How AI is changing what satellites can do in orbit

Hinzugefügt: 2026-06-04

[00:00:05]As more satellites move into orbit, the focus is shifting from getting hardware into space to what those systems can do once they're up there.

[00:00:14]Companies are using artificial intelligence and real-time analytics to process data directly from orbit. This includes tracking wildfires, monitoring changing conditions on Earth, and improving how information moves around the world. There's also interest in what comes next. Some are exploring the possibility that future computing and data processing could happen in space itself. The capabilities being developed today are expanding what's possible with space-based technology and opening new opportunities for orbital operations.

[00:00:48]We were just talking about the predictive piece of this and what collecting all that data in space is going to mean for the future and for prediction and right um and honestly for making Earth, you know, better what that how muon plays within that concept.

[00:01:01]>> So I just want to piggyback on what Dylan was talking about with the uh data centers. I think we're already seeing the kind of inference stage where we're seeing our systems that are up there today doing the uh AI analytics. We're also the only satellite company to so far announced a partnership with SpaceX for the Starlink laser terminal so that we're linking directly into Starlink. So it can be like broadband in space and these spacecraft become just another data node like it would be in a data center. And so I think for companies you end up compressing where that node is in space. uh and that allows you to do the kind of uh things that make it much more operational and it's not just like you're saying sort of reporting data down and so for example we have firesat which is a global uh wildfire detection and tracking system which we developed in concert with Google research and with uh earthfire alliance which is a not for profofit and that will uh enable a fire as small as 5 m x 5 m to be detected and tracked uh perimeter intensity of fire and path of fire for public safety fire operators globally. And we're aiming to get to one hour global revisit within 24 months and eventually to 15 minutes. And so the kind of always on connectivity that you get with intra satellite links and lasers and the kind of processing allows us to move all of that uh decision-m up into orbit where you've got these always on kind of systems that aren't going to be sort of slowed down by wind conditions or cloud cover or things like that.

[00:02:42]>> I mean that's really impressive when you're talking about 5 meters by 5 meters especially when you're talking about something like wildfires and prevention and detection. um the numbers and I've had conversations with folks for example um you know within government in the US right like last five years alone something like a trillion dollars worth of economic impact if you start to layer in the knock-on effects I mean it's it's a huge issue >> uh that governments can't seem to get their arms around especially when we do start to talk about climate change so just to put it in perspective in terms of five meters versus five meters using data that you're pulling from face. It's It's kind of amazing.

[00:03:20]>> We think it's it'll have a big impact.

[00:03:22]And to your point about the trillions of dollars of infra infrastructure that's threatened globally and increasingly so, it's sometimes difficult if the budget doesn't come from a single entity to get that put into a particular new capability or technology. And >> yeah, >> that's where I think the public private partnerships of uh private companies and eventually some government entities can come and then benefit from these new technologies. Okay, before I jump further into public private partnerships, >> yeah.

[00:03:50]>> Um, so then going back to I'm I'm jumping around here, but going back to data centers from space, given the fact you are already working with Starlink on some of this capability, couple of years, is that your timeline as well listening to Dylan and thinking about what Elon's been saying? I I'm sure Elon will surprise us all but we believe that there are you know a number of engineering elements to be tackled in terms of you know radiation thermal dissipation and then really also getting the economics right. So yes it needs starship level and I think that we need to just build out the capabilities again first on inference and then on training.

[00:04:26]I think it's uh much like a a lot of the panelists have talked about today, the adoption of a lot of the AI capabilities requires humans and organizations to uh change the way that they operate. And we see the same thing. We're just trying to build the tracks and the infrastructure so that when it's going to be cheaper and more cost-effective to do uh training in space, you can do that and it's irrelevant to the particular hyperscaler or company that's uh deploying their capabilities. So public private partnerships.

[00:04:59]>> Yeah.

[00:05:00]>> Just the fact that and we see this in the US uh with government contracting shifting. NASA was really the front lines of it. And I think about what you're doing with Star Lab and just the fact that you're even they sort of changed um the process for low Earth orbit to enable a commercial space station like that to be built and developed. Um and now we're starting to see some of that contracting come in with the lunar economy, this nent lunar economy, too.

[00:05:24]But you're also starting to see it happen on the defense side as well. So what does that mean for Voyager and how does it change the economics for companies that are operating in space?

[00:05:34]>> Yeah. So you mentioned on the defense side, for example, Golden Dome, which is the missile defense shield the US is contemplating for North America. Um you're right, a lot of this has to do with creating incent uh incentive alignment between what the government's trying to do and what the private sector needs to do. So oftent times the c uh the government will be the initial customer. They'll set the requirements.

[00:05:57]They'll provide the initial funding but then they'll anticipate that the market will grow beyond just the government. Um so yeah I think it's a good model. I think um you know the days of cost plus contracting I think uh are thankfully dissipating because I don't think we got great outcomes with that.

[00:06:16]>> Uh we didn't get great innovation. uh frankly we didn't have an large enough industry to attract the capital that we needed. So I think these new procurement models are helpful. Uh we also now in the US have the space force which is really acting I think as a uh bridging the gap between the department of defense and NASA civil space and military space uh and leaning very heavily into innovation. Uh and it's not just the US. I mean Europe is looking at different procurement models as well.

[00:06:46]And I think um you know right now I think there's something like 75 space agencies around the world. So I think these models are helpful uh in countries that are looking for a way forward. Uh but I think in my view public private partnerships are the best of all worlds.

