck in my time at Instagram and Meta, words that he said on an earnings call years ago when he was asked about Facebook's success. And this poster to me is what delivering for customers really means. And your strategy sets your general direction, but then you put your head down and you do good work, focusing on data and on customer feedback. And we're not perfect yet, but I'm really enthusiastic about how customer-driven Planet is becoming. And I think there's a lot of value building under the surface that isn't all visible quite yet. So a set to making daily change, visible, accessible and actionable guides our strategy. Let me remind you how we think about that and show you some examples of the progress that we've made just in the last 12 months. I'm also going to give a few sneak peeks into where we're going next. So making change visible. It's all about our unique Agile Aerospace capabilities and the proprietary data that they create. We operate many times more satellites than any of our competitors. We make them faster and more cost effectively. And we built Pelican at a rapid speed. You just saw it sitting at Vandenberg, ready to go out or first tech demo. And we did that all while building out a new hyperspectral satellite with Tanager. And we moved quickly from identifying a market opportunity to testing and building and launching. And when we do, we automate all of it. We automate satellite commissioning and management, and that allows us to scale up with customer demand. And in just the last 12 months, we made a huge amount of progress here. So Will talked about the Pelican tech demo. The Tanager payload is already here at the lab. It's being assembled as we speak, and we have a fantastic response from customers in our early access program. And we continue to improve the quality of our underlying product. So we've made multiple major improvements, the band alignments on Planet scope this year, in addition to lowering latency, increasing our data storage efficiency. And quite literally, Planet's scope gets better for our customers every single month. And Dito Sky set. So we shipped software that improved the Constellations capacity, made it more efficient and improved image quality and latency as we did. And the best part about doing this kind of foundational work is that it doesn't just improve our core product. It improves every analysis, every model and every decision that the product powers. Planet's ability to make data accessible to businesses and the governments around the world expands our TAM well beyond the traditional geospatial market. So we just talked about our core data, a daily 3- to 4-meter global scan of the earth, SkySat high-res tasking and now our archive of over 1 billion images going back over 7 years. But the even bigger opportunity is in moving up this data pyramid. From imagery to data and APIs to machine learning, time series in Planetary variables. The history of this industry is humans looking with their eyeballs at imagery. But the future is about computers running algorithms and analysis to automatically identify objects and patterns and time series. We're seeing this with our AI products. where we can automatically identify new roads and urban development or ships in open water across millions of square kilometers or with our partner in synthetic, which I'll show you in a little bit. Through our acquisitions of VanderSat and Salo Sciences, we're offering planetary variables, which are new categories with that fuse our optical data and of other sensors with different characteristics, microwave data, LIDAR data, even RADAR or STAR data like you just heard. And these products, they don't look like pixels. They're measuring an actual geophysical property, something like land surface temperature or soil moisture or forest carbon. And they're shaped more like a matrix or a time series, something that anybody who knows how to use Excel can use. And this is why with each step up this data pyramid, we're not just reducing time to value for customers. We're also meaningfully increasing our TAM. And we made a lot of progress here, too, over the last 12 months. So this spring, we completed the integration of VanderSat and relaunched their core planetary variables as part of our API. We rebuilt our analytics pipeline, making them faster and more accurate. And in the process, we extended our chip detection model out further into open water. We've developed new AI models that upgrade the accuracy of our cloud masks, and we have more coming soon. And we're in a beta period with a new product called Analysis Ready PlanetScope. Which is a temporary and spatially consistent radiometrically harmonized co-registered version of PlanetScope that's all tailor-made for time series analysis. So as we do all of this, and we build up the data pyramid, we're making it easier for others to use this work as a foundation for their own products like you just heard from Swiss Re. So they're standing on our shoulders to make more sophisticated products more quickly. And as we do all of this, we have an opportunity to speed time to value even further for our customers and partners. Because so far, everything that I've talked about has been about data. right? We've talked about Planet scope and SkySat. We've talked about Tanager and Pelican, even AI detections and planetary variables. That's half the battle. The other half is in workflows around the data. And as mundane as that sounds, it's fundamental to driving adoption and retention. So just like we want to build up the data pyramid so that folks don't have to continually recreate the wheel on analysis. Our customers are asking us if we can simplify the way that they integrate and understand and manipulate data. So imagine you're an agricultural customer. You're getting PlanetScope scenes, you're getting planet fusion data and you're getting SkySat images. But now what if you want to quickly generate a time series of NDVI values across all of your fields. Filtered by country and group by region. But if you want to correlate soil water content and crop yields over time. Or imagine you're a defense customer. You have a broad area of PlanetScope. You have SkySat tasking and you have ship detection. But maybe you don't want to just detect ships, you want to count them over time, you want to establish a baseline and you want to alert automatically anytime that value goes anomalously high or low. So what's needed here is really like a cloud software platform for the data pyramid itself. And each step we take along this path is something that our customers can start to do with a few clicks that they used to have to hire software engineers and check in code to get done. So this makes our platform stickier and makes it more valuable for our customers and partners. This is probably where there's been the most work under the hood, a lot of which is going to come visible over the next 12 months. But some that you can see today are, for example, making -- we are making weekly improvements to our account management and quota systems. These are small but critical changes that make our customers' lives easier. We're providing customers real-time insight into tasking, including weather previews, real-time quota management, sun angles and more. We've upgraded our GIS integration so you can do more with planned data from within ArcGIS or QGIS or Google Earth engine. And of course, we completed the synergized acquisition. And that's what we're spending more time on. So let me pivot to now looking forward, and I'll talk a little bit more about Sinergise. So Sinergise as a company started 15 years ago. They were doing geospatial consulting. And as they build solutions for their customers, what they found was that they were solving the same problems over and over and over again. And so they did what any sensible group of engineers would do and they automated those pieces. And that help them serve more clients and serve them faster. Now you rinse and repeat that for years. And what they realized they had after a bunch of iterations was a powerful geospatial cloud platform. So they named it Sentinel Hub. They opened it up for public use and they began charging customers for storage and compute. They called out processing units, just like an AWS. So now Sentinel Hub has these top-notch geospatial access and analysis capabilities. They leverage planet era data sets like optical and SAR and more. And of course, they have all the public data sets like Landsat and Sentinel. So whether you're a data scientist analyzing daily agricultural data, a web developer building a new EO solution or you're trying to integrate SkySat into a GIS app or into an iPhone app. Sentinel Hub helps you get there faster. So I'll show a demo here. And going to show you how easy Sentinel Hub makes it to do analysis. So here we are the government of Indonesia. And we're trying to monitor and enforce the forestation. So we're going to start by looking at publicly available data over the area. So for example, Sentinel, Sentinel scans the earth every 5 days at about a 10-meter resolution versus planet daily 3 meter. And Sentinel Hub makes it easy to quickly look at a time lapse of any imagery over any period. And as we do that, we're like it's really cloudy. Like there's some deforestation happening, but there are a lot of clouds that get in the way as often happens in tropical regions. So now we'll switch to planet data. And when we look at the time lapse, what we're going to find is we see the power of daily imagery because suddenly, you start to see much more granular change. You have fewer clouds, and one of the things you can tell is there's definitely clear cutting happening here. Now we go in further. And one of the ways, if you're looking at imagery to make vegetation really pop and really stand out versus nonvegetation is to leverage planets near infrared bands with NDVI. And in Sentinel Hub here, you can see you can construct combinations of bands and different ratios called ban map, and you can see the results in real time. You can even make this quick slider is giving you a quick before and after preview. All right. So now let's get more sophisticated. Sentinel Hub exposes its own imagery aware scripting language to developers called evalscript so developers can write Quick Script that do custom transformations and logic. I'm not going to go into all the details, but the point is Sentinel hub makes it easy for anybody to write algorithms that run map reduced style across all of the imagery in real time. In this case, to do custom logic on deforestation. And here, you can color based on the time the deforestation occurred and turn this into more of like a data biz output. So analysis that previously required you to set up huge imagery pipelines, you can now do with a script in the cloud and you can iterate in real-time over terabytes of imagery. And it's been my experience, my entire career, that when you tighten the loop between questions and answers, you make more innovation possible, and that's exactly what Sentinel Hub enables. So now you can continue to iterate on this evalscript. You can build business logic into it, say you're tuning thresholds for what's considered okay and what's not. You're highlighting the most egregious bits here in red. And with that, you can stream the result directly into a dashboard into a GIS program, even into an app on somebody's phone. Another cool thing is, this is a 3-minute demo I probably showed you like 5% of the power of Sentinel Hub. We only just touched on evalscript. We didn't get into time series analysis. We didn't talk about batch processing. We didn't talk about their upcoming ability to host machine learning algorithms. Sentinel Hub and planted together open up a world of possibilities and you're going to increasingly see them not as 2 separate things, but 1 deeply integrated cloud platform for our customers around the world. Sentinel Hub also shines as a way to explore and analyze planetary variables. So all of our planetary variables will be integrated and available within Sentinel Hub this year, including some of the new ones that we're really excited about. The upcoming forest carbon products that Will touched on and that Andrew Sole will go deeper into and this new automatic field boundaries product. Field batteries are fundamental to any agricultural workflow. So whether you're talking about precision agriculture for like a large ag tech like Bayer, ag insurance for Swiss Re. or anything in Between. So you have for example, the CAP program in Europe, where European countries are providing $50 billion in subsidies for farmers who follow sustainable agricultural best practices. These subsidies are, of course, at a field level. So each country in order to recognize these subsidies needs to know their field boundaries by field across their entire territory. And if you look today, what they do is they ask farmers to upload them or they have people looking at imagery and segmenting fields by hand, both of which are time-consuming and error prone. Our automatic field boundaries product can do this all automatically. And one interesting thing was this was in our road map as Planet. And when we brought the synergized team on, they had already been working on it because they had felt that demand from customers as well. So now we've got our teams working together to finalize this product, and we've seen a lot of interest from a bunch of customers already. And lastly, on the product front, let me talk a bit more about AI and why it represents such an incredible opportunity for us. Before planet, there was no such thing as a daily scan of the year. It didn't exist. It's a new capability for humanity. Planet data set is recorded macroscopic change across the entire phase of the earth for going on 2,500 days now. But in doing this, we've created a new problem, which is there is so much data. 13 trillion pixels per day that we bring down from our satellites. So if you care about broad area change, you need this data, but how do you slice through all the terabytes to get it exactly the bit that you want to know about? You probably don't have 7 years to manually look at the entire world every day. But if we could do it with AI in 2 minutes, I'm pretty sure that would work for just about everybody. There's a revolution happening in AI right now that's making all of this possible and it could not be a better time for us. So I want to show you a bit about what AI is capable of and some of which -- some of these things we've developed internally, others are strategic partnerships with folks. So let's start with this. These are detections over the 3 million square kilometers of the South China Sea. With AI, Planet and our partners are able to provide insight into where ships at Sea on any given day, even where there are aircraft in flight. So in this view, what you're seeing is detections from the South China Sea over 1 week in July. So let's zoom in. Here's an image captured by one of our doves. This is a 450-kilometer patch of the ocean, part of that original 3 million square kilometers that was just showing you. And if you're tracking maritime activity, there are things in this image that you need to know about. But without help, you're trying to find kind of a needle in a haystack, right? It's time consuming, and it's error-prone. Now it turns out there are 3 objects in this image. Can anybody spot them? Some people are seeing one. There are a couple that are there. This is where our AI capabilities shine, right? We can quickly locate what's invisible to the naked eye. We could scan the 3 million square kilometers, not just this 450 square kilometer patch but 3 million square kilometers and locate objects of interest in minutes with individual confidence scores about each detection and additional metadata like the object speed and the altitude. So, go one level deeper, you can see what these are. And you see planets AI, lets you hone in on the insights you need. So you don't get caught by surprise. You don't miss