Episode 44 of the Space Industry podcast is a discussion with Michael Aspetsberger, Head of Aerospace at satsearch member Cloudflight, on building digital products that incorporate AI.
Contents
Episode show notes
Cloudflight is a digital innovation and solutions provider working across many sectors, including space.
In this episode we discuss the applications and processes involved with building space-related digital products that incorporate AI. We cover:
- The key roles that data processing algorithms play in satellite missions and services
- How AI can help when hardware alone isn’t adequate and you need to use routines that go beyond standard toolkits
- What companies need to think about when they are building and developing products using AI
- Other ways that AI and advanced processing can be utilized in space mission development and manufacturing
You can find out more about Cloudflight here on their satsearch supplier hub.
The portfolio of Cloudflight
Space data processing and analytics
Cloudflight’s space data processing and analytics services are designed to create bespoke data products to maximize the chances of mission success.
Cloudflight works with prospective clients following an agile approach; assessing, evaluating, and proposing possible solutions. The end result is a minimum viable product (MVP) that is tailormade to the client’s (and their data’s) specific needs, maximizing the chances of mission success.
Supporting new space business models
Cloudflight supports new space business models by helping companies test and develop service concepts suitable for the modern market.
Cloudflight has the expertise to harness digital tools and capabilities in order to develop and evolve business models better suited for today’s market conditions.
Cloudflight can also provide access to additional data streams to satellite operators and startups to support any R&D department, thus helping in finding new, innovative business models to maximize the success of a company.
Algorithm development and optimization
Cloudflight’s algorithm development and optimization services are designed to improve processing and mission outcomes.
Algorithm optimization is the process of finding the input parameters or arguments to a function that result in the minimum or maximum output of the function.
One of the main tools to achieve this is machine learning (ML), an approach that enables the efficient processing, organization, labeling, and categorization of information, making it easier to learn from and/or apply to real-world problems.
One example of algorithm optimization in action is Cloudflight’s partnership with GRASP, the world’s most complex algorithm for aerosol detection. GRASP is a scientific research project that uses Earth Observation (EO) data on the effects of aerosols (from sandstorms, volcanos, wildfires, pollution, etc.) on the global climate.
Predictive maintenance
Cloudflight’s predictive maintenance services are based on the company’s expertise in aerospace manufacturing, a field that strives for “zero failure production.”
While high-resolution, multi-dimensional datasets describing the condition of a machine or building component offer an unprecedented level of quality, they’re also challenging and time-consuming to inspect.
Cloudflight integrates its knowledge of big data analysis and process automation with technologies such as computer vision, augmented reality (AR), and machine learning (ML), to build applications for intuitive and accurate quality assurance inspection workflows and to optimize manufacturing processes in general.
Episode transcript
Please note that while we have endeavoured to produce a transcript that matches the audio as closely as possible, there may be slight differences in the text below. If you would like anything in this transcript clarified, or have any other questions or comments, please contact us today.
[00:00:00] Hywel: Hello everybody. I’m your host Hywel Curtis. And I’d like to welcome you to the Space Industry by satsearch, where we share stories about the companies taking us into orbit. In this podcast, we delve into the opinions and expertise of the people behind the commercial space organizations of today who could become the household names of tomorrow.
Before we get started with the episode, remember, you can find out more information about the suppliers, products, and innovations that are mentioned in this discussion on the global marketplace for space at satsearch.com.
[00:00:32] Hywel Hello and welcome to the episode. Uh, I’m joined today by Michael Aspetsberger, Head of Aerospace at Cloudflight. Cloudflight is a digital innovation and solutions provider. The company works across many different industry sectors, including space.
Now, today I want to discuss a real sort of trending topic for many people. Different levels of the value chain in the industry. And that’s this concept of building digital products that incorporate AI using space data or based on space-based data.
And, um, I think there’s lots of, uh, applications to this sort of technology and there’s lots of different ways that companies are assessing it and testing it. So, uh, it’ll be good to get into that detail. So Michael, hi. Welcome to the podcast today. Thank you very much for being with us.
[00:01:14] Michael: Thank you, Hal, for having us here. I’m really excited and I hope we have an exciting, uh, podcast ahead.
[00:01:19] Hywel: Fantastic. Great. Well, let’s get into this, um, this discussion. So firstly, my question is, in, in your view, what do you think are the key roles that data process and algorithms play in satellite missions and, and services, and how can AI sort of accelerate their development?
[00:01:34] Michael: So, as you have mentioned, I mean we as Cloudflight, we have seen many different areas over the years, Right? We have been working with the public agencies, We’ve converting with the private new space industry, both big corp, uh, and small startups. And the one thing that we have really seen is that the data processing is the, the heart of the value chain, right?
All the data you record from space, you need to transform it to have some benefit for your ultimate customers. And so the data processing really is the driving force behind generating that value. And what we have seen there is really artificial intelligence can be a way to be accelerating the development of this data processing pipelines it, instead of having to spend an awful lot amount of money and, and effort and, and resources into trying to understand certain aspects of the system, you can really try and prototype quicker with AI, right?
You can try many more things more rapidly than you could if you would model the entire system, uh, using some, uh, classic procedures. And at the same time, also with the new capabilities that you get artificial intelligence, uh, you can also compensate for some of the deficits that, uh, your, your specific hardware hand has.
[00:02:47] Hywel: I think. Um, Drivers as being seen in other industries as as well. I mean, you mentioned rapid prototype in there. Well is one of the key value ads of something like additive manufacturing. This what company’s talking about all the time, and this is on the data processing side, so, so that’s, yeah, very, pretty interesting.
But this relationship between software and hardware is important. So again, in your view, how can AI help when hardware alone isn’t adequate, isn’t enough in a satellite mission, you need to go beyond and you need to, you know, use routines to go beyond standard toolkits.
[00:03:19] Michael: Yeah, absolutely. So if you compare against like the classic, uh, way of building the, the, the big satellites, right?
In the, in the, in the past it was, you would try to get the best possible thing you can, right? You would try to get the best possible pointing for your satellite, the best possible calibration with the FUS and whatever. You try to really have an extremely well calibrated and well-built instrument to have an A package that can generate the maximum value out of what you have in orbit, basically.
Now, when you want to build a full constellation, that won’t work, right? You have budget constraints, right? You can’t build a hundred satellites with the cost of a hundred or 200 million each, right? You have to make compromise. The constellations. They’re not only driven by pure scientific demand or by what the taxpayers or whoever decide, but really by, by the industry.
You have to understand what is the, uh, what is really the benefit or the value. That is to be generated ultimately, and you will make some compromises, right? For some early demonstration, it might be sufficient if you have, uh, a degraded pointing or if your instrument calibration is not perfect, right? All that matters is that you can generate something and, in these fields, when your data is not perfect.
I think this is particularly one of those areas where with AI you can try more novel approach to some other strategies. For example, if your pointing is not accurate, By the advances that, um, have been done over the last years, you can now use some AI co-registration routines or you can use some feature extraction routines to try to better understand what place of earth are you looking at and how does it match with, um, historic observations you have done before.
Right? And try to, uh, sort of build that, the product based on the data that you have attend. At the same time, there’s other techniques such as, uh, data assimilation and, and also data fusion. And, and all the way to super resolution where recent, the recent technological advances in AI can really help you make a difference in the product that you have.
[00:05:20] Hywel: We talk about all different aspects of, uh, earth observation industry and, um, ensuring that you are actually imaging the location that you want to is a key part of this.
[00:05:30] Michael: That’s really critical for, for getting a good product. Because if you’re doing some crop monitoring product right, then you’re monitoring the parcel of your neighbour.
Uh, whatever index generating out of it, it will be extremely beneficial to him, uh, but not to you, right? So it’s really critical to get a good geometric accuracy and it’s got really interesting, really important to get a good radiometric accuracy for the, the more sophisticated algorithms you might be running on it.
But then again, you need to have a product first before you can get to the point about being picky on the, on the data quality side.
[00:06:00] Hywel: Yeah, absolutely. Absolutely. And this is what you mean about this hardware and the software, you know, going hand in hand and working together. I’m very pleased. Also, in the previous answer, the previous question, you used the word compromises.
Because this comes up in every podcast, every space mission. There’s only so much that we can launch, that we can operate, that we can power, and that we can fit into the, uh, the, the volume of the satellite and everything involves a trade-off and optimizing the entire system for the key performance criteria that you want that will earn value to your mission or to your service.
So working within those constraints, how can data processing help or accelerating data processing, how can this help to reduce costs and potentially also mitigate the impact on the environment? And do you have any examples you can share with us on this?
[00:06:48] Michael: Yeah, absolutely. I mean, data processing is specifically the component on the ground, right?