[00:07:03]I think everyone wins. The government as a buyer gets better outcomes, better value for their money. Uh, I think the industry uh creates better alignment and I think the capital markets like it as well.

[00:07:14]>> I was talking to Peter Beck from Rocket Lab about this last week, somebody I'm sure you both know. Um, and one of the points that he made was that what this new commercial space era is enabling is countries that perhaps didn't have their own homegrown in-house space capabilities to now outsource it to to companies. Uh, and your thoughts on that? Yes, I I I think we're seeing the same thing and it's really starts with what Dylan was talking about with the US government transitioning to really relying more on commercial space because the innovation flywheel that they can tap into uh allows us to bring greater capabilities to bear and as you know Peter Beck was talking about every country or group of countries needs some amount of sovereign capability and so if you can tap into the innovation you get from a range of private companies in the US and and elsewhere, you get the the very best of what can be developed. It's not this traditional cost plus environment which really didn't deliver.

[00:08:18]And when you're moving from billion dollar school busiz satellites that take a decade to build to this proliferated low earth orbit, it's a cornucopia of options. We were just a space symposium as we've all been to over many years in in Colorado Springs and uh the the the national government and military attaches from allied countries all around the world are there because they want to tap into what commercial space is able to do for them. One other thing, Morgan, that I think we can anticipate happening, which I think might be quite interesting to follow is of these smaller space agencies creating more of a marketplace around that demand and aggregating that demand. So, I would anticipate either new companies being formed or maybe consortiums being formed around this. Uh because I think um if you look at the need for that in the market, I think I think it's highly likely that'll come about. Hm. With that, let's open it up to questions from the audience and see if we have any.

[00:09:18]We've got a gentleman in the front row here.

[00:09:22]>> I'm Claus Nzo with EMTTD Lab. U my question is with people like Alon Mus putting up this idea of orbital data centers. You mentioned space radiation that can quickly degrade or destroy all the computing equipment even before data centers. What are your views on the trends and particular innovation in that space? Yesterday we had some biotech speakers here and they talked about using AI simulation models to discover new drugs. You know, do you see similar things happening to mitigate the the radiation challenge?

[00:09:50]>> We're definitely using a whole range of physics models that employ AI to design whole constellations before we ever cut metal. And so I think it's used both in the design phase and then I think to the point you're maybe alluding to in orbit.

[00:10:05]You're when you've got systems that are looking to make decisions as they're coming uh up on new information and tip and cue other uh systems, you need to have that those being able to operate independently. And so we're seeing that already across a number of customers.

[00:10:22]>> Yeah. And just on innovation, microgravity research. So a lot of research happens on the international space station today in biioarma in things like perfect crystalline growth and and the like and the reason more research doesn't happen on the international space station is it's really supply constrained. So when the commercial space stations come online I would anticipate really a flourishing around these research areas. uh you literally can grow a perfect crystal uh in space and if you bring those seeds uh back down here to earth you can grow a larger perfect crystal around it. Uh you can print organ tissue on orbit. Uh we've actually we meaning humanity have printed a human meniscus for your knee.

[00:11:06]So you can imagine a future where your stem cell uh you know stem cells prints you a new kidney in low earth orbit and comes back down and uh there's no uh risk of organ rejection because it's actually your exact uh match. So there's a lot of really science fiction type stuff that's enabled by the fact that we now have uh heavy lift rocket capability. Uh our space station for example will be single launched to orbit. So imagine a space station you construct, you test, you configure on the ground and in one launch, it's in orbit and functional. Uh, you know, that's the future we're going to live in.

[00:11:46]>> Do we have other questions?

[00:11:48]>> Thank you very much. Uh, Alisa from Amcham Singapore. My question is, I'm going to preface and say it's a little bit of a challenging question and perhaps not a current reality question.

[00:11:59]Um, what is your best prediction for when humans will make it to Mars or if humans when and if >> so humans just to clarify humans landing on Mars or orbiting Mars because it's two different Okay. I I would say um in the 2030s it'll happen. I think it's probably late 2030s. Uh it's a very difficult problem to solve. So just to give people an idea, the International Space Station is 400 km away from the surface of the Earth.

[00:12:36]The moon is 400,000 km and Mars at its closest approach is 60 million km away.

[00:12:44]So it's a much much much more difficult technical problem.

[00:12:47]>> I would just say that you know we're focused on low earth orbit and cis lunar which is around the moon. So we're focused on much more practical uh applications just now.

[00:12:58]>> So just to wrap this up, what is it maybe on your term goal, what do you think it takes? Do you think we actually come into the end of this decade and we see a permanent presence of humans on the moon?

[00:13:11]>> Oh yeah, absolutely.

[00:13:12]>> It's closer than most people think. I think >> I think we'll have humans on the moon by the end of the 2020s and we'll have some lunar base. It'll probably be an inflatable habitat with some life support. Um, but I think deeper into the 2030s, 2032, 2033, you'll be able to sit on your porch in upstate New York and look at the moon and there'll be lights on the moon because there'll be people living and working on the moon.

[00:13:39]>> Wow.

[00:13:41]>> And I think we will have satellite infrastructure that is no more remarkable to us than as we think of data centers today. It's really going to be distributed uh broadly.

[00:13:52]>> All right, on that note, Dylan Taylor and Gregory Smeirn, