Off-Nominal - 189 - Glow in the Dark Plants (with Awais Ahmed)

Episode Date: March 20, 2025

Jake and Anthony are joined by Awais Ahmed, Founder and CEO of Pixxel, to talk about building a constellation of hyperspectral imaging satellites. Honestly, we mostly try to help Jake get his head around the electromagnetic spectrum.TopicsOff-Nominal - YouTubePixxel | Hyperspectral Imagery and Space Data CompanyGoogle leads $36 million funding round for Pixxel - SpaceNewsSpaceX launches 131 payloads on Transporter-12 rideshare mission - SpaceNewsFlatland - WikipediaFollow AwaisAwais AhmedAwais Ahmed (@awaisahmedna) / XFollow Off-NominalSubscribe to the show! - Off-NominalSupport the show, join the DiscordOff-Nominal (@offnom) / TwitterOff-Nominal (@[e-mail]) - Spacey SpaceFollow JakeWeMartians Podcast - Follow Humanity's Journey to MarsWeMartians Podcast (@We_Martians) | TwitterJake Robins (@JakeOnOrbit) | TwitterJake Robins (@[e-mail]) - Spacey SpaceFollow AnthonyMain Engine Cut OffMain Engine Cut Off (@WeHaveMECO) | TwitterMain Engine Cut Off (@[e-mail]) - Spacey SpaceAnthony Colangelo (@acolangelo) | TwitterAnthony Colangelo (@[e-mail]) - jawns.club 🐘Off-Nominal MerchandiseOff-Nominal Logo TeeWeMartians Shop | MECO Shop

TLS and go for main engine start. It's a matinee Thursday, Jake. Matinee Thursday. Matinee Thursday. Look how light it is in my room. A secret, a secret pre-recorded show on the same day as a live show. And so there's going to be all the keen-eyed viewers are going to be looking for what clothes we're wearing in the in the shows and notice that. The state of our facial hair.

Yeah, the state of our facial hair. Everything's going to be the angle of the sun. Tell us one show you think this was also recorded on. Yeah. But that's fine. We do that sometimes. Although daylight saving totally screwing up the angle of sun, right? Because now this is when it's off nominal early for you.

Yes. So. Yes. And we're spanning like a lot of time zones today, I think, right? We really are. Yes. Where are you calling in from? I am in Bangalore, India.

Bangal. So that is the voice of... Are this off nominal show we've done? I'm trying to think of... Ooh. You might be a record holder. Maybe. didn't we have someone

Well, we had Peter Beck Yeah We had Peter Beck About the Google Map measure this out I thought I thought we had one one time Where I guess like wasn't from that area But like did the show from like Hong Kong or something

I thought we had that once didn't we I still think we need to Google Maps measure distance this out All right That's homework So India Indians actually right on the opposite side of the Globe Okay

and probably be farther. Hard to get farther away. Yeah, because I'm in Philadelphia. So my antipode is like west of Australia, like where they think that Malaysian flight ended up. It's like the exact opposite out of earth for me. That sounds like a conspiracy theory waiting to happen. So then, yeah.

Didn't it happen like 10 years ago? That was a conspiracy thing you're 10 years ago. Anyway, let's talk. We've got you here today, uh, Pixel with two Xs. is the topic and probably some more origin story. Jake was reading, oh, you worked at these other places too. So I feel like we'll probably have some origin story stuff that will be fun to talk about.

Jake, did you bring a special off-timed edition drink? I did, yeah. So it's 10 in the morning for me. So I decided not to go all in on alcohol right now. So I have a cup of coffee, which is normal. But I did slip a little something interesting in it. So I picked this up. This is Orchata liqueur, which is a new thing for me.

I love Orchata. So this was like, okay, well, I'll try this. It's like, you know, it would be like a Bailey's kind of thing. I was like, it'll be good coffee. Pop it in there. You didn't lay off entirely like you made it sound. No, just like a little, you know, just a little kicker. But I had a sip already.

I got to say this is less Orchata liqueur and more bubble gum liqueur. That's what it tastes like. So I don't know. This may not be a repeat drink. That's a swing and a miss right there. It's a little bit off nominal, but yeah. I've got a La Cologne Draft Latte. I'm following you on the coffee front.

These things are absolutely delicious. This is my, this is such a treat. These little cans, just delicious. So if you ever see one, I would recommend grabbing it. If you're walking through a convenience store and you're like, oh, shit, that's the one that Anthony drinks. It's delicious. That's my strong recommendation.

Oh, wait, did you bring anything in particular? Are you got any drinks on or near you? Water, I mean. I mean, based on where we're, the time zone you're working in, that's probably a smart decision. So, yeah. We're the other end of the day. Yeah. Where should we start, Jake?

You let you want to just do origin story stuff? We haven't had a good origin story in a while. I think so, yeah, because, well, so, I mean, so you're the founder of this pixel company, which is based in India, and you have a, I mean, I think that is interesting in and of itself, because we don't hear, you know, we're not swimming in Indian space startups in the market day. So maybe we just start with that. Like, who are you? Where did you come from?

And how does you arrive at doing work like this? Absolutely. Just a little bit of clarification. We are a U.S. Indo company. So we do have an office in El Segundo, Los Angeles, Degando. And we have an office in Bangal. But obviously, the origin of it started in India and then we kind of expanded to the other side of the world. But in terms of the amount of work that we have.

do it's almost equal both in the US and India so we call ourselves a US Indo or Indo-US company but the original story starts almost exactly six years ago the end of February 2019 now like most of probably your viewers and yourselves I was obsessed with space as a kid wanted to be an astronaut at some point in time and then an astronomer and an astrophysicist so as things got more real and I got a taste of the real world it went from astronaut to astroastroar astronomer to astrophysicist. But then life had other other plans for me. But I ended up going through these competitive examinations that you have in India that I sort of talk about

and got into an engineering institute that was in the middle of a desert almost. Right. And so you have this campus and then there's like a desert for four hours and then you have New Delhi, which is the closest metro city that that you had. But that is where I think I got actually way for working with space technology. So there was a team, there was a student team called Team Anandth, Anand means infinity in Sanskrit, and they were working on a small satellite, a cube sat at that point in time with the help of scientists at the Indian Space Agency, YSro. So basically, Yusro had the student satellite program. They used to work with a few universities and train and teach these students on how to actually design, build, operate

a satellite. So when I got to know that there was a team like that in college, I was like, okay yeah I got to be a part of that and and it was basically like a four to six month process you have to write an exam and then they take you through about four months of study and then once they know you're serious about it after another set of interviews you're in that's what I did for my first year and by the end of my second semester I was in the team and so I spent a lot of time doing that building satellites and you know working working with the hardware that we had rather than going to classes and then there was an even more interesting experience that was in college this was in 2017

This was when Elon must start tweeting about the traffic in Los Angeles and then we said, okay, there's a concept called the Hyperloop that we should sort of test out. We'll have this vacuum tube near our headquarters in SpaceX. Let's ask teams from around the world to come build their Hyperloop pods, raise within that tube and see what pans out. And you know, like most engineering kids at that point in time, this was back in 2017. I was also a fan of what Elon had done with SpaceX and with Tesla. And so we're like, yeah, is this a chance for us to be able to go and meet Elon?