I mean, as I said before, the data processing per se, is the heart of the value chain, right? And you can do some data processing in space, right? There’s more and more edge computing going on there. So there is certainly a space component to it. But the big data problem typically is on the ground, right?
Because you’re collecting the data and you’re reprocessing the data over time. And in many cases, the, the algorithms that are in play there. These are the very scientific POCs, the proof of concept you did a long time ago, Right? And you tried to get the fit right and you, you expanded on it for those that were working fine.
And often there, the focus on getting those proof of concept done was not to get it to be done efficiently or get it done fast, but it was to generate the actual product out of it. Right? Like the quality of the, the product was the, the primary focus. And now if you bring that poc, the prototype to production, and particularly you most likely are going to do it in, in the cloud, right?
Because cloud gives you an abundance of resources, right? There is no shortage on, on that side. A simple way of bringing that to production is just to randomly cloud and throw as many hardware on it as as necessary to produce the results. And there’s certain costs attached to it when you want to have like near real time product.
And there is a different cost attached to it. When you want to have a new generation of, of data product, a new algorithm version coming in, because you need to look at the entire history and, and regenerate the results. Of course you can throw a lot of more resources on it and try to solve it that way, but as you said, that’s going to drive your costs up.
And one of our core offerings to, to the European Space Agency, for example, was to accelerate algorithms. Right? Because the, the thinking there is quite simple. If you invest a certain amount of time in making it faster, you will reduce the, the time it takes to run to, to do the data processing per se.
Right? And for as long as the initial investment is a quantity, that is handleable the benefit that you continuously generate when you put this in operation. Will basically be being offset by this, right? So if you improve something by, um, a factor of two, uh, you will have half the, the infrastructure cost going forward, at least on the processing side, not storage, but processing side.
If the initial investment was there only a few, 10s of thousands or a few hundred of thousands, that could really be something. That, um, has a big impact on you in the long run on the financial side.
[00:09:16] Hywel: Particularly if it’s a constellation, right?
[00:09:18] Michael: Absolutely. Now, historically, I admit that we have only been looking at the, the financial side, right?
Because many of the organizations, of the companies we talk about, they’re in the early stages, right? And they, they are looking specifically to optimizing their costs because that is their, they’re basically their runway, right? Whether they, they still have, uh, time to grow or not. But at the same time, the maturity of the cost that you spend in the data center is on the energy side.
Right. And this was already in the past, and it’s going to be much more going forward now with the, the current, um, situation A across the globe. The energy shortage. Yeah. Yeah. So at the same time, when you reduce the, the cost on the infrastructure side, you’re also reducing the, the amount of resources that it takes you to generate it. Right? Like the electricity that ultimately powers the, the data center. And so improving those algorithms and making them more efficient, making them run quicker. And maybe also, you know, optimizing algorithms is not just by trying to replace the hardware, they run on with more specialized hardware that’s maybe more efficient.
It’s also about redeveloping the algorithms, reduce their complexity, right, And make them more maintainable and. More easy to use to resources attend. So all this, by reducing the infrastructure resources, you also, uh, reduce the strain on the environment because ultimately the power that you need in a data center needs to be generated somehow.
And there is only so much energy you can get out of renewable energies like wind or solar. So whatever we do to reduce the costs, And in turn, the resource there. It also has a positive effect on the, the environment for us.
[00:10:55] Hywel: Yeah, that’s great. I think that’s a really interesting story. You know, the, as you say, the, the global situation at the moment has thrown into sharp relief, the costs of energy, of power in anything from your home to your business.
And at the same time, we’ve seen the, uh, the effects of climate change continually hitting home and companies are looking at how they can become more sustainable, say they are more sustainable, all these problems, and ultimately it comes down to reducing resources if you can.
[00:11:26] Michael: Absolutely. Because you also have to be, be frank here, I mean, many of the constellations are being launched, many of the products that are being generated there.
They’re often for purposes of monitoring specific characteristics down on earth, right? And many also with the intention to provide more insight into climate change and the environment and how well we treat the resources we be that we have on the ground. But at the same time, right? If you have those bold goals, which I fully support, right?
You shouldn’t be sort of compromising them, but with the way how you treat the environment yourself by by, by processing massive amounts of data and storing massive amounts of data. So you should be sensible in how you do those steps to make sure you are not the bigger impact on the environment than what you’re trying to, to monitor and improve.