something that you needed to know. Now I'm going to give you another look. In this case, this is work that we've done with our partner synthetic, and it's going to show you what's possible with planet data. This demo is going to leverage generative AI. And what they do is they enable searching in real time. So this demo is about looking for missile silos. And technically, these are -- actually, we're going to be looking for environmental covers that protect the missile silos, people call them bounce castles, but the model knows nothing about missile silos. In fact, we're literally starting with a web image search over here on the right and finding an image off the web, giving -- putting a little square around the bounce castle, the missile silo cover that we want, and then we hit go. And now what comes back and in real life, this takes a couple of minutes. It takes a couple of seconds here in the video. It's a set of hits across the area in question of things that look like what we search for. And they're color coded by similarity. And one interesting thing is to note the variety here, right? We search for a rectangular white object. And what we're getting back are rectangular things, square things in different colors. We're seeing circular silos. And we can quickly fine-tune our search. This is what's happening here by telling the algorithm which matches or the kind that we want, and then we leave alone the ones that we don't. We can save this surge, we can sort of create different categories, and now we refine and we can start exploring the results over here. So we've gone from a web image search to this in about a minute. So let's see how this looks. All right. First detection, that looks like the thing that we search for. Keep going. So does the next one? Right? And you can zoom out a little bit, and you start to see there's a regular pattern of these objects in the desert in the middle of China. Now we can some in somewhere else where we found another bunch of detections and see what we see. And we find in this case, these are the circular silos that we found. So the algorithm is really finding different variations of the original object that we queued it towards earlier, which again came from a Google search. And so you recap [indiscernible] Visisearch for an object the algorithm had never seen was never trained for across millions of square kilometers in China. And the results are looking pretty good, especially for our first pass that took us 2 minutes. And then you can continue with this kind of iteration. So we could go back and tune on the right-hand side what we're searching for giving it finer-grained understanding of the results that we like and the results that we don't. And this kind of human in the loop AI is exactly what our defense customers are looking for, right? Because we can combine the best of what an analyst can do because they bring context. They have local understanding. They have mission awareness. They have a lot of context. And we combine that with the best of what computers can do, which is about massive computation to detect patterns and similarities. And if you think about what I just showed you, this was looking backwards, right? So we sort of -- in 4 minutes, we fine-tuned the search looking backwards. But now imagine playing this forward. So we've got this search that we've trained on missile silos in this case. And now imagine that each day, the planet imagery comes in. So an image of the entire world comes in, we automatically run that search over that imagery, and it's going -- and it can alert the customer for new detections of new missile silos being created, automatically, daily across any expensive land owner, right? We literally just showed you how to make this possible within 4 minutes. And that's 4 minutes that before the combination of Planet's daily scan of the earth and synthetic generative AI, this had never been possible. The other thing to remember is this was a search for missile silos. We started with an image on the web. We could have equally found runways, planes, flights, cooling towers, bridges, anything else you can think of. And although I gave a bunch of examples in defense, this broad area of search extends to civil government and commercial use cases as well, right? Imagine tracking construction of oil derricks or mining activity, geologic activity, floods and more. It's a new world and these AI techniques underscore and complement the power of Planet's daily scan. And just to touch on AI more broadly a little bit. GPT 4, I think, if any of you have played with it, it has probably blown your mind, but I don't think we're even at the steep part of the S curve yet. So we are about to see GPT 5 coming soon from OpenAI and Microsoft. We are about to see a new foundation model called Gemini coming from Google. You have Facebook fast following with their sort of Android strategy, keeping open source. The world is accelerating and it's about to get faster because GPT 5 and Gemini are both going to be multimodal models. That means they're going to understand imagery and video just the way GPT-4 understands text. Now they don't train on our data. I'm sure they would love to. But imagine what you could do if you combine the model that could understand and describe and point out novel features and imagery with the world's most comprehensive and up-to-date imagery of the whole planet. So it's a super exciting time in AI. It's a very exciting time also to have the world's most comprehensive and up-to-date proprietary geospatial data set. So that was a lot on our product strategy, right? The improvement that we made over the last 12 months and some of the exciting stuff that's coming soon. I talked about our philosophy of execution and iterative improvements and the compounding value of getting a little bit better every single day. This stuff really matters to our customers. And here, you can see a consistently growing customer base. And by the way, this does not include synergizes customer base yet. But more than a growing customer base, it's a happy growing customer base. Our Net Promoter Scores are now consistently in the 60s. And remember, that's on a scale of negative 100 to 100, where 0 is a good score and 50 is an outstanding score. And our CSAT rates are now consistently nearly at 100%. And is the testament to the work that our team is doing day in and day out on product improvements as well as our cross customer success. It's also an indicator of our ability to drive retention and upsell among our existing customers. Another great indicator for retention is customers adopting multiple Planet products. As you can see, there's a consistent trend, consistent upward trend towards adopting multiple planet products over time. So for example, you have traditional geospatial customers. They start with high-res tasking. But over time, they find their way to PlanetScope for various AI solutions like automated road and building detection. Other customers start with PlanetScope and base maps, and they realize that they need planetary variables or they want to tip and queue off of change and they start using high-res tasking. So the fact that we have multiple constellations, the fact that we have a broad array of products like planetary variables and AI. it creates surface area for us, for retention, for upsell with customers. And this is strategic because up-selling current customers is a really efficient way to grow. And the fact that we are consistently achieving an NDRR of over 115% the last couple of years, shows that we're getting traction there. But obviously, our recent growth rates have not been at the pace that we expect for the business and addressing that's been a top priority. So we spent a lot of time reflecting and a lot of time analyzing data. I'm not satisfied, None of us are satisfied, but I'm also a big believer that success comes from how you handle adversity. And so we've been taking these headwinds, and they've been a learning opportunity for us. For one, as a young growing company in a new market, we were in the habit of following every lead that came our way. So we know our data is useful in just about every vertical. So it's no surprise that we have leads kind of across the spectrum. But when you follow up on all of them, it makes sense when you're small, but you start to be a $200 million revenue business, you can actually inhibit scale. Because when you expand into too many different verticals at the same time, it leads to time spent on proofs of concept and small deals, you can lease to nonstandard deal sometimes that tie up internal teams and create opportunity cost. It even makes basic stuff like sales enablement and education a tougher thing to do. So to motivate some of the changes we're making, I want to walk you through a bit of data from our pipeline. So this is one breakdown of our pipeline. The colors here show the verticals, so civil government, Defense and Intel, commercial and research. The pie chart on the left shows you a breakdown by vertical of all of our qualified pipeline opportunities. The next 2 on the right show you the mix for opportunities greater than $100,000 in ACV and greater than $1 million in ACV. And in order to be counted here, a deal has to have already been systematically qualified and scoped by a member of our sales team. And you can see a lot here. I mean, first, one of the things you see is that we really do have a strong mix across defense and intelligence, civil government and our commercial verticals. When you look at the 7-figure deals, you can see that they've shifted towards government opportunities, which is reflective of the mix shift that we've talked about on our recent earnings calls. When you look at the 6-figure opportunities, there's a strong mid 6-figure deal set with commercial actually edging out civil gov for the most. And then on the left, you see this incredible number of small commercial deals. Nearly 60% of the total pipeline by deal quantity, not by dollar value. And the challenge is our tendency has been to spread ourselves too thin across too many of these smaller deals. So as we took a deeper look at this data, we saw an opportunity to simplify how we're going to market and service of faster growth. We're focusing our go-to-market around 3 core principles. The first is reducing complexity. So the main thing here is that we need to standardize and automate small deals as much as we can. Today, any time a salesperson spends on a $15,000 deal is time that they could and probably should be spending instead on a $250,000 deal. So we're leveraging Sentinel Hub and other partners to speed our process on small deals. In the future, we're going to be directing a large number of our smaller deals to Sentinel Hub with its powerful APIs, it's self-serve documentation, and that future is coming. So as of this morning, Planet's offering new packages that allow customers to buy multiple different sub-$50,000 packages of PlanetScope data and Sentinel hub processing units directly from Sentinel Hub's website with transparent pricing without ever needing to go talk to a salesperson. This is great for people to get started quickly, but it's also an avenue for these smaller deals that doesn't require a salesperson, doesn't require papering a deal, negotiating terms and so on. So as of today, it's available via a wire transfer and pretty soon, we'll open up credit card and PayPal. The second principle is to scale through partners, especially in our noncore verticals, right? We know our data is useful across nearly every commercial industry, but we're never going to be experts in everything. And we're never going to have solutions in every market. But our partners can. So we have a large partner ecosystem today, but we've often asked them in the past to front-load the risk with an upfront purchase. That made sense when we were smaller. But at our scale, we can look to better share risk and reward. That means aligning our business models with theirs to make it easier for them to get started. And it also means more meaningful upside for both teams when we win together. And then the third principle is about once we reduce our time spent on small commercial deals, once we offload more of our noncore vertical deals to partners, it means our commercial time can spend more of their time focused on bigger deals in our core verticals of civil gov, defense and intelligence and agriculture. This is an enterprise sales motion, and we're increasingly focused on selling solutions in these verticals, not just data. So let me talk a bit more about this. As we've expanded our product portfolio, we have these higher-level products like planetary variables and AI. We're seeing an increasing ability to sell and package solutions, not just data. At our last Analyst Day, I talked about this as part of our intention, but we didn't have the product portfolio in market yet. With the launch of planetary variables and some of the strategic AI work that led -- talked about, now we do, and we're just getting started. So this is a shift for us, but it's one that we've been driving towards for a while now, and it's one that we're really excited about because it can meaningfully shorten sales cycles and speed customer adoption and time to value. And to be clear, we don't plan to develop every solution. When we sell a solution in which our data is a key part, we're driving demand and usage of our data and we're helping the customers see value faster. So whether it's our partner solution as with synthetics broad area search or our own Planet benefits. And a great example of this is with civil government. So we talked earlier about this $50 billion cap program. Sinergise actually has offered an end-to-end solution for this program called area monitoring. It combines planets data, planetary variables like field boundaries and AI algorithms, all packaged in a single interface, so a country can manage this program across their territory. And we see similar opportunities coming with a program called EUDR that's focused around deforestation that Andrew will talk about soon. Our partner, Livio, as another example, they package planetary variables and PlanetScope together into an interface into a user interface designed to highlight vegetative encroachment. So think about linear infrastructure like power lines, where a stray falling tree can cause down lines, can cause forest fires, can cause billions of dollars and damages can cause loss of life. So it's exciting to be at a place where our expanding product portfolio with things like planetary variables and AI, is allowing us to build these more complex offerings. So both internal and partner led, and it's helping us and our partners deliver solutions faster and more repeatably. So that's just a quick look at where we've been and where we're going and how we're focusing our go-to market. I want to end real quick by reiterating our long-term vision. We began as a satellite company, but our future is a software platform company. And today, we're partway there. We're a data company, and we're a pretty impactful one. But as we really open up this market, building up the data pyramid with planetary variables in AI and building a geospatial cloud platform around it. Together, these all augment our proprietary data. They make it easier to use. They make it more powerful. They build sticky workflows and create retentive customer relationships. These kinds of products enable us and our customers to go from selling data to selling solutions, which are easier to understand and they're more directly connected to a customer's goal. So it's an exciting time. And with Planet data being used to address some of the world's most important pressing geopolitical problems and climate-related challenges, this is a future that we're extremely motivated to create. Thank you. And now I'm excited to introduce you to Paulina [indiscernible] from our commercial team; and Mark Rosenberg, who is a research data manager at CAL FIRE. It's going to be a great conversation.