And so we said, okay, let's give it a shot. So we applied for it. Didn't really seriously think we would get through. But to our pleasant surprise, we got through the first stage, and then we had to go through a series of preliminary design review, critical design review. And finally, we were given the green light as one of the 20 finalist companies to build this hyperlopart.

And so now we have to figure out how to actually build something that no one has built before. There's no chat chisput at that point in time. Oh, shit, we won. Now we have to actually do this. There's no how to guide on Google. you can just like basically find a single source of saying how you build a hyperlopart. Now then you figure out from first principles exactly how do you build the levitation system,

the breaking system, the entire structure and we managed to do that in a span of three months in Bangalore at that point in time during our summer holidays. We brought that part to Los Angeles at the SpaceX headquarters and we were able to present to Elon himself and the rest of the SpaceX team and you very very crucial and formative learning experience for me. But you come into the origin story, it was while we were there at SpaceX. It goes on a tour of the SpaceX factory. So looking at those rocket engines being built and the Falcon 9 booster that they keep in front of their headquarters in Hothorn,

that was the Eureka moment for me that, wow, I think a private company is doing all of this. And it was started as a small company a couple of decades ago. And from there, I kind of made up my mind that, look, if we as a team of students could build a Hyperlopod, satellite should be easier because there were a lot more resources on it. It turned out it was not the case just because there's more resources and more people that have worked on it that it was easier. But, you know, naivety gets you started. So came back from there into college. I read about everything from asteroid mining to rockets to space stations.

And then a lot of it was not really possible for a college student without a lot of money. But satellite data and imagery seemed like there was a lot of it coming down. And while working with different sources of data, realize that, okay, existing satellite data doesn't cut it. we need something better. And so that's what we're doing at pixel. We're building and launching the world's highest resolution hyperspectral satellites, which help us see 50 times more information than current satellites. Nailed it.

Great. How did you settle in on that as the direction, like hyperspectral imaging in particular, right? Because I was thinking back to late 2010s, right? There's planets up and running with probably a couple hundred satellites at that point. You got Maxar working on their worldview, legion satellites still that's been going on for years what was what were the kind of like metrics that you took at the market to figure out which kind of imagery you're going to go after yeah I think look honestly I was just trying to figure out any problem that I would

work on in space right it wasn't like a typical startup well is there an actual problem I'm like no I want to work in space let's let's try and cram this interest into a problem that there might be but I think planet and what they had done was a lot of inspiration right there figured out how to use these shoebox-sized cubesats and sent, as you say, 200 of them or so to space and be able to downlink there. And Maxar had been doing that for the last three decades in the 1990s. But I think while working with this data came across three resolutions we talk about from space. You have spatial resolution, temporal resolution, and spectral resolution. Now, spatial resolution is where Maxar really shines 30 centimeters, 50 centimeters and so on.

How clearly can you see and zoom into a thing and zoom into an image? That's one. Then you have temporal, which is where planet did really great job where they said we will put 200 satellites and we will have a daily imagery of every part of the planet. But what was missing was the spectral resolution. How much spectra? How much more of the electromagnetic spectrum can you see? If I have a broken bone and I go to a doctor, a doctor can't use a normal camera to be able to see what's happening inside. You have to go to the different part of the electromagnetic spectrum, see it through x-rays. So while this is not x-rays is on the other side, the infrared spectrum, being able to split light into hundreds of wavelengths and be able to see what. see it. I think that was the calculus

of figuring out, let's not go and compete with what MaxA is already doing, let's not go and compete with what planet is already doing, let's try and figure out something new that will add as an additional metric or an additional dimension to what satellite imagery is already providing. But in fact,

that's what we tell the investors actually, but the honest part was, I came across hyperspeckle imaging when I was reading about asteroid mining and at some point I really want to go and mine asteroids that's out there and hypospactyl imaging was one of those sensors that you use to actually see what's on an asteroid, you know, is it made of metal, is it made of something carbonaceous, does it have, you know, water vapor in which you can use to use as fuel?

And so when I was reading about that, I was like, wow, hyperspectal imaging, can we actually turn this back on the Earth? And turns out that people had already done that. So we had NASA that already launched a hyperspectral satellite. So the concept was proven. And the plan was like, let's start with looking at the Earth and then hopefully sometime in the future we can look at other planetary bodies. So I have to have to always preface what I'm going to say because this the hyper spectral stuff like melts my brain a little bit.

Like I'm not good with the with the spectrum. I you know, I think I should be like all the different stuff I've done reading about Mars science and all that kind of stuff. But still just like I just, I can't do it. So just correct me if I say something really dumb here. But my question. When you say something really dumb, leap on me. No, so you talk about, you know, find this place in the market where there is a gap, right? And it's hyper-spectoral.

My question is it, is that because there was just nothing on that side of the spectrum, like it was all visible light or something else? Or is it because the existing hyper-spectoral imagery didn't have enough resolution? Like, are you going to a new place on the spectrum or are you on the same place just adding more zeros to your resolution? Where on the spectrum are you as Jake's question? The spectrum is your spectrum. Hopefully not on the artism one. But I think, so the spectrum was already being covered. So you actually had Sentinel that covers all the way up to, I think, 2,200 or 2,300 nanometers,

which is in the short-fueled range. But it does go with only about 13 wavelengths in total. So there's a lot of gaps in between the electromagnetic spectrum. So the electromagnetic spectrum is a continuous spectrum. Humans can only see in the visible range and only about three very discrete wavelengths. We are not seeing the continuous spectrum. We're only seeing red, green and blue as huge blobs of these wavelengths. And then Sentinel, which is the best in terms of the multi-spectral resolution, has about 13 or so.

Maxar also has about 6 or 8 or so, right? But they cover the range. They covered the visible range, the near infrared range, and the short wave-intharid range. So you're going out further in the infrared range. What was missing was, you know, 13 wavelengths or 8 wavelengths or not enough. You need hundreds of wavelengths. And the reason for hundreds of wavelengths is you are splitting that continuous spectrum. into very minute slices.