[00:12:11] Hywel: That’s great. Actually, that leads into my next question. So, and you, you’ve mentioned a few things that companies need to think about when they’re developing products, you know, based on these sorts of technologies.
You talked about the idea that the proof of concept is a different situation to the operational system and approaching your, your sustainability goals and, and ideals with, with these technologies in mind can help you achieve those aims. Do you have any, any further advice for companies you know, more generally based on your work in the industry?
[00:12:40] Michael: I think the one advice I’m giving the most of the times, and I’m, I’m always giving it because I think it’s the most valuable one that I can give, is to really try out things right and try to find a fit on the market.
Now that’s, that’s much easier said than done. So I know that saying that per se is, is, is kind of empty. So what I, we are not trying to give examples, what I’m trying to say is you will need to run many different prototypes for your product, right? You need to have a way to rapidly try out new algorithms, new data products, generate new indicators, new information from the data that you generate.
And what I see in many different startups is often there is a focus on having something in space, right? Uh, trying to build a satellite or trying to build a very specific, uh, instrument. And then there is the expectations that once you launch that instrument, once you launched its satellite, uh, the business will come automatically.
But in, in, in nine of out of 10 cases, that’s not going to happen. Right? You need to have a really, a really niche product that nobody has done before to make that work, right? In most cases, there will be some kind of product that is somewhat similar. Or at least some sort of observation that is similar to the one observation that you generate.
Often, the in orbit characteristics of the observations you make, uh, might not be really what you had intended initially. You also had to make compromises to get it up there, right? So what I’m trying to recommend is when you have something in orbit or there is some data that is similar to what you want to have in orbit, try to build your product based on that.
Try to have your algorithm prototypes in place and try to have many different there because most likely the product you’re going to be successful with ultimately on the market. It’s not going to be exactly the one product you had envisioned initially, right?
So trying out many different results there and be able to rapidly adapt and extend on them. I think that is key to have the success that you will need in the early stages. Once you have that in place, once you have the found one or two of those cases, which are giving you a fit, then you really need to bring that into production, right? And you need to extend on it, but only do that for those where you see really a fit.
Now when you want to try out many different prototypes and when you want to, uh, polish them up to bring ’em into production, that is a quite labor intensive, um, effort, right? For one, you need to find the right staff to do it, which is not easy. I mean, we ourselves are very limited in finding good stuff cuz it’s specifically in the IT side, uh, and specifically the IT on earth observation side.
You’re always in a tight competition on the market. It’s not easy to find stuff there. And at the same time, you don’t want to move your core staff around, right? Because you want to have the core staff focus on your core activities in, in, in your organization. So you don’t want to put that at risk by just trying 10 different new products too.
And so our recommendation on that side is we don’t recommend to simply buy some off the shelf solution or buy some completely external component for it. Our recommendation really is to enter into a corporation there and try to use outside resources when you can, when it helps you, when the staff is the limiting factor.
But try to keep the the knowhow in house, right? You don’t want to be vendor locked on a specific organization or a specific product, because that will expose you for your further growth, right? It will limit you. But bringing the right party in to um, sort of cooperate with you on, on trying out these different prototypes can be a way to stay flexible and agile because that way you’re also not tied up with a long term engagements because I mean, hiring people and, and firing them, if something doesn’t work out, that is going to put a lot of stress on your internal organization as well.
But if you have some, some cooperation going where you are flexible in the way how you, how intense you work on a certain topic, that can really be a way to rapidly throughout new approaches and see if it fits or if there is a market fit or not.
[00:16:48] Hywel: Yeah, you need do need to strike that balance between what you can do in house and what the third party vendors do.
And the smaller the team, also the, the more crucial kind of key staff turnover can be because the people who leave can have a huge amount of knowledge and it’s very difficult to replace them rapidly.
So, perfectly understand and you emphasize versatility and flexibility, and I think that’s very important for companies to consider if they’re trying to be commercially focused in, in today’s space industry that involves, of course, as you mentioned, third party suppliers and providers at different stages in the value chain.
How does then maintenance and operations factor in of such products, you know, with that value chain, with different people playing different roles?
[00:17:31] Michael: Bringing some prototype to operation to true 24- 7 operation to true, truly an automated business is not an easy task, right? I mean, many people think all you need to do is just let it run on the computer, right?