Hi, everybody. As Kevin mentioned, we've got Mark Rosenberg with us today. I have the pleasure of working with this gentleman to my right. He through his work at CAL FIRE has taken on many roles over his 30 years in forestry analysis, including CAL FIRE's fuel reduction, tracking application, which we'll talk a little bit about today, supporting informed climate change policy around California. And then more recently, Mark has helped implement community wildfire preparedness programs within CAL FIRE across the state within the office of the state fire marshal. And today, he's now working with Planet. So to kick us off a little bit for those that might not know, can you tell us a little bit about what CAL FIRE is and what's the work that you do there?

Well, thank you for having me, Paulina. CAL FIRE is an entity within state government of California. We're actually the California Department of Forestry and Fire Protection branded CAL FIRE. And we have primary responsibility for fire protection across 31 million acres within California, a 100 million acre state. We also manage natural resources and regulate timber harvest on private -- state and private forest lands. And we have the state fire marshal who is concerned with Firesafe engineering and community wildfire preparedness. In my role within CAL FIRE is as a GIS manager in the community wildfire preparedness and mitigation division, I have 9 staff that were embedded within several division programs, managing utility, line clearance, regulation, looking at defensible space around homes, looking at post-fire damage inspection, tragic, but also something that we do. And we have community wildfire planning and mitigation. So it's our pre-fire programs.

So it sounds pretty simple, pretty basic. Can you take us back in time a little bit to how your relationship with Planet started?

Sure. I've been paying attention to Planet for a while. Global coverage, daily imagery is pretty useful. We've had several meetings with Planet over the years. More recently, though, there's a lot of emphasis on fuel reduction activities and keeping track of them and reporting on those. So I signed up for an event at a conference with Esri, the software company and you and I met and talked about a project that I was thinking about related to tracking our fuels reduction projects. We have a goal to get to 100,000 acres of fuels reduction a year. And so for that, we have to provide some reporting. But the most interesting thing about this project for me is transparency for the public on our projects. There's a lot of unmet information needs, and we're trying to address those by giving people a picture. And it took a while for us to get a contract written and awarded, which is pretty typical for a state government. And I just remember soon after we started talking, you gave me a call, we had a fire up in Northern California, the McKinney fire and you guys offered to provide some imaging for us, and that was a really successful collaboration with us providing smoke forecasts and helping to direct when appropriate times might be to take a sky stat picture of our fire and send it over to our wildfire integration and threat center.