So you think of a loaf of bread, and instead of cutting like three big loaves, which you can't really stuff into your mouth, you're having very thin slices that basically cover that entire spectra, which is what gives rise to those hundreds of lines. And that's what hyperspectral sort of gives. So you are basically able to capture the minute differences in the spectra

due to that thin slicing that you would not be able to get if you had a very coarser resolution. Right, right. So if Sentinel's doing, I'm going to make up number. They're doing 10 and 20 and 30 nanometers of wavelength. You're doing 10.1 and 10.2 and 10.3 and 10 point, like you're filling in all the little spaces in there.

Exactly. So the width of each of the wavelengths of Sentinel, for example, is 60 nanometers or 100 nanometers, so it's very wide. Whereas we have between two nanometers to 10 nanometers, right? So basically if they only had one wavelength, which is from 900 to 960 nanometers, we basically would have at least six wavelengths, 900 to 910, 910 to 20, 20 to 30, and so on. Gotcha. Okay. All right.

So you're filling in space at a much bigger, much higher level of detail is the short story. Yeah, yeah. All right. So when you were approaching the market, right, you have all of your examples where government institutions launching satellites in the hyperspectral space. Was that, did that get that way because it was very expensive to build those satellites with this resolution before?

or are you exploratory in that, like, all right, who's actually going to buy it that isn't a government, or even if government is your customer? They're commercial customers, agriculture is one I always see related to hyperspectral. Have you kind of figured out that there was a market there that was underserved, or was it a, we know it's there, it's just too expensive to serve it? So the market was there because in the end you're trying to solve for the use cases. So, for example, if half a trillion dollars worth of crops are lost every year because of crop diseases and best infestations, it's because you can't detect them. And multispectral satellites cannot detect a crop disease or a best infestation before the symptoms show up.

So, you basically need to have roughly a three to four week advance notice that multispectal imagery was not providing. So when we started, there were probably three or four hyperspectral satellites that a few space agencies had launched. So you had one that NASA had launched called Earth Observing One, which had a hyperspectral camera called Hyperion. This was decommissioned around 2017 or so, right? So we had a bunch of data that it had been down that we could work with to see what hyperspectral actually shows us from space. But it was at 30 meters resolution, it probably cost more than $50 million, maybe close to $100 million. It took about five, six years for them to sort of build it. So it was expensive.

It took time, but it proved that the concept was. And then around 2016, 17, 18 and 19, you had two more satellites that went up. One was the Prisma satellite that the Italian Space Agency launched, which was also at a 30-meter resolution, which also cost tens of millions of dollars in many, many years. And then you had the Indian Space Agency launching Hysus, which was Iverspectal satellite, that again was 30 meters in the resolution, took many, many years, and tens of millions of dollars. But then when we looked at it, we were like, look, 30 meters doesn't really cut it. When we talk to customers in the agriculture domain, in the oil and gas domain, in the mining

domain and the government, they told us that, yeah, we have these two or three sources of data. One, it's not regular enough. It only comes in once a month or so, and they can't really task that satellite for their area. So that was a big problem because if they can't task your own, you can't task your own area, you can't solve your own problems. But even on top of that, even if they could have tasked that satellite, 30 meters

resolution was not good enough to be able to see some things. If you're looking at a oil and gas pipeline, 30 meters is too big for you to know exactly where you're looking at and where something might be. And if you're looking at Africa or India where you have smallholder farmers, you can't even distinguish between multiple farms to know which farm has a problem and sort of, you know, what is being grown where. So the resolution was, I think, the bigger issue and we said, okay, 30 meters doesn't cut it. We need to be able to get to 10 meters at least because that's what our customers told us. We reached, you know, I sent out cold emails and LinkedIn emails to a bunch of folks and ask them, hey, you're procuring satellite data today. You're paying something to them.

What are you paying them? Are you going to buy hyperspectral? Is it like a need for you or is it a nice to have? and if you were to pay hyperspectral, would you pay a little bit more for it? And the feedback was kind of overwhelmingly positive to our surprise. They said, oh, we've been experimenting with hyperspectral, but it's been on drones, and that doesn't give us a large area, so we need it from a satellite. But you need to give it at a 10-meter resolution or ideally a 5-meter resolution.

Now, building a 5-meter camera was not possible with the technology even 10 years ago or 7 years ago. So that was a part of what we had to figure out to get into 5 meters. So the resolution part is important, but even before the resolution, it was figuring out, How do you cram that into a small satellite? Because a small satellite is cheaper to build, faster to build, and therefore also cheaper to launch, because then that leads to the business model. If you're spending $100 million in building a satellite,

and you can only generate that revenue over five years by the time your satellite is ending, you've barely made money on that satellite. But if you're spending a few couple million dollars on a satellite and then you are operating it for five to seven years, then you've recouped the cost within the first year, and then it pays back more than what you paid for it. That was the rough calculus when we were looking at it.

So one, we have to make it cheaper and smaller. And second, we had to increase the resolution while doing sort of five meters. And we decided to take it from first principles and break it down and figure out how to build that. I'm always interested with a service like this. And I have the same kind of like fascination with Planet too, which is like you have this sort of like three segments of the business. Right. Like you can go after like the big government kind of things, which is always lucrative. But, you know, it's a small market in that sense.

So you get, you know, huge injections of cash every once in a while. And then you have like these kind of larger organizations, larger corporations, but probably what you talk about, like, you know, oil companies, mining companies, these larger groups that need this kind of data. But I'm always kind of fascinated when you get closer to like the retail side. And Planet kind of does this. Like, you know, an individual farmer can go to Planet and be like, I want to buy a picture of my farm every day for the year so I can see what's going on there.

Do you plan on kind of getting into that sort of realm as well where it's like down to an individual, an individual buying this data? Is that retail satellite imagery, I guess? It's what it is, right? Yeah, you can do that today, actually. You can go to aurora.pixel.spase. You can create an account, and you can start tasking our satellites. So we are obviously a few years behind planet in terms of the maturity of the business

and what scale we are at, because they started in, I believe, 2013. We started in 2019, and we just very recently launched our commercial satellites. But we do have the same capability. Yes, you have a few retail customers coming in, sort of trying it out, but that's not really where the business is made. Obviously, we make more money in the B2B deals that we have. Always. But absolutely, I think anyone can come in and sort of use it on the platform that we have.

In fact, we've gone a step further today and integrated an LLM in the platform as well. So you can just actually go in and type it out and it'll tell you exactly how to place an order. And in a couple of weeks, you can actually type out an order and it will place an order for you without having to go through a clicking your mouse phase. Wow. I'm setting up our organization account right now, Jake. Oh, are you? Doing it.