And it works fine, but that’s not true, right? If you want to have a true 24-7 offering, right, that is resilient, that customers can rely on. You need a different type of thinking, right? You need to, first of all, make sure that your, your system is running fine, right? So that there is enough redundancy and there’d be enough error recovery built in so that it can survive.
But at the same time, it also requires organizational rethinking, right, Or reorganization in a sense. When you’re operating 24-7, you also need to have on call personnel available, right? Because I mean, space is the space industry specifically. It’s not a local business, right? It’s not that you’re sitting in, in Germany or in France or in the United Kingdom and you launch a satellite, you’re launching a constellation and your customer is sitting next door to you, right?
It’s sitting in the same time zone. In many cases, that is not the case, right? Because the space data, it’s global, right? So you could. Some customers sitting in, in Canada or in, in, in South America, in, in India, all the way in China or Japan. Your data will be applicable to them, but they’re sitting in a completely different time zone.
So if they have any urgent request or if the system goes down for whatever reason, uh, you need to be able to respond to this. For this, you need to have no-call team that can, can react to this. And you need to have mitigation procedures in place. What happens if you are, if the data link is down, or what happens if a specific bit of your data processing pipeline fields, how can you recover from it?
You need, you need to have some strategy in place on what you can do to recover quickly to re bring the service back up, and then have, have enough time to investigate and, and fix it in the longer. And now we have had multiple customers on this side, right? We built products for them, we operated products for them.
And as I said, it’s, it’s a difference in, in how you handle a scientific product to an operational system. And it’s essential there to make sure that you, you understand your system. Uh, it’s also essentially how do you understand the external inferences that impact your system? And if you rely on external partners, you need to make all to make sure on their capabilities.
Not just go, I don’t know, look up in some directory and look for the, the most inexpensive, the cheapest call center you can find to sort out production for you. Because ultimately your, your customers, if they rely on you, they trust you and they’re willing to pay a premium price if you can offer a premium product.
[00:20:21] Hywel: Absolutely. That’s, that’s great. Thank you. So we focused, you know, almost exclusively on the use of AI and advanced process in satellite data generation and because that’s where the vast majority of the value generation is in the industry when we look at these sorts of missions.
But how else can such solutions be utilized in space, mission development, or manufacturing, those sorts of areas?
[00:20:47] Michael: Right. I mean, manufacturing is an entirely different business, right? I mean, most of what we discussed, they just said it was more on the downstream side, like on the payload, in the ground segment, on, on how you can use data to generate value to your customers. But if you look at it from a pure, uh, technical point of view, if you’re manufacturing components, you also have machinery.
And the machineries, it’s, it’s not just that you turn them on and to produce something and you’re taking the output and you’re happy with that, but the machines, they, they’re having more and more sensors, uh, being placed in. Why? Because you want to better understand how the production is is going, whether there is any congestion or where the bottlenecks are, how the production chain is generating the products.
And because you know, when you generate any instrument or any component or any, any piece of machinery per se. What do clients they’re interested in is really to understand when they can get it right. When is the time by which you have completed manufacturing? And also they’re interested in the quality, right?
How good is the instrument? Is there any, any defects in it? Uh, so for, for your clients, this will be essential. In understanding how they can work with it. And now these sensors, these machines have more and more sensors, right? And understanding the data out of that sensors can be essential in understanding what these machines are actually working, how they’re doing right now at this very moment.
And the way how you can analyze that data, it, it’s going to be massive amounts of data. If you have a continuous sensor running on a, on a 24-7 production line, because you might have sensors that measure the, the throughput. You have sensors that measure the, the certain strength of certain activities. So there’s an abundance of data coming together and evaluating that data.
It can be obviously tempting to just look at it once and then throw it. But if you want to have some sort of history behind it, because many machines, they don’t stock from one day to another, or they don’t make any faulty production from one minute to the other. It’s something that often degrades slowly over time.
You need to have a way, if you want to be proactive and, and you want to understand how your machine is doing right now, you also need to have that history, uh, in mind, right? So you can’t throw that all the data away. You might be able to aggregate it to some extent, but you shouldn’t be throwing it away.
And the way to really understand that data and to assess that data. This is really a place also where artificial intelligence can be of help, right? Because some of the problems that machines face when they have some production malfunction, you can anticipate, right? Because out of experience, you know that every, every, so many cycles, you know that a certain piece of equipment will likely have issues.