And that fuel reduction -- wildfire fuel reduction activity, monitoring it and assessing it is the project that we're working on together today. That's the focus of our work together. Why is it important to monitor these? You already said visibility is a huge part of it, but what did that process look like prior to using Planet? And what does it look like now that you're leveraging Planet data?

Yes. Well, we're still working on leveraging planet data through our contracts. So we're just starting evaluating and then we're going to pilot the imagery for a full year. Prior to working with imagery and our current status is really GIS mapping. We have an application, it's enterprise geo database. We have field units, they go in and draw their projects over a map on screen or they're taking their GPS unit out to the forest, where they're doing fuel breaks or prescribed fires and then bringing that information back about what they did and drawing that in or uploading that to the application. So they're really not using any current imagery for their -- we have a 2-year-old imagery that they can look at but there's nothing really current to see where they've been doing their work.

So a longer process. And sometimes, I know we talked about accuracy of what's being mapped.

Yes. Yes. If you can't see a picture, they don't always do the best delineations on the map. And sometimes, they -- they [indiscernible]. They want to draw the area that they're going to do. But that's not what they did that, month or maybe even sometimes they don't get it all done in year. So having a good picture about where they did get the work done and what's left to do is really important, both from a work management perspective as well as from public transparency.

And then if you could also just to wrap this up briefly talk a little bit about what that catalyst was between leveraging Planet data previously and leveraging it now.

Yes. There's -- -- with the increasing frequency and severity of fire in California, there's a lot of emphasis on working on the landscape before the fire come. So protecting communities with fuel breaks are working out prescribed fire out in the more wildland areas. So, one aspect is the hunger for information about our activities and that the amount of activity we're doing has really increased. Also, I would say, over the last 5 years or so, CAL FIRE has gone through and been implementing a strategic plan for geographic information systems and using that for decision-making, data-driven decisions. And so you're seeing not just within CAL FIRE, but across the Natural Resources Agency and state government sort of this emphasis on technology and technology solutions and the ability to integrate that technology into our operations is increasing. And some of that's strategic, some of that's just the marketplace has really improved. And I've been using imagery for looking at fires for a long time. And you said, be able to request some imagery, then you would download it. It would take hours to download and then you would pull it up and it was all smoky and you really couldn't see what you were trying to see. But with streaming products and detections and a little bit more service from your professional services program, which is a real big deal for us. When we try to enter the market and use some of this technology previous to now, we really -- we would have had to put together multiple contracts to get professional services on top of imagery. And now a lot of that is built in with Planet and you guys are able to provide a lot more service on top. And we're talking about image enhancement algorithms against the imagery and also delivery because, as I mentioned, that delivery was really a bottleneck for us.

And then really quickly, where do you see your partnership with Planet going in the future?

Well, I expect that our -- the image solution that we're looking at will be really successful. We're in the pilot phase now, but if we get adoption across the state with our units and they find value, I expect us we'll be continuing to capture imagery moving forward to provide a tool for our unit staff we're doing projects as well as for the public to see where we're doing work. And the goal -- our target is 100,000 acres, but the target for the state is 1 million acres. And I also could see a lot of additional requests coming for places where our fuel reduction projects have interacted with a fire and we're trying to evaluate effectiveness. And so I think those will be places where, well, we've got this contract. We're planning to do all of these, but every once in a while, we'll get a fire and someone would want to request imagery for that. I anticipate some of that facility.

Well, we're looking forward to it. And thank you so much for joining us for this conversation. I'm going to invite Kevin back up here. Thank you everyone.

All right. It's my pleasure to introduce Robert Cardillo, who is our Planet Federal's Chief Strategist and Chairman of the Board, also a former Director of the National Geospatial Intelligence Agency. Robert.