Hold on. Doing it live. We're going to get a hyper spectral image of the next meetup is what's going to happen. See exactly how drunk Jake was. Yeah. What color did he turn? What hyper spectral color did I turn? Okay, so let me ask.

I was going to ask a hard. Yeah, because I still have to set up our organization. Yeah, I was covering for you, man. Come on. I'm going to ask you a really tough question then. What stops planet from adding a hyper-spectoral camera to their satellites going forward and then in totally eating your lunch? Like, what do you do to stay competitive with that?

Yeah, I think the good thing is that this industry, if you're building hardware in a space industry, it's by default you have a moat because it's very, very complex and time-consuming and capital to build something. Right? So even for a company as biggest planet. So let me actually, I don't have to speak in hypotheticals. Planet actually is wading into hyperspectral. They launched the first of their hyperspectral satellite last year. I think October was the month.

It was a 30-meter resolution satellite, 3-0. We have a 5-meter satellite up there. So we are about six times better while being cheaper and sort of all of that. But I think the reason we could do that is we decided very early on to build a really strong team of people that had worked on hyper-specicle, and there were not a lot of those folks out in the world. And we decided to completely focus our efforts on building that. and then go at it and we spent the last three and a half four years doing various iterations of hyperspectral camera we had to start at 30 meters like everyone else did like planet did like uh NASA did like isro did

and then we got into a second iteration at 10 meters resolution and that took a lot of effort and development from 30 to 10 and then finally that gave us all of the info and learnings we needed to build a 5 meter camera um so planet absolutely they they are a very talented team have massive respect for them so i believe if they devote their time and and money and effort they can do it but it will take them 12 to 18 months to be able to replicate sort of where we are because that's just how long it takes to test out on hardware, see how it's performing, use that knowledge to iterate and sort of build and launch it up to space. So even for a company of their scale and size probably an 18 to 24 month catch-up period, I would say, and that's what really means that while we focus on hyperspectral, we can keep that gap going.

I would probably only give SpaceX, you know, probably they can do it in 10 to 12 months. They're the only team that moves at grain speed. So I would say 12 to 24 months is roughly what it will take anyone to catch up because it's just so complex to build it and so complex to test it and then you can't just shortcut that. So you just got to keep moving then. Oh, that's the name of the game, I guess, huh? Cool. So then the reverse question, can you later add visible spectrum to your cameras and eat panel flight? We can have a multi-speccal cameras.

We can do it in the same satellite, but yes, we can replace the camera and, yeah, with a hyperspical camera with a multispectal camera. Absolutely. But I think we don't have a technical advantage there. Planet has done a good job with their satellites. We have a bunch of other companies like Black Sky and Satellogic as we're doing that. I think we do have a very big cost advantage being that we, you know, build a lot of our satellites in Bangalore. So we can do it for about one-fifth of the price to one-tenth the price of what it takes in the U.S. for some of our competitors to do.

So maybe at some point, I'm not saying, no, not right now. We want to focus on hyperspectral, but maybe in two, three years, we would want to expand to another. Yeah. So what are some of the, like, interesting use cases for this? You know, we've talked a little bit of kind of like hypothetically, you know, farms, mines and stuff. But, like, what is the, what is something that you just like,

you can't get anywhere else? Like, what really makes hyperspectral, like, such a valuable commodity, right? Great, let me talk about one of the agreements that we have in India with the Ministry of Agriculture. Now, India is a little bit of a peculiar nation because there's very, very small farmer holdings, right? You have less than probably a 10 meters square plot that one farmer is doing and then just beside that someone else, and they're all growing different stuff. So unlike in the US and Europe where you have some level of continuity that if wheat is growing here, wheat is growing sort of everywhere. If corn is growing here, there's a bunch of corn that's growing.

Not the case in India. So you need to be able to identify the exact. species of crop that is growing where and also the subspecies of crop. You know, is it this kind of rice or that kind of rice? Is it this kind of wheat or that kind of wheat? And I think so crop species identification becomes important as a census standpoint because they need to know how much of a crop is being grown this year and what is the potential yield going to be for every species.

And because of the spectral signatures that have a spectral camera can capture, it's the only data set that can tell you specifically with species of crop you're looking at and similarly for forest as well, which very specific species of tree you're looking at. I think that's very unique use case there. The other example alluded to earlier was crop diseases and pest infestation. So, right now with multispectral data that planet provides or other provides, you can do something called a NDVI. You can generate an NDVI, a natural differentiated vegetation index, which tells you how healthy the crop is. A rough indication of whether it's green, yellow or red, healthy, somewhere in between and not healthy.

But you can't tell exactly what's causing that and why is that. With hyperspectral, let's say you have wheat rust that has started some place, you can very specifically say that this crop is wheat and then it has rust disease and very specifically identify that from some other disease. And this happens about three to four weeks before the symptoms actually start to show up in NDVI. Before NDVI can tell you, you're actually looking at exactly what's happening within the leaves of the crop and that is a crop disease. And three to four weeks is a big, big difference for farmers to go and sort of make a difference in. how much it has spread. So that's a couple of examples in the agriculture domain. Maybe let's take mining.

Everyone's trying to figure out who has rare earth minerals, you know, is there cobalt here, is there lithium, you know, do we have uranium or thorium or anything else for nuclear purposes? Hyperspectral can identify exactly which mineral you're looking at, like very specifically. A very cool example I like to take is if you just search USGS, which is United States Geological Survey, Afghanistan, hyperspectral. You will see that the USGS flew a hyperspectral airplane over the entire country of Afghanistan probably a decade ago when they were still sort of occupying it. And then they found $1 trillion worth of new mineral deposits through hyperspectral data

that they were not earlier finding it through other prospecting methods. Imagine doing that for the entire world, knowing exactly which mineral is where, and so that the mining companies don't then have to go in deforest areas where they don't need to be looking at. So I would say that is a good example in mining. In oil and gas, it's about oil leaks and gas leaks. Whenever there's a methane leak, we can identify methane leak because it's invisible otherwise.