Or maybe there is even some information from you, from the supplier on, on this bit and pieces. But in many cases, problems specifically for machinery that is not standard off the shelf, but it, uh, is being custom-built. Many of the mistakes or the, the, the problems coming into life, they are not known upfront, right?
So you need to be observant. You need to watch the, the various metrics that the sensors generate and try to detect anomalies, right? Try to find what could be an issue that’s causing that hasn’t been there before. That’s maybe the reason for why other metrics also turning sour, right? That is really a place where also artificial intelligence can help you to better organize the data, better detect patterns in the data, and ultimately, um, help you to have a better understanding of what the situation is and also give you predictions on, on how things can go forward.
So, although it’s an entirely different industry, I would say, right, many of the strategies that you can employ to generate value out of data for earth observation, for example, um, some of those strategies you can also apply to, uh, improving the production in, in a manufacturing business.
[00:24:56] Hywel: Oh, fantastic. Thank you. That’s great insights. Yeah. And you could clearly see the impact of the different sectors you work in at Cloudflight like this.
Which yeah leads me to my final question. What are you currently working on at Cloudflight? And, um, I always ask conversion of this to, uh, everyone of our guests. What are you most excited about in the next few years?
[00:25:17] Michael: So, at Cloudflight I mean, we work with so many different things, right? That there is so many industries, specifically in the space industry. What we see is artificial intelligence is getting more and more present, right? So it’s, it’s not just used. Very small or isolated cases, it’s being used even in in cases where you wouldn’t think of it in the first hand.
Right. At the same time, also cloud resources are now standard, right? So doing stuff on premise. Is something that in the space industry we barely stumble upon, right? Many of the, the small companies, the startups, they have a limited budget and spending budget on building your own data center, that is typically a no go.
We see that there is, uh, a point when they reach a certain size, right? And maybe the constellation is five or six satellites. You start thinking twice about, um, using standard cloud resources because of the price issue that comes specifically with storing data. But in most cases, cloud, there is the way to go and, and also for the classic industry that was probably more conservative in, in moving things to the cloud, that they’re seeing the change there and they’re trying now to also be more agile in terms of their own internal IT resources.
So cloud computing, it has arrived and I think it’s here to stay for the next years. At the same time when it comes specifically to like Internet of things and, and assessing sensors, edge computing is something that’s more and more coming, or at least that where we see more and more projects, um, uh, coming ahead to us.
Cool. The, the clear quest there is how can you reduce the big things on the ground? How can you make smarter decisions? How can you make smarter assessments directly on the sensor basis, uh, just to be more, more dynamic, more agile in certain cases. And there’s going to be a little bit of conflict there.
Because you know, the data that you discarded early on is a data you can never recover. So it, it needs sensible choices, how to go forward there. But in many cases where privacy, right, where, where data privacy, where there are really critical data about persons or individuals in play, then these edge computing can be essential to even enabling any business model.
Blockchain used to be a hot topic a few years ago, then it has faded a little bit. Uh, we see there are small resurgence because trust, generating trust for your products is, is something that is still present and, and, and nothing else has been found that could, uh, compensate for this. So we see a slight resurgence in there.
And then there’s also a lot of other different technologies, also more exotic ones that we are working out. So it’s, it, there is a big abundance of different, uh, technologies that are being tried out in the industry now, and it’s a little bit difficult to see clear trends there, what I’ve already mentioned before.
But what I do see in, in, with many of the organizations that approach us, and also with many of our own internal folks here, is that the environment matters, right? And in the past you would only think about a, what’s the cost attached to it or can it be done or can it not be done? But we’ve really seen a change in the way how people think, right?
It’s, it’s really okay, I can do it now, but what does it mean to the environment? Right? Can this approach be sustainable, both in terms of the, the financial side and the, the resources it takes, but also on the environmental side, right? Because more and more people are seeing this as their main concern going forward, right?
And even now, in times of when we are discussing a lot about the energy shortages and, and how can we heat our homes over, over winter, um, you would think that this is completely off the table. Because, you know, you might say this is only like a, a first world problem, right? Like when you’re rich, you, you can care about the environment.
But it isn’t right. I think even nowadays people are very focused on this activity and try to keep it on the table. So really trying to improve the sustainability, uh, the environmental sustainability of solution is something where we see, at least at people working at Cloudflight, this is something they truly care for.
And we are glad that it’s the same for our customers and what I’m personally really excited about is I’d really like to see how the constellations go forward. I mean, there is, constellations is pretty much in every, uh, new space startup, a new space company I talk to, they all have their eyes set on a constellation, right?