Thanks, Kevin. It's good to be back with you all. I was here 2 years ago for Investors Day just after I joined Planet. But let me give you -- remind you about some context so you understand my perspective. I'm what's known as a lifer in this profession that we now call geospatial intelligence. I actually began that path -- that career early in the Reagan administration. I've earned this here. And I was -- what was known as a photographic interpreter. And what that meant was, it was my job to peer through a microscope on to a light table, at a flat of film, actual film that had been captured by a satellite that took that photograph, injected the capsule, dropped that capsule through the atmosphere, deployed a parachute got caught by an Air Force transport somewhere over the Pacific and a Trappice. That sounds incredible. I happen to think it sounds amazing because think of the ingenuity and the risk that it took to do all of those things. And the reason I tell you that story is, 1, to let you know where I come from, but I tell you, two, is that in those days, that's just 2 generations ago, the value proposition was, I have an image and you don't. Now this was all owned and operated by the government. It was a monopoly. It was highly classified. And so this was the cohort. And so for decision-makers in the United States and our allies to be able to see something that the adversary didn't know we could see or to hide something that they could not see was the advantage. And I say all that to set up my discussion this morning because while the image is always critical. It's no longer sufficient for the advantage that we need going forward. So again, with that as background, I am the former Director of NGA and now a new Chairman here at Planet Federal. I spoke that day about my excitement about what we were going to do over the next 2 years, and it was just really refreshing and rewarding to hear Will and Kevin and our customers show you what has been accomplished in the past 2 years. But my point for this talk is to -- also to share with you my enthusiasm, my excitement for what's ahead. And I am going to focus on the government in general, but really narrow my conversation around that defense and intelligence community that I came from. And to share with you my anticipation about what we'll do together. So with that, this is one representation of the difference from that world I described when I opened my talk, a lonely satellite operating its space at top secret levels. We're in a period where you might consider the sensing of the planet to be nearly ubiquitous. That's not quite true, but there's so much sensing that's going on now that there is a risk that if you don't provide a frame of reference, if you don't provide that context that you could actually confuse a customer or a partner through data indentation or overwhelming. We use the term Tsunami of data in the government from time to time. When I was a director, we had challenge coins, and my coin had the agency seal on 1 side. On the other side, it had 4 words. And they were content, context, conveyance and consequence. And the reason I chose those 4 words are as follows: as Will mentioned at the start today, if you don't have the data, you don't get in the game, right? So nothing happens without that data and Planet has been and will be a data provisioning company. But as I said and as shown on this chart here that, that data, if it's not framed, if it doesn't have a context, can, in fact, be confusing. Actually Deputy Secretary work onetime called my mission at NGA that I need to move the community from chaos to coherence. And this is exactly what we're trying to do here at Planet as well. And then, of course, me knowing something but not being able to convey it to you, doesn't do you any good. When I was in government, that conveyance usually was a piece of paper or a set of papers. They have a staple in the upper left-hand corner, and it would stay top secret at the top and top secret at the bottom. You heard and saw Kevin demonstrate and our customers demonstrate today. It's an iPad, it's an iPhone, it's a laptop. It's a signal on a computer screen. So a whole different way to convey. But at the end of the day, that consequence, okay, that outcome is why we're here. Can you make a better decision? Can you employ California fire assets in a way that save lives? Can you recover more quickly after a disaster? All of those things lead to the consequence that we're pursuing. So I mentioned my world. That's actually a modern look. You don't have the light table anymore. We have a screen. But the reason I bring it up is because I've mentioned before that we did some back-of-the-envelope calculations when I was the Director of NGA. And I said, let's total up all of the imagery that's coming into the building or into our agency in a given 24-hour period. That's obviously all the government provided imagery. It's our allies imagery that they would share with us. It's commercial imagery such as that we were getting from Planet. Airborne was included as well in Spaceborn. And then we applied the metrics of work that it takes to extract information from those images if you pursue the human model, the old Cardillian method back in the early '80s. And we did the math and we said, okay, to get through all of today's imagery, we would need 6 million imagery analysts here in the agency. And while that kind of excited me for a moment, I know that the Office of Management Budget wasn't going to provide the funds for me to hire said 6 million, but it really made the point we have to move to machine first, and we have to move faster. And oh, by the way, those calculations were done about 6 years ago, and I think it's easily 10 million now. And so the -- yes, there was some tension in the government about how and when to move to machines, how good are they, how reliable are they? Can they get us to the right answer with the fewest amount of errors. But as Kevin and Will have described this morning, the commercial technology in this space has not just caught up but really gotten ahead. So anyway, all of this now allows us to pursue value in a different way, which is how can we use computers to answer the questions they're really good at. What, where and when? And save your human, save your experts for the really difficult questions about why and what's next. And it's that human and machine, not machine versus human that I think is -- that I know is the way forward. And so we're all -- we're doing all that so that we can set up the next phase, which is what do these detections mean to me? Do I need to do add some confirmation, are there sensors, what assets can I use to confirm the data or the detections that I'm getting? And now what decisions can I make? So as we go from there, let me give you an example. This is a Chinese military aircraft. On the left side is 2019, on the right side is 2023. And if you look at the difference between the left and the right, you can see clearly to the eye, there's been some major improvements made to this Airfield. And that's obviously of import because those improvements lead to capabilities and capacities. But once you do that baselining of that construction, of that expansion, you then apply that daily scan monitoring to begin to develop patterns. What's normal? What's abnormal? How do I know when an exercise might be occurring? Has there been a pre-deployment for a future plan? And again, I'll say, even though I have lived in the imagery world the whole time, I'm also well aware and appreciative that imagery only tells you part of the story. Again, it's the frame of reference upon which you need to add other sources of information. But this, in fact, what I like about it is the sequence shows literally that time machine. How the evolution of that air field has occurred in the past, and as a lifelong analyst, I know that those past events and realities can, in fact, be indicators of the future. So it leads me to this story, which has been well told, but the reason I wanted to bring it up with you here again today is to remind you of the difficulty of what you've seen and what I'm going to show you again. On the 15th of January, on an island off the coast of China, Hainan, some group of scientists, and/or military officials launched a lighter-than-air balloon. And they did it for some reason, for some collection purposes, for some surveillance purpose on the 15th of January. But when that happened, very, very few people in the world outside of those involved with the launch knew that it had happened. On the 2nd of February, some 17 days later, you probably saw on your Twitter or X feed or in your New York Times app, Chinese Spy balloon over Montana, some pictures taken of this large craft about 65 feet in diameter with some sort of package hanging underneath it, slowly floating across the fields in Montana. Many questions, right, diplomatically to begin to ask from our State Department militarily from the Pentagon. And from the intelligence community, the forensic questions came and the first amongst those question is, where did this balloon came from? Now I'll go back to this map. That's the challenge, right? You know where it is now on the 2nd of February, somewhere over the skies of Montana, but you don't know where to look for it. And as Kevin demonstrated in our partnership with Synthetic we made the case, the theory that those prior locations of that object are somewhere in the PlanetScope archive. And again, using Kevin's math, we're now talking about 200 trillion pixels between the 15th and the 2nd. And by the way, given the size of that balloon, probably 100 pixels are what we were going to look for. So 100 out of a couple of 100 trillion. And again, this was early days. We had not done this before. And as Kevin mentioned, we didn't have a model of the balloon and didn't need one. But we did need to tell that algorithm kind of what to look for. So imagine sketching out a balloon shape and then going through the kind of procedures that Kevin did with the Silo example. Well, that led to this. And not only did it was detected once or twice, but it was detected 9 times in the archive from that point on the 2nd of February over Montana, all the way back to the 15th of January just off the coast of Hainan. And I have to tell you that what -- I mean, there's so many things that are amazing about this, it -- to me it brings to life the theory of the value of the archive. Because again, on the 15th of January, no one knew to look for a balloon launch off Hainan Island. And yet, once we knew some 17 days later that we then needed to go back and look, the fact that we had the archive and the partnership, we were able to unlock that history and unlocking that history became the value proposition. So to me, not only did it then answer that first question, where did this balloon came from. But then you could push forward. And you could begin to set the algorithm to say, have there been other Chinese balloons or how many and where did they launch from? When did they launch? Where did each one go? And so you can be proactive in the future to, again, set conditions to give you a notification when the next one might be launched. So take that capability, let's go back to China. In this case, you've got 3 air fields, 3 different military air fields, and you apply that approach and the teamwork between Planet and Synthetic by understanding not just the gross activity at these air fields, but the types of activity. So transports, fighters, tankers, et cetera. And you'll see in the columns, the different counts that occurred at different airfields because let's face it, every military is a military of systems. And so nothing exists in isolation. So you may have tankers at one base that are responsible to keep the bombers at another base, a loft during exercises and/or combat events. And you need to be able to know the activities of both because they're coincidence, okay, and they're working together is what tells you, gives you indication. So that's what you see in kind of the main screen. But I'd also like to direct your attention to the 4 images on the right. It says unusual aircraft detected. Well, that's an example that this partnership with this algorithm is that, yes, we told it to look for tankers and we told it to look for fighters. But it also found aircraft that didn't fit anything we told it to look for. And in my world, those are unknown unknowns. Didn't know to look for them, didn't know they were there, but the algorithm brought them back and said, you might be interested in, I found something else with wings. But it didn't have the sweep or didn't have the shape that you asked me for. To me, this is very exciting about going forward because let's face it. As much as we like to believe we know about adversarial capabilities and threats, there are always unknowns and to have the ability to tease those out and bring them to your attention is what is possible now. I've been in the Pentagon quite recently with very senior members of their leadership. And when we show them this example, the phrase we got back is, Detections as a Service. And frankly, I couldn't improve on that description. This is the breach dam in Ukraine that occurred some months ago. And you'll also recall that after the breach occurred, there was a great debate about who was responsible for the damage. And both sides accused the other. And I'm not here telling you that this image, okay, can just positively say who is responsible. But again, it allows that frame of reference to provide other sources of information to allow to have the international context, okay, for the conversation about attribution. And it's just yet another example of what Planet's been able to do to expose and to hold accountable what's happened in and around Ukraine, from the pre-invasion through some of the indiscriminate targeting that's been done by Russia. And then finally, I would say, too, that back to the Pentagon, that there has been more and more discussion about using this kind of transparency as the deterrent. Now it clearly didn't deter President Putin for making his invasion, but you can imagine using this kind of light transparency is a way to elevate the conversation in advance of future events. So I showed you this graphic upfront. And as I've mentioned a number of times, we're very proud here at Planet of what we've been able to do in the past few years and about -- and what we're stepping up to do here soon. I also want to tell you that given my role at Planet Federal, this is in very much aligned now with where the government, especially the defense and intelligence community is moving. Because, yes, will they continue to be for cures and buyers of our data, of our images, Yes, of course. But more and more, we're seeing the demand signal, and we're seeing budgets move to outcomes. And so where the listing -- we want to subscribe to activity. And we want you to provide us with the results of the imaging and the analysis in ways that just inform us on the activity. And again, Planet will continue to serve both of those demand signals. And again, if you think of that 7-year archive, it grows every day, and it gets more valuable every day. And as we identify other partners and other capabilities here inside of Planet to unlock that value, to me the upside is very, very large. So we're pivoting, we're maturing in line with those customers -- those key customers in our defense intelligence communities. And as they begin to solicit for those outcomes, our services become their solutions. So I hope you can see why I'm so excited about the combination of what you're seeing on the screen, which is the continued evolution of the planet sensing capability that both Will and Kevin showed you this morning, but also about the way that we're driving the value from that remote sensing capability. So it's both. It's both the sensing and sense making. And as somebody who's joined the community a long, long time ago to try to contribute to a more stable, more secure world, I'm sure I couldn't be in a better place at a better time than right here and right now. So with that, I want to thank you for your time this morning. As I welcome a friend and colleague, Dr. Jeffrey Lewis. There you are. Jeffrey, come on up here. Thank you. Jeffrey is a professor at Millberry Institute of International Studies. And among many roles, Jeffrey leads a research team that uses public and commercial data, such as planets, earth observation data to track the spread of nuclear weapons around the world.

Weird gig, right?

Indeed. Jeffrey, welcome. Thanks for joining us. Why don't you just start by expanding a little bit on that top-level introduction.

Yes. So as you say, I run a research team, and we make extensive use of earth observation data to track the spread of nuclear weapons around the world. I mean it's -- it's a classic problem because we all want to know what's happening, yet this stuff is supposed to be a secret. So we don't really have the option of just going to North Korea and saying, "Hey, could you show us the nuclear missiles."

You tried that?

Yes. Well, other people have, but you don't see them again. They just disappear after that. And -- like we have graduate students and there's some like institutional review board things if you get them thrown in prison. So much safer to go ahead and look and study these things using satellites.

Right. And so when, how did you first become associated with Planet?

Well, I think I was associated with Planet before Planet existed because I knew Will and Robbie when they lived in D.C., and were working on all manners space things. So it really wasn't until Planet got going though, that Trevor Hammond reached out and pointed out we've done an analysis. And he said, we have this huge database of 3-meter images looking at this particular thing you're looking at, and it moves. Would you like to see it move every day for a year. And I was like, yes. So that sort of has touched off what has been at least from our perspective and incredible partnership that's allowed us to do things that I really never dream possible.

And I suspect Jeffrey being academic and a researcher, there's not a data set in the world that you wouldn't want to explore and to potentially leverage. What is it about Will and Robbie's and Planet's approach that's been maybe particularly valuable to you and the research team.