We can also see an underground oil leak. We can't really see the leak, but what we can tell you is that the soil characteristic has drastically changed. So this soil signature that we were seeing earlier, which was just soil, now has changed to this, which is because of the oil penetrating there, means that there is underground leak. So maybe a few examples that I can give you, but apart from that there's many more. We can identify how hot a volcano is to blowing up and sort of where the lava is. flowing, we can look at potential risk for forest fires, we can identify borders and see,

you know, who's moving around real tanks and fighter jets and so on. So these cases are many, but the data set is the same. Jeez. How does the mining one there work, like in terms of identifying minerals? Is that purely things that you can see from the surface, or is it similar to what you're talking about with the soil quality and the indications that you get on minute changes? Yes, so we can penetrate with hyperspectral data, so it's not a penetrative data set. We can only see what's on the surface. But most, most minerals have, you know, something on the surface. We can't tell you how big the ore is for that. We would actually relate and prospect it and see how big the ores. But you can tell that, you know, this mineral is a typical.

It's a typical kind of situation. Exactly, exactly. I've been found here. There are quite a few others that are not found on the surface, but then you have what you call indicator minerals for those. So, you know, if I recall the example correctly, I think wherever you find diamonds or gold, you also find Keoliite around. I would say we're able to find Keolini around. I would say we're able to find Keolian nitrish on the surface, you have a better indication that there's likely, you know, gold or diamonds around because of the geological processes, they are formed fair, you know, in fairly similar regions. Hmm. It feels like some Minecraft knowledge right there. I feel like that felt like I was playing Minecraft again, how to find diamond goals.

Can you imagine going back in time, like only, like, only a hundred years ago and like going up to a farmer and being like, hey, how do you find out if your crops are diseased? And they go, well, I walk out there and I look at the leaves and I smell this and I check this. So, you know, they're like, well, what if I told you that in just four generations, we're going to look at it from space with a machine that flies around the earth at 28,000 kilometers an hour? How do you like that idea? And Nose a month before you do. Yeah, like, can detect it. Actually, it was a very cool example.

So there are a few companies doing gene editing for seeds, right? So you have these seeds whose genes they edit. So whenever it's stressed or whenever it has a disease, it glows up in the infrared range. and hyperspetal cameras that we have can see the infrared range. So for satellites, basically, it's glow in the dark that, you know, it makes it even easier because of the gene editing that's there. It's very cool. That's wild.

All right. That is my insane fact for the next month. Yeah, so did they, I guess somebody realized they could pull this off, and so then they could make use of, was it hyperspectral data or infrared imaging that they were doing already? What was the, they had to know both of these things are possible. It's my problem, right? This is like a chicken or the egg problem.

Like, they had to know they could get this imagery so that they went and crispered it up in that. What could you like take a like a handheld, I don't know, you just walk around the field and shine out on leaves? Oh, nice. You can use a hand-red spectrometer to be able to actually see these things. I think the golden, the golden egg basically is if you can actually have crops really glue in the dark in the visible range when something's gone wrong. so you know visually but until then we are stuck with it glowing in the infinite range

yeah that one might be a little harder to solve also just sounds awesome generally like the plants in my house to do that like I turn off the lights the plants light the hallways like that just sounds pretty rad generally so there's a big market for that if yeah not just farming

yeah we can have our genes edited so every time we are stressed we glow up we'd have a lot of glowing people over on the sea of Earth. You would ruin a lot of dark sky zones, that's for sure. I'm still thinking about that. Blown dark plants. Edit the genes to make them to look.

Gene editing so that the satellites see them better is like such 2020s thinking. That's so good. I love that. I feel like there's a black mirror episode you can make out of all this. Can we talk about the hardware a bit too? Because I'm curious, you know, when you were, you're talking about controlling for cost and figuring out how to fit it in a smaller package. Was there a particular component or thing that you had to solve that was the trickiest about

that? And did it drive the ultimate size of the satellites that you're building? It did, yeah. So I think we took a little bit of an opposite approach to Planet. When they started with their dove satellite, they said, okay, satellites are getting standardized, so we will pick a 3U-Cube size and try and cram as best a camera as we can into this size. So the satellite size decided the camera size that decided the image.

quality. Whereas we decided, okay, I think, you know, they've already done the temporal resolution, but we don't need to sort of standardize it. Let's ask our customers what they want. So we first decided what the image quality needs to look like. That decided what the camera needs to be designed around it. So I think that was really where we could control how big the satellite could be, how cheap it could be and so on. So basically what you're building is a spectrometer. So in James Webb Space Telescope, you know, the second image that it captured was being able to say that there's water vapor content in this exoplanet that's many, many light years away. The reason we are able to see that is because it's a spectrometer.

You're able to see what spectra is coming in from every molecule and it gives you that. So, hyperspecial image is basically a spectrometer that image is done. Now, there are multiple ways of building a spectrometer. The most common way and the only way that was used earlier by NASA, where the space agencies was to build a spectrometer where you have the optics. So you need to have lenses and mirrors that capture the light and try to focus it to a detector. But in between the detector, which is converting your photons into electrons and then you everything that we see as zeros and ones, you have a spectrometer, which basically means you

have some things like gratings or prisms. So, it's like a prism. So you have light coming in and then a prism basically, you know, makes it rainbow. Similarly, it's prism but for hyperspectral where you have gratings and prisms that is splitting the light into many, many wavelengths. So this is a very voluminous contraption, the way that you sort of need to do that because you have to split so many wavelengths of light and then you're losing a lot of quality of imagery because the dispersion of light is causing the quality to drop and go on.

We decided, I think, we looked at and researched what was happening in material science and detectors and spectrometers and all of that, and we thought I think there's a different way to do it from space. And before us, no one had sort of done that. You have the optics which remain the same. You need to basically capture the photons and then focus it. So that remains the same. The detector is a detector which converts photons into electrons. But instead of all of this contraption in between, we said we will just deposit these filters on top of our detector, which due to their material properties reflect all the light, they let only a specific wavelength go through depending on their thickness and depending

on the materials. So, if you lay out a series of filters, basically, you know, think of it like yellow sheets one on top of the other, a combination of them will let light through of different wavelengths. So the detector now is capturing hundreds of wavelengths because of varying thickness and material properties and so all of that. So basically all that to say that instead of a very voluminous contraption called the spectrometer, we decided we built filter-based type of spectral cameras and that took a lot of time and effort, but that made the camera much smaller, because you don't have all of these things in between. And after we had spent the initial R&D money on sort of figuring it out,

the camera also became cheaper per system to be compared to a spectrometer. And so the smaller camera means a smaller satellite. A smaller satellite means a cheaper satellite. And therefore, a smaller satellite also means lower launch cost because you're paying your launch cost on the basis of how big your satellite is. And are there any tradeoffs with that using that instead of dispersing the wavelengths across the way the spectrometer does? Is there any drawbacks or things that you're losing by doing it through the filtering mechanism?