I barely meet one that says, Hey, we only want to launch one satellite, and that one is going to the core of our business, pretty much anyone says we want to have a full constellation, and I’m not even talking about the mega constellations from SpaceX and, and Emerson Kiper, et cetera, but I’m talking about like smaller ones, like maybe a hundred satellites, maybe 50 satellites only.
So I’m really excited how that works out. I’m excited from a business perspective. Even if you use very inexpensive satellites, it takes a lot of resources to launch them. Right? I’d really like to see how that runs out. I’d like to see it from environmental perspective. I mean, obviously the space in space literally is in a sense unlimited, but there’d will be capacity issues being seen.
The situational, a bareness in space, right? Seeing, trying to understand various space debris and, and how that impacts your constellation specifically, or the constellation of others is a really hot topic, right. There’s also the issue about the orbiting satellites, right? And, and the impact it has on the atmosphere.
There’s the issue about light pollutions for astronomers, so there is many environmental impacts it has when you launch a constellation. I’m pretty sure that there can be, a way can be found that works for everyone, right? And that also makes that thing happen. But it’s going to be an interesting way forward.
But ultimately, the one thing that I’m really most interested about, about all the constellations is really in the data they generate. Because if you look back 20 years ago, or even 30 years ago, the few you had on Google Earth, right? Or or, or Bing Earth or bing maps, whatever, it was a special thing back then, right?
You had maybe like one satellite crossing over your location. Two weeks or so, or every once a month maybe, whatever it is. And now with the constellations, you are seeing over the very place that you live at maybe 5, 10, 20 different views on it. And not just once per day, but maybe as much as 3, 4, 5, maybe even 10 times a day.
So the way how we can see our earth is going to be changing so dramatically. Now that data is, is not the same data as we had before, right? In the past, most of the data was owned by the government, right? So, or by the public agencies. So you could have accessed it at maybe even free or at least no cost. Now we have commercial data coming in.
So I’m really excited about the possibility this brings. And at the same time, we also have to see how well the data is being archived because the information behind it, it’s going to be incredible.
[00:32:18] Hywel: Absolutely, absolutely. And in more than the visible spectrum, of course.
[00:32:22] Michael: Absolutely visible, near infrared thermal radar, maybe only like the, the atmospheric sounding.
You also have not only the, the few on the ground, but you also have the position of ships, of aircrafts, of things on the ground. So it’s so many different layers of information. If you can combine them, I’m, I’m a hundred percent confident that there is new products in there to be explored and new business models to be found to be ultimately successful with that.
[00:32:52] Hywel: Excellent. That’s, um, really exciting, uh, vision and I think that’s a great place to wrap up. Thank you, Michael. I think, um, you’ve shared some really good insights today for our audience, the discussion on rapid prototyping of new products and space, uh, compromises the companies have to look, making missions and services more sustainable.
I think its very important goal in and of itself, as well as the benefits, the multiple benefits here it brings to, to companies, uh, the share some great advice on setting up a new product on new business, and yeah, bringing in your insights from other sectors and looking forward to, to what could happen in the future. So from all of us at, at satsearch, on behalf of the listeners of the Space Industry podcast, I’d like to say thank you very much for spending time with us today.
[00:33:31] Michael: Yeah, thank you. It was a blast here. Really awesome job you guys are doing and leading the knowledge and connecting people. That’s really great. And yeah, I’m looking forward to the challenges ahead of every one of us.
[00:33:41] Hywel: Fantastic. Thank you Michael. And, uh, to all our listeners out there, thank you too for spending time with us today. We’re very pleased to have you remember, you’ll be able to find out more about all the, um, information that, uh, Michael has shared with us today, the companies, the innovations and resources satsearch.com and in the show notes for the podcast. And, um, just please stay tuned for our next, uh, episodes coming soon. Thank you very much.
Thank you for listening to this episode of the Space Industry by satsearch. I hope you enjoyed today’s story about one of the companies taking us into orbit. We’ll be back soon with more in-depth behind-the-scenes insights from private space businesses.
In the meantime, you can go to satsearch.com for more information on the space industry today, or find us on social media if you have any questions or comments. Stay up to date. Please subscribe to our weekly newsletter and you can also get each podcast on demand on iTunes, Spotify, the Google play store, or whichever podcast service you typically use.