Yes. The thing that is hugely valuable, really is the combination of the enormous amount of 3-meter data and then the ability to flexibly task if you're going to tip in queue. So there's this whole wide world out there. And if you're only looking at the world through high-resolution images like you have to know where to look and you have to know where to look in advance. And that can be really tough. And so having this kind of baseline capability where you're looking at the whole earth almost every day, clouds being cloud gives us the ability to monitor change in a way that is really remarkable because I think back to an early bit of research we tried to do with Russia's nuclear test site. This is in the news a lot lately because it looks like Russia might be moving toward resuming nuclear explosive testing. And we wanted to buy an image of that test site, which is way above the Arctic Circle. And so in our pre-Planet days, we contracted with an Israeli company to get an image of the site, which didn't exist on Google Earth. And it took something like 3 months, several thousand dollars and the satellite just made pass after pass after pass, and it was just clouds, clouds, clouds, right? And so we now monitor that site with Planet because we're getting daily 3-meter images, where you can see all kinds of change and then that can queue you to know like it's now time to put a high resolution asset on the thing, and they have a high enough revisit rate that usually, we get a break in the cloud.

I'd love to hear your perspective. Again, we heard from Will and we heard from Kevin today about the excitement, okay, in the AI world, especially in the past 6 to 12 months, right? Apparently, this leap forward to capability. Have you been able to leverage that on your research team? And if so, how has it helped you advance your understanding?

Yes, we are just starting to try to integrate machine learning and AI, but it's ultimately going to be absolutely essential because the cool thing about the data is there is so damn much of it, right? The downside of that is like your 6 million employee problem. I would love to have a 6 million employee team. I'm sure that would be really good from like...

That's a lot of grand students.

It's a lot of grad students. I probably wouldn't oversee them all. But it's just too much data to look at with your eyes, right? And so a lot of what we do is signature based. And so we really need to find ways to get through that data, really, really efficiently looking for those needles that we know are in this enormous haystack.

Given your long, long history, Pre-Planet history with the company, given where Planet's going, right, with the next generation of high res via Pelican, which I don't know if you heard is down at VandenBerg as we speak. So the tech demo will be launched next month. The Tanager center is downstairs as we speak, being integrated as well for launch next year. What -- what excites you about the next 2, 3, 4 years with -- as Planet evolves its capabilities?

Well, there are a number of things that I'm really excited about. I mean one thing is just the shift to multi and ultimately hyperspectral data. A lot of times, I think people have this idea when you're -- I mean, you know this is not true because this used to be your job. But people with like normal lives, think that you get a satellite image, you look at it and you see the thing you're looking for. But it's not true. My favorite example of this is the Cuban missile crisis. No actual pictures of missiles, right? What you have is pictures of sites, right? You have tents, you have trailers. And so a lot of what we do ultimately is inferential, having multispectral and ultimately hyperspectral data helps us do a lot more of that inferential work. So for example, if you're trying to deal with a Chinese missile Silo that's camouflaged, right? The hyperspectral data can really help pull that stuff out. If you're trying to analyze the facility to know if it's operating or not, the -- you're looking at like turbidity of water allows you to know which way the process flow runs through. So there is just this enormous amount of data that's available about the world is getting captured. And I think a lot of that's going to exist in terms of multi and hyperspectral.

Yes. I couldn't agree more, especially in hyperspectral. What excites me about that is even though we tend to add that word imagery, it really is sensing. I mean it's in the truest sorry, sense to the word, meaning it's taking measurements, okay, across those spectral bands in ways that we can't interpret. What makes me hopeful about it is there will be no -- even no interest at all to try to go look at it. You know what I mean. Give it to the machine, set your algorithm and your approach to say, "Oh, no, I'm only interested if these 3 detects happen because those 3 things mean this chemical production or this crop threat for disease or whatnot". And I always worry that sometimes that when we label things imagery, right, people even -- well, let me go look at it, to your point, and find it. So it gives me great hope that the hyper spectrum will unlock answers we didn't even know to ask before.

Yes. Well, you know this movie Event Horizon? It's an old Sci-fi movie.

Yes.

That's because we are old, we know this movie. But not everybody does. I always show there's a clip -- I showed to my students where one of the main characters has gauged out his eyes, right? And he says, "Well, we're going, we don't need eyes to see". Right? And I think that ultimately is the future, right? We're not going to be looking at all this imagery because there's just too much of it. It's really just data, right? And when you start to think about what you're really trying to find out, which is what change is happening on the earth, you're just treating as 1 giant data set, you are going to look for what's normal, and you're going to have a computer tell you when there is anomaly or something that's abnormal, and then that's going to go ahead and queue you in. And again, I think one of the things I run into is people do not understand the fundamental value of the 3-meter data, right? Because they're used to thinking of an image of something you look at. And so the sharper it is, the better it is. But ultimately, when you are a good analyst, you can know what to look for even in moderate resolution imagery. And then the other advantages you get of having an enormous database of it getting it almost every day, right? Those advantages really start to arise. So like we use that to track Chinese missile test.

I love the end of that and yet I'm still stuck on just the imaging of the beginning of the gauging of the eye.

I mean I just gave you a nightmare fuel, right?

Can I also say though that I love what you said about the detection indicating an analytic thread or a theory of the case. One of the things I strongly believed and learned as an imagery analyst that sometimes, the answer is what's not in the image. Something is missing, right, that would need to be there or should be there to indicate a certain intention. It could be a supply vehicle they don't have or it could be they don't have the fuel to -- and so -- and look, I'm sure there's ways to tune the algorithms to help us -- what is missing from this picture as well. Having said all that, I don't want to miss the opportunity on this stage, Jeffrey, you put my former hat on. And thank you for what you and your team has done. I'll be frank. There were times I would open up the newspaper inside my top secret facility with my top secret badge, and I go, "Oh my goodness, what has Jeffrey done". Right? That -- my immediate reaction that shouldn't be on the front page. But then take a step back and I go, "Oh, no, no, no, it has to be part of the conversation". Are there things that the government will protect, right, for security reasons or have a smaller sliver of advantages. But I just wanted to thank you and your team for what you do to advance on the state of the profession or understanding of the world but do it in a way that's educational in multiple senses, obviously, for the institution. But a long way to say, I believe strongly that the more informed our citizen REIT is the better citizens they can be. So for all of that, I just wanted to thank you for that. And please don't stop.

Well, I appreciate and as long as Planet continues to partner with us and provide such amazing data, we won't.

We'll commit to that if you'll commit to your end of the bargain. All right. Thank you very much. Thanks, Jeffrey. With that, I have the great pleasure to welcome to the stage, another friend and colleague, Planet Impact Officer, Andrew Zolli. Andrew.