You just need to use more lines that I think double the number of lines of the sensor compared to normal spectrometer. But you basically have everything advantages. Your signal to noise ratio is bigger because you're not dispersing light. The camera is smaller. Eventually it's sort of cheaper. So you're not losing anything on the quality and you can make up for the number of lines. So there's the only tradeoff that sort of comes with that. Yeah.

Do you think that has, is this an area that is going to like come from the space industry and be useful not on satellites too? Like is there, could you see this going into either other packages or other industries in a beneficial way? Yeah, yeah, I think absolutely. I think hyper, spectral cameras have a lot of... iPhone 17. Hopefully not on an iPhone.

But I think you can use it in the medical domain quite a bit, like Hyperspectal. spectrometers are used for identifying cancer cells from normal cells because they reflect different spectra. The glow of the dark, yeah. The glow in the dark, cancer cells, yeah. I saw that when they designed COVID in that lab, they made it glow in the dark, actually.

Yeah, exactly. Yeah, but it's also used in the food processing industry. You can basically keep it in the food line and then see if a fruit, fruits that are coming out of there are rotten or not from the inside or basically any other kind of food materials. So I think those are the two industries where there's a lot of scope for uptake.

That's cool. So tell us about your first couple launches here. So you have some hardware in space. I want to hear about what, well, I think you had like a first or second one kind of a while ago. And then recently you've got like the big three pack, you know, of the demonstration set. So yeah, tell us about that. Yeah, so we started with a demonstration satellite in 2021. That was a 30 meter satellite.

So we matched what NASA had done and a few of the species. things are done. We did it at, you know, about 120th the cost and a smaller size and all of that, but we were standing on top of giants, NASA basically invented this technology and so on and so forth. So, but 30 meters, we said we're going to match what was out there, but we're going to do at a fraction of the cost, and we're going to do it at a very small satellite. That was done in 2021. And then we used the learnings from that to launch two satellites in 2022, which were of a 10-meter resolution. So we made it three times better. But all of these were demo satellites. They were not meant to last there long. Their lifetime was roughly about 18 to 24 months. They were supposed to beam down just a few

images, you know, a week and then, you know, the revisit time over any area was roughly two weeks or so. It wasn't great for commercial use cases, but it proved the entire technology that we wanted to give us a lessons. Finally, all of those lessons culminated as building a five-meter camera, which went on a hyperspectral satellite. These were our commercial satellites. So we're calling them the fireflies. There are two constellations that we're building the fireflies and the honeybees, the honeybees will come later, they're bigger, better versions. But as of now, the fireflies were what we built last year, and there were six of these satellites.

Why did we build six? Because with six satellites, we can revisit any point on the globe every 24 hours, which means if I'm taking an image of Philadelphia today, I can come back and re-look at it in exactly 24 hours. And that holds true whether Philadelphia has changed with New York or whether it's Los Angeles, whether it's Tokyo, Auckland, anywhere in the view. So that is when six was important to us because it got. to a level where we could deliver a daily image like planet does to any of our customers

anywhere and do it globally. And out of those six, in January, we put three of those satellites up there. So, three more will go up around April or so next month when the SpaceX launch goes up, and three went up in January with the SpaceX launch. The good thing is all those three satellites are working. They have a seven-year lifetime compared to an 18-month lifetime that the demos had. They have more than 200 times daily capacity to beam down data and imagery. and to capture a lot more area.

So I think that's the launches that we have done until date. In fact, a very cool fact is in 2022, when we launched our 10-meter hyperspectal satellite, there was one more hyperspectal satellite on the same Falcon 9 rocket. This was a satellite that was being built by the German agency, DLR. The difference was that satellite program started in 2002. 2002. I was five years old then.

We started building that satellite in 2021. their satellite cost close to 100 million euros to build over the course of those many years. It weighed close to 1,000 kilograms and it was a 30-meter resolution satellite. Our satellite weighed 20 kilograms. It was three times better in terms of its resolution at 10 meters, and it took about 14 to 16 months to build, right? So I think that's the new space industry where you can sort of move fast and sort of, you know, it would have failed for sure. We didn't put in the same level of reliability with that probably the,

the bigger satellite did, but I think when you move fast and do something, then you can move you better faster. So yeah. Yeah, yeah, that's the name of the game right now. That's the grand unifying space narrative right now. These companies move faster. Yeah, I'm curious, because you went up on a transporter, right? It was like a ride share thing, right?

How does that work with your, I always like to hear it. customers talk about this just because the concept of rideshare, I'm fascinated by it. It seems to be like such an interesting point in time right now. We're at this kind of like golden era of ride share. But, you know, the drawback of course. And occasionally they sound like total messes. And occasionally they sound like total messes. Everybody comes off at the same time and it's chaos.

Yeah, yeah. And like, you know, your satellite comes off and sees another hyper spectral satellite. So you kind of like elbows up and, you know, get them, get it on the way a little bit. Bumper cars. See if you can knock it out. But no, just getting to. Getting to your target orbit is always, of course, the tradeoff there, right? Because everyone wants to go somewhere different, but the rocket's going to one place.

How do you handle that? And I'm just kind of curious to know your perspective on that tradeoff. Like the cost savings you get from a ride share versus the extra work in time or whatever you have to spend to move your satellite to its destination. Yeah, I think that's a good question. So SpaceX is by far the cheapest launch provided in the world. right now. So, like maybe a history lesson before SpaceX came in, the price per kilogram for a right share was roughly between $20,000 per kilogram to $25,000 per kilogram. And that was a cheapest out of it. And then SpaceX came in and we will do these ride share launches and the price is very

straightforward. You just make sure you come on time, you fit sit in your slot and we will put you in an orbit and you pay us $5,000 per kilogram. So now inflation adjusted everything now is roughly about $6,000 per kilogram, but it's still about one-fifth of $25,000 per kilogram that we had to pay before. So that changed the game and now companies are trying to bring it down but you can't without losing money. So the options are either you get on a right share with someone else and you're still paying two times to three times what SpaceX is charging and you're still a right share

and SpaceX is launching so often now that you have four launch opportunities which are dedicated right shares, more opportunities if you're willing to go on someone else as a primary satellite. So there are very regular launches and so you know that SpaceX will go up on time and generally it won't blow up. That's not the case with others where you don't know if they're going to go up that often because you're dependent on the primary satellite. The other option is you can book out a dedicated launch vehicle for the smaller satellite

like a rocket lab, you know, electron rocket for example or any of the new ones that are out but then that becomes way more expensive because you're not paying for an entire rocket. You can launch only about 150 to 300 kilograms depending on which rocket you're looking at and they end up costing between $6 million to $8 million or $10 million per rocket because you have to pay for the entire rocket. You can't have a right share but you control the timeline, you control exactly where you're going to. Instead of paying about and we did this cost calculus it's roughly five times to 10 times more expensive to do a dedicated launch and it is to go with the right share.

So the trade-off ends up working for us because we just add an additional bit of additional fuel on this satellite so that we can move up and down in the altitude if we need. You know, the dropers are a certain altitude. We just need to make sure that we collect for the altitude and go out on and then we operate from there. Plain change maneuvers are obviously different. We just tend to choose those SpaceX launches which go to a sunsynchronous orbit because we need to be in a sunsynchronous orbit.

We don't go on the launches that are going to the inclined orbit. And generally it works out because most imaging satellites have to go in sun synchronous, the rest of the folks go on incline. And as long as you've built your satellite to make sure it can survive in space and it doesn't matter if 100 satellites are coming out of there in orbit or so, you can make sure that you're communicating with your satellite like you've been able to do all the last few weeks. But I think it's like it's basically a train. You just have to make sure your satellite is ready on time and you show up to this basic facility on time and they will deliver you from point A to

point B. Now you figure out where you want to go from point B. Yeah. Yeah, I guess with certain key orbits, you know, sun synchronous being one of them. I imagine, although geo would theoretically be one, although I don't know how much use there is for smaller stats out there. But, you know, these kind of like lynchpin orbits where they're like, there's a very specific inclination you go to and everyone's in the same spot. And that's why it's useful, right? That seems like the optimal use case for ride shares. Because like you said, like you could have an entire ride share where everyone wants to go the exact same inclination. And, you know, there's no compromise at that point, right?

almost every imaging satellite that has global coverage requirement goes to a sunsynchronous orbit, that's like a given. You just know how to figure which of the rockets are. Then out of the four ride-share rockets a year, at least two of them or sometimes three of them go to a sun-signulous orbit, which is a fairly good optionality. Yeah, yeah, yeah. That's the popular one.

It's interesting, right, because some imaging constellations like having a couple of satellites at that mid-inclination orbit, right? Black sky likes to have those 45, 50-degree inclination. they want to cover where most of the humans live on Earth more frequently. So they've got a couple, I think they have a couple of Sun-Secretists, but then they've layered in their mid-inclination. Is that combo something that you see in the future,

or is the revisit time being able to compare, like, one-for-one over the course of a day or several days, the benefit there from being in Sun-Sikikis, that you're getting the same local time every time you come through? Does that outweigh the more frequent revisits you get at mid-inclination? Yeah, I think it depends on what kind of constellation you are. We are more of a monitoring constellation. We monitor almost every part of the landmass every day.

Like, we're going to take an image of the Amazon rainforest almost every week, regardless whether someone's asking for it or not, right? Because we have a lot of capacity and we need to be painting the earth to see how the snapshot of the earth has changed over many months and years. With black sky, they are more of a tasking satellite. You know, you have clients coming to them and telling we need to image this, this, this, this. and so they need to have the revisit capability to come over that area because they are more focused towards defense intelligence because our infrastructure that's the kind of satellite that they're building and launching.

So, I think it depends on the use case. So for us, the core constellation will be a synchronous satellite because that gives us optionality anywhere on the globe, as well as to not just to be able to deliver an image that has been tasked, but also just monitor the earth and what's happening. But once the core of 18 satellites is built out for us, we also plan on, you know, springing around a few in the intelligence. land orbits because then we get very highly visited over certain areas that then we can use for tasking programs. Yeah, you can be opportunistic about those, whereas the others are more kind of steady state operations.

Yeah. Yeah, because you're right. Like, you know, there are some of the use cases we talked about, agriculture being one of them, that will be concentrated closer to the equator than it will be to the poles, right? We're going to have more agriculture happening in mid-nottitudes. I don't know, man. It's getting better to grow shit higher up on the earth these days. That wind goes expanding a little bit.

Well, hopefully areas can find all the methane leaks and we can slow that one down a little bit. Yeah, I think it's beyond our control now. We can detect all the leaks that are happening, but then the Earth's on our different trajectory. Yeah, still got to plug them, right? Oh, wow, cool. Okay, so you said there's another launch coming up in April or so

we're looking forward to? That's right, yeah. April or May, depending on when SpaceX sends that up. the exact date isn't sure up yet, but sometime in the next couple of months. That's cool. All right. So you've got three went up. The next three are coming.

You said 18 is the initial constellation. Are you building that whole batch as a single run, or are you still working on the later half of that? Built six already, so that was the initial run. Now we're building 12 mode. And then is there a kind of cadence that you expect beyond that? I don't even know how many kilometers up you got, right? These are not coming down. Sunsigris, you're typically going to be up.

six, seven, eight hundred kilometers. I assume you're somewhere up there. Five, all right. But you're not in like the, I guess you are kind of roughly in the Starlink territory where you're expecting a certain amount of de-orbitz over every couple years? We are, we get a lot of, you know, there's like space traffic control, so we get alerts every time a satellite is going to whizpasters, and by whizpastas, it's still a kilometer or two kilometers away, but in space it's close. And so like almost every other day we get like a conjunction analysis that there's a potential for clash with a Starlink satellite or some other satellite, mostly it's a Starling satellite, but thankfully, you know, either they can move out of the

or we can out of the way, and in most cases we are passing a few kilometers away anyway. It's just like a, you know, risk-free thing that they do to just give us an alert. But we are in Starlink territory, but Starlings now also operate a little bit lower, I think, 500, to be precise, we're a little bit on the higher side, so it's not as crowded. Yeah. Is there, so, but does that come with, you know, are you expecting these to last five or seven years, is there a deorbit date that you have in mind when one of these comes off a launch? Like, how long do you expect that to stay up?

Seven years is what we've designed it for. So there's a design life. So we're hoping that on the minimum they stay up there for seven years. And then the fuel runs out for them to be maintained. So we need to maintain the altitude for which most of the fuel is kept there. So that's seven years. After that, we can still continue to operate the satellite as it keeps coming down. And so we also keep a small amount of fuel at the end for deorbiting purposes.

So once, let's say, seven years of operations at the same altitude, maybe two more years or three more years of operation as it's gradually reducing its altitude, and then as it gets towards 450 kilometers or 400 kilometers, and then we use the final bit of fuel remaining to bring it down completely to about 300 kilometers, and then it burns up in a couple of weeks. So when you think of that from a sort of like a tech roadmap perspective, right, when you've got these 18 that you're going to launch, but the generation after that when these are going to start deorbiting,

how do you think about the timeline for what tech is going to go in that next run of satellites? have some time to learn from this round and adapt to whatever you're going to build next, or what kind of lead time do you need to actually start constructing those? Absolutely. So I think the six go up this year, then we will launch 12 more the next year and the year after that. And so by mid-2020-7, we will be completed with our first phase of the constellation. And then we are good for five years.