Hello, and good morning, good late morning to everyone here and to everyone who's listening online, it's terrific to be here with you. I want to just say I was past a note that it is not technically required for you to gauge your eyes out to get value for a hyperspectral imagery. We just want to make sure everyone is on the same page. As our Chief Impact Officer here at Planet, I have the great pleasure of leading a team of technical experts in areas like biodiversity assessment and forest and land use and climate and scientific research and development as well as humanitarian and human rights work around the world. This team works with every part of the company, the entire commercial organization, all of the other parts of the technical organization to help realize our overall aspiration, which you can see actually written in big letters on the other side of the front door at Planet, which is to use space to help life on Earth. Now I do want to say -- just to call -- we have a couple of threads together before I give you the kind of the main bulk of my remarks this morning, and that is just to call a couple of things out. First of all, sustainability and climate in particular, are the subtext and often in the foreground of all kinds of customer conversations and all kinds of customer use cases that we actively work on today. These considerations are at the forefront of agriculture. They're at the forefront of land use. And you heard in some of our customer conversations earlier today with our colleague from Swiss Re talking about insurance, the subtext of that work is how do we ensure novel forms of risk on a changing planet. We heard about -- with our colleagues at CAL FIRE, we hear about how important it is to reduce the physical risk of climate exacerbated wildfires, not just in California, but that's a problem all over the world. So I want to say the things you're hearing are very much today, those are today problems. They are today solutions, they are today consideration. But it's also the case that we're going to see all kinds of new markets forming, new applications, new drivers around these kinds of considerations. And that's what I want to spend a lot of my time with you talking about today. It's not just the stuff we do because we are a public benefit corporation because we want to drive to mission-aligned outcomes but because we are also establishing the markets and the approaches that will be used to scale solutions in the future. So scaling these kinds of solutions has now taken on an incredible level of urgency. That's the other thing that's happening. We're bringing these capabilities forward at a time when, say, the intergovernmental panel on climate change, the IPCC, which is the UN body of experts that gives us our consensus perspective on the context of climate change today. has told us that we are living in the most consequential decade with relationship to the climate that humanity has ever lived in. This is a decade in which we will either lock in or lock out the worst effects of climate change for sure. And that means that every day that we proceed and there are about 2,500 days left in this decade that every day that we work is, by definition, the day in which we have the most agency. And you can see that fact, Will referenced it earlier, this past July was the single hottest month in recorded history. The ocean temperatures off the coast of Florida were measured at 101.1 degrees Fahrenheit. There were SSAs in the New York Times like what's like the swim in that water. And my favorite headline, which we didn't put up here was from the onion, the [indiscernible] at the onion, who actually said, even sharks deserve a hot meal every once in a while. I hope they weren't talking about us. I'm assuming they were talking about fish, but you never know. But the real consequences of that kind of impact on the climate are mind-boggling. They will be measured in massive coral die off, they'll be measured in a dramatically intensified hurricanes and typhoons. Now in order -- one of the other themes you've heard today that I want to kind of draw some lines is like a couple of keywords you'll hear over and over again. And one of them is about signals. It's about creating demand signals. It's about creating reliable signals from this massive information, which we've now collected but the subtext of the -- one important part of the subtext of the narrative of AI is our ability to create rapid, high-quality, scientifically accurate indicators around which we can form new kinds of behaviors and fundamentally form new kinds of markets. That's because addressing this kind of planetary change is bigger than a single company. It's bigger than a single country. It's bigger than a single segment of the world. It requires us to scale existing solutions and create new ones and fundamentally create markets and market like structures where supply and demand can be met where indications of what's working can help drive scale and the data that Planet provides and others will be providing, plays a key role in helping these kinds of new market-oriented solutions to these kinds of issues of climate change form. The data we produced helps those markets operate in ways that are transparent, in ways that are coherent and ways that are ethical. And I'm going to show you a couple of examples of that. Now technology and policy form the kind of trigger and scaffolding around the formation of these kinds of markets. And you can see this coming forward from Europe right now in a huge raft of relevant initiatives. Today, these include things like the Norway's International Climate and Forest Initiative, which we've been very close partners with over the last few years creating data sets that allow us to commonly understand what's happening in our forest. You heard reference to cap that's the common agricultural policy. It's a huge program in which Europe is driving and incentivizing sustainable agricultural practices at scale. I'm going to talk about this third one in a minute about what's called the EU DR, the EU Deforestation Regulations. So that's Europe's effort to require that commodities and products based on commodities that are imported in and through Europe are not linked to deforestation. And then finally, CSRD, and I know it's a little bit of an acronym soup, but that's the EU's corporate sustainability reporting directive. That puts new requirements on companies to actually measure their climate and sustainability related impact. Now the thing to understand about all 4 of these is they are huge creators of markets. They are going to spur compliance with these regulations, all kinds of technological capability, and it is fundamentally unachievable without the kinds of data and products that Planet is producing. Like what we're seeing is the world saying we have to move with all boot do speed across these next 2,500 days to fundamentally drive new behaviors across global supply chains. We have to incentivize behaviors that reduce emissions. And the only way to do that to fulfill any of these obligations is with the tools that Planet is producing. Now -- could we -- could you roll ahead -- thank you. Now the one I really want to spend time on with you today is related specifically to the work that we're doing in forest. You've heard this is our latest planetary variables suite of products. And I want to tell you about how they work and then I want to show you them in action in a moment just for -- in a few seconds. So today, as was mentioned, you can see a lot in a high-density time series about forest change, right? We can see every active deforestation as it's occurring. We can see the trees being fell, right, because you're watching every single day. But in this new world in which we're going to see these new regulations, it's important to understand other dimensions of forest change. It's important to get at this kind of structure of forests. So we've been developing these products that combine small strips of airborne LiDAR data. And so for those of you I know that many people here know what LiDAR data is, but for those of you who don't, that's the instrument. We sometimes see them on self-driving cars here around the bay. It's that rapidly spinning. It looks like a siren. What it's doing is it's spraying around the car and collecting points of information to build a 3-dimensional volume metric model of what's around the car. It's what allows self-driving cars to avoid accidents. So if you take that instrument and you point it at the ground and fly it in an airplane, you can get three-dimensional strips of information about the volume of the forests that are below the airplanes. And it turns out that there's really good correspondence between the satellite imagery that planet collects at the 3-meter resolution and those LiDAR collects. And we can use the tools of AI to take the places where we have both and then extrapolate to the places where we only have satellite imagery. And that allows us to build a volumetric 3-dimensional model of the biomass of the forest. Now why is that important? It's important because when you understand the structure of the forest. You understand how much carbon is there. That carbon is really important for carbon accounting, for instance, it's going to be really important for helping to fulfill the mandates of the EUDR. It's also critically important for meeting the needs that we heard from CAL FIRE, but where are the fuels that might burn? So you can imagine how important that is to utilities to understanding the kind of structure of forests, right, to understand where is the defensible space around my power lines. It's also critically important to understand when we're actually working on finding the forest we're taking out those fuel loads over time, are those policies actually being fulfilled? How do we measure the risk of wildfire? How do we measure the -- and ensure against the risk of wildfire. Those are all critical questions that involve utilities, insurers, regulators, all of those actors need a common picture of what's happening in the forest. And that is to say nothing is the people who are issuing carbon credits today. There are all kinds of new markets around carbon crowd, so I'm going to talk about them in a second. But let's take a minute just -- I want to show you what these products look like. Because you -- just open up the video for us. That's great. So we're zooming in here on a piece of California for us. This happens to be a place where you've got commercial forestry. There's been fire on the landscape. There's been some illicit activity on the landscape. And as you can see, one of the things we love to do at Planet is to just to show continuous before and after. The fact is that we can train algorithms to find all the places where there's been change at this landscape scale. But then by applying these algorithms over this forest Canopy, we can actually see the 3-dimensional structure of the forest carbon over time. And that time element is critically important. That's actually what you're going to see here. This is above ground carbon for this very same place. And that carbon is both an asset and a risk, right? It represents something that could burn. It also represents something that can be sold if it's managed correctly to secure carbon on the landscape scale. There are going to be all kinds of applications for these tools. I want to share with you this particular application was built by a partner of Planet we've worked with a long time called Upstream Tech. This is a product called Lens and it's specifically designed to take for change and landscape level change like this and make it accessible to these end users. So we're going to see all kinds of applications like this, and we're already seeing significant demand for these products from all these different sectors. Now there are countless applications, as I said, for these, but I really just want to highlight too. You can come back out of the demo, if you wouldn't mind. The first thing I want to talk about is involuntary carbon markets. Now some of you might have heard recently that there have been really pointed questions raised about some carbon credits. And that -- those questions about their long-term climate benefits, whether or not they're really delivering have caused crisis of confidence for issuers. We've seen that the market for carbon credits begin to kind of wobble a little bit. And these kinds of tools are going to be absolutely essential for ensuring that we deliver trusted, high-quality science grade credits and that we monitor them over time to make sure that they're delivering. This is the thing that sort of acts as the guarantor of the market. That thing that Planet can do is not just provide transparency, but in so doing, provide trust. Okay. The other one I want to talk about is about the realm of EUDR. And just to give you a sense of how big this particular opportunity is. Oh, sorry, can you guys go back? Yes. Sorry, I'm not sure what happened to the slide, so I'm just going to tell you the statistics rather than show them to you. Starting in 2025, we are going to see these regulations take effect where the EU is going to mandate that companies that are trading in cattle, cocoa, coffee, oil palm, rubber, soy and wood are not going to be allowed -- they're going to be required to do a high-grade due diligence to ensure that the importation of these products