We're not going to look at replacing them for the next five years. But what we will continue to do is we went from 30-meter resolution to 10 meters to 5 meters, and now we're trying to go from 5 meters to something much better, maybe one-me, meter or submeter. So the next batch of satellites, if we will replace, there won't just be five-meter satellites. Maybe we will launch more five-meter satellites as well, but along with those we will get towards one-meter satellites or something better, where the resolution has gotten better now. But that's probably, you know, we'll have three years or so to sort of figure out how to get there because it's not immediate.

It's just such an interesting timeline where you're, you know, like, it's, you know, in the tech industry, it's like, I'm going to work on the next innovation and it'll be ready when it's ready. and but with this kind of satellite constellation, it's like, well, I know the clock's running. So I've got three years to figure out what I can figure out before I got to start bending some metal. I got to invent something in the next three. Yeah, yeah.

There's a little bit more time pressure than like, man, I hope I can figure out how to make an iPad, you know? It's like, there's a harder constraint on it in a really interesting way. Yeah. And then like, do you get, you know, if you don't, if in the event, not you, I'm saying you, but not you in particular, right? but like if a satellite designer, if you don't make enough progress on how good you can make the next version,

do you just build the one that you just built again because it's cheaper to run that again than it is to like kind of build a halfway completed new version? If it's still print money. I think it would be both. We will definitely replace the 5-meter satellites with more 5 meters because the resolution requirements don't really change or not. Use case still remains the same. So 5-meter multi-spectral has been going on for the last three decades, right?

and while there is better resolution imagery that's available on the market, the 5-meter imagery still sells. So you're right in the sense that we would keep on the machine for the 5-meter data to keep coming down. But at the same time, you know, when we are trying to replace them, we would rather have something better and not just the same thing. But you can always launch the same thing. If the technology hasn't reached a point where you can do something or something new hasn't been invented. So 5-meter, I think, as a resolution for multispectral or hyperspectral, will always continue to remain the case unless you. one meter or something better, it becomes so cheap and so quick wishes that five meters, there's no point to it.

Right, right. Yeah, and then even then it's interesting mix where, you know, maybe you're kind of baseline monitoring like you're talking about stays five meter, but the opportunistic tasking satellites are at one meter because if somebody wants a picture of what someone else on the earth is up to, they probably want the most resolution possible, whereas monitoring in environmental situations or changes over time is fine at five meters. Yeah. Yeah.

Yeah, I guess the use cases aren't like, it's not objectively better to, have higher resolution, you know, presuming the costs aren't identical. So I guess, yeah, the, hmm, that's interesting. Yeah, okay. You feel you're feeling better about imagery, Jake? I do, yeah, yeah, I mean, I'm thankful that we sort of danced around and didn't really get too deep into what hyper spectral is because that was where I would have my lunch taken from me, but.

You still don't want to think about too much. We just, we just make the assumption that we all know what we're talking about when we say hyperspectral, and then we can just move on to the business stuff. That was great. That's perfect for me. It does sort of feel like trying to explain somebody the fourth and fifth dimension, right? You're right. Yeah, exactly. Well, I know red, green and blue. Like, is there more? How much more could there be? I can't imagine it with my mind's eye. Like, well, what color is it? It's like, nah, na-na-na-na-na-na.

Can't answer that. You've read the Flatland story, I assume, right? Yeah, yeah, yeah. And Jake? Flatland? You don't know about Flatland? No. Read it. It'll explain your struggles with hyperspectral imagery. Okay. I don't even know if I could do a good Cliff Notes version. So there's like, when did it start? It started like line man or whatever, right?

And then circle goes through his plane of existence or something. Oh, okay, okay. Globe goes through the circle's plane of existence. There's like the other shapes of other dimensions are passing through the other shapes dimension in a way that they only see it as a whole thing. Just look it up. Read the freaking computer. It's like a digital act in interstellar as well.

It's like a 3D representation. of time as the fourth dimension. There it is. That's a better description. Speaking of which, I watched that Lightyear movie last week with my son, and I realized Lightyear is just interstellar made by Pixar. So, check that out. That's the

Buzz Lightyer origin story. It's just straight up what it is. So, yeah. See how that review sits for all the parents out there. Yeah, okay. Cool. Well, I waste, this has been enlightening.

It sounds like you're doing some really cool stuff. Congratulations on launching the Sats. That's awesome. And hopefully those customers start rolling in like crazy and you can do some wacky stuff. And when you have the glow-in-the-dark plants figured out, we'll have you back on. And we can...

Totally. We can do that. What's your hiring situation, right? You said you've got a president in India and in L.A. area. What's the... Should people be checking anything out? I'm sure a lot of people that are listening are always looking for new days. Yeah, we're hiring both places.

So we hire Pan-US. We hire predominantly in Bangalore and India. We have roughly about 30 people in the US right now in total, but they are in Los Angeles, they are in Colorado, they are in the Washington, D.C. area. So, I mean, if you're working on hardware, if you're working on image analysis, if you're working on business development,

we're always looking for folks. And same in India, you know, we're always looking for more satellite engineers and, you know, analysis engineers in Bangalore. Cool. All right. Check them out. Nailed it.

Jake, I'm pulling up our calendar because we've time shifted this episode. So I believe next week as you listen to this. Oh, yes. Jason Snell is coming on the show, Jake. This is a Anthony achieves a long-term dream of his own podcasting heroes coming on his podcast. Yeah. So look at that. We're completing the liftoff nominal crossover.

We've got Stephen Hackett. And then now we've got Jason Snell coming on, dusting off the old space. takes from Jason. So I don't know if you've read a lot of Jason Snell in your life. I don't know how much of a snow. I listen to the kind of guy you are.

But yeah, yeah, I'm not, I'm not into it like the way you are. You're not as much of an Appleman as I am. No, no, that was the crossover that never reached me well, that was all going on, right? Yeah, totally. Awesome. Well, that's what we got, folks.

Always thanks again for hanging out with us. Awesome. I would love to have you back on to talk. about the Indian space economy generally because I would love your take on what's going on. We didn't talk about launch out of India at all. There's some launch startups happening. I would just be super intrigued to pick your brain.

So we should do that next time. Very fun, very casual compared to most podcasts and fun as well. Thank you all so much. We'll see you later. Bye. Hi, everybody. One, two, three, four, five, four, three, two, one,

and the test.