or the production of these products or any products that were derived from them -- so this is not just raw commodities, it's finished products. It's trees and paper, just to give you a sense like that. They're going to have to certify independently that none of these commodities were produced as a consequence or in connection with recent deforestation or forest degradation. And the only way you can do that is with these kinds of geospatial tools. According to the European Commission and this organization, Eurostat, sorry, up to 1.2 million companies are going to have these kinds of requirements on them. So they're going to have to independently certify, and that certification is going to have to run all the way through the supply chain. It's going to cover $130 billion worth of traded goods. And the anticipated cost just of the annual anticipated compliance costs are north of $5.6 billion a year. So the EU has said, we have to dramatically accelerate climate action. We have to leverage global markets, and we have to enforce climate-sensitive trade. And again, the only way to do that at a global scale because it involves commodities from Indonesia and Brazil. It involves places all over the world is with this kind of global monitoring capability. There are many actors, there's entities like commodity providers and aggregators. Those are the people that collect all those commodities, the shippers, the government bodies, they're all going to need a common picture. Okay. So while the EU is creating new market imperatives, we're also seeing this kind of work monitoring carbon beginning to be extended beyond the realm of forest into other places. And I want to tell you that the second one in the area of what we call coastal blue carbon. Now that's the carbon that exists not in deep in the forest but at the edge. It's the mangroves and Sea grasses and salt marshes. It's these things. that where huge amounts of carbon are contained. In fact, actually, the independent scientific assessment suggests that these kinds of ecosystem secure carbon at a rate 10x greater than the mature tropical forest that we're used to monitoring. In fact, a scientific assessment in nature in 2021 valued the carbon that these ecosystems secure. Let's see if we can do this. Yes. At more than $190 billion a year, plus or minus $30 billion. Now whether that number is at the high end of the low end, I can guarantee you 1 thing. It's going to get bigger, right? Because we're going to need to affix all this carbon. And again, carbon markets are going to play a critical role. Now we've seen already demand from corporates who are interested in investing in these kinds of areas, and planets data can act as a critical guarantor of the supply side. There's actually more demand for coastal blue carbon today, then there are coastal blue carbon credits to be issued in support of that -- meeting that demand. And in order to deliver against that, we're going to see entities like Planet play a really critical role. Now to figure out the details, we've been working very closely with the Nature Conservancy at the scientists and conservation of the nation conservancy to understand how our data can play a role in monitoring these kinds of assets. This is really forefront work. It's important to know that -- could you go back a slide, just 1 second. So mangroves have been mapped, but at low resolution. Sea grasses have just been mapped but with no time series and no one's ever done the kind of salt marshes work that we've done before. This is all at the very beginning of the next stage that we're going to see. So if we go forward, this is what the most common public source of the measurement of mangroves in this particular case in Papua New Guinea. Now using our data, this is from public data sources and it's an open data set. This is what the scientists at TNC were able to discover. They actually were able to map 3x the number of assets. And it's important to remember, those assets are all fixing carbon, but they're also potentially the sources of the issuance of new credits. That represents a different kind of climate financing pathway and the formation is also critically important because if you have 3x the assets to put on a market, you are going to attract market players into that market in a way that you haven't before. So really phenomenal kinds of outcomes here. Before I go any further, I want to just pivot to the last little theme to give you a sense of these are like the elements of work that we're doing that are coming next. This is our preview of coming attraction. And the last area I want to talk about is in disasters. So we have seen from across our ecosystem, from across our customers and partners in every different industry intensifying interest in understanding how our data can be used, how our tools can be used to blunt the sharp edge of disruptions to prevent forest fires from occurring in the first place to respond to them more effectively when they occur. And there's a really simple reason why. And you can see it in this chart from 1980 to 2022, this is the increase in just the $1 billion-plus scale large disasters in the United States alone, right? If we are solely in the business of doing payouts and trying to recover from a disaster curve that ever onwards goes up into the right, we're going to have real problems just managing our budgets at public agencies, in civil governments, in disaster funds and in insurance. So we've got to be able to use these tools to effectively respond to disasters, and there's a lot of reasons why climate exacerbates these issues. It makes dry players -- places dryer, it makes wet places wetter. So it means they're more likely to burn and more likely to flood. But there's a big reason why we see that increased growth. And that's for this chart, which is called the bull's eye effect. The reality is that on planet Earth with every passing day, on an ever more crowded planet, there's just more people to be effective. There's more stuff on the earth. There's more infrastructure on the earth to burn. A disaster that cuts through a city in 1950 is going to cause substantially less damage than in 2100 or 2025 or 2050, just because the cities themselves have grown as a consequence. But no matter the reasons why the underlying cause, the impacts of disasters have become a major focus. And we have been seeing this really interesting set of applications of AI to these disasters. Now this is a story that I'm going to pick up with what happened about 18 months ago when we saw the war in Ukraine, within a few weeks in Ukraine, Planet and our colleagues at the Microsoft AI lab, AI for good lab, began working on a new system that allowed humans to assess a raw satellite imagery to determine whether or not buildings in the pictures had been damaged. So just presenting before and after images and letting train satellite experts, some of the 6 million folks that Robert wants to hire, at the UN say, okay, we can see that there's damage in this assessment. And then using that data to train algorithms to monitor every single building in the entire country on a weekly basis. That information went directly to the Secretary General of the United Nations and directly affected the UN's humanitarian response to the conflict. But then the story changed a little bit. That took about 6 weeks of consolidated work for us to set up. And then we deployed exactly the system. By the way, this is a picture of one of those -- one of the pictures. Could you go back 1 slide just really quickly. I don't know if you can see this in this little red line here, you can see that the common reports that they had of damage, this is to water and sanitation and schools and hospitals. The human beings were finding 6 incidence of damage, the algorithm once trained was finding 53. So vastly more ability to process all of that imagery. This is exactly the kind of tool that Robert was talking about. Now we took that tool, it took about 6 weeks for us to develop, and then we deployed it again. In this case, we deployed it in the earthquakes that happened in Syria and Turkey, you might remember those. And then we deployed it again. And again, and eventually, we came to deploy it in Lahaina, as Will mentioned at the top of his remarks. In providing the information here, every time we launch this tool in response to a disruption to try to understand which buildings had been damaged. The amount of time it took to deploy dropped from weeks to days and from days to hours. And ultimately, here, it got directly in the action cycle of the frontline responders at the American Red Cross. Our colleagues at the Red Cross that, normally, we have to process so much information that we couldn't possibly do this. Casey Richards, who's a Senior Director for situational awareness at the Red Cross said, the volume of information that can be synthesized would take us days and days. And at that point, the disaster would be over. But she said in this model, it's allowed us to operate in a much more consolidated period of time. And as soon as this started to occur, we started to hear the phone ring from all kinds of actors who want to be able to take the information and deploy it as quickly as possible. Now, I want to share and there said all those things, I didn't actually know the slide was still in the deck. Okay. Last thing I want to tell you about is one other piece of the work that we're doing in humanitarian response and response to disasters that I think is going to be transformational. And that's work that Planet and the Microsoft AI good for a good team and a specialized public health organization called IHME. They're a world-class public health organization based out of the University of Washington. We're working together right now to map populations around the world at the highest level of spatial precision and temporal precision. And why does that matter? It matters because knowing where people are is critical to saving lives. You've heard a lot of discussion today about where is the carbon, where is the water? What's happening on the landscape? Where are the objects? Those are all really important questions. But ultimately, the denominator for many of these problems is how do we help people deal with these problems. So what we're doing is we're providing satellite imagery and Microsoft is providing the analytics and IHME is providing the spatial demographics that allow us to produce the most up-to-date models of where human beings are living anywhere on the earth. This is an incredibly important thing. It's less important here in San Francisco, which is continuously monitored and very information dense. But if you look at the second and third and fourth most populated cities in say, a country like Niger, that happens to be the country, Zinder in Niger. They haven't been measured for decades, and they're growing at 5%, 10%, 15% a year in some cases. This is Niger, the last time a census was conducted. This is the growth that's happened since. And what we're able to do is use these algorithms to produce spatial models of population that all of that red stuff -- sorry, all that purple stuff, all those are indications of where human beings are living today, they weren't living in yesterday. So if you want to deliver public health. If you want to understand climate risk, if you want to understand disease burdens or you want to understand where people might migrate to, it's a foundation for all of that kind of humanitarian action. So what we're doing now is taking those population models, and we're taking these other layers to begin to understand risk in ways that we fundamentally couldn't before. Now these are all, as I say, previews of plumbing attraction. This is what is happening at the forefront of the application of these tools. It's a really wonderful time to be working, as Robert said. In each of these and many other cases, Planet is developing the products and partnerships that will build new markets and bolster existing ones, all while fulfilling our mission as a publicly listed public benefit corporation as an impact accountable organization, I like to say. I believe that we are proving that doing well and doing good are not at odds, but actually are fundamental strategic complements to winning massively in a future that is really worth living in. So with that, I'd like to say thank you for listening to me and I loved to explain -- to share all of this with you. And would you please welcome -- join me in welcoming up to the stage our Chief Financial Officer, Ashley Johnson.

Thank you, Andrew. In case you're wondering why our employees love working here. Everything that Andrew just described is what just excites us every day to get out of bed and keep working really hard at everything that we're doing. I know Pavlovian, you're probably opening up your models right now, but actually, I get to do something fun first, so you can hold on those. I'm actually going to introduce you to one of our customers. So Melanie Desjardins is the Director of the Northwest Territories Center for Geomatics. Melanie has worked in public service since 2001 and has deep geomatic expertise through her works at Parks Canada, Aboriginal Affairs and Northern Development of Canada and the government of Northwest Territories. So hopefully, Melanie is online. There we go. Melanie Welcome, and thank you so much for taking time out of your day to join us at our Investor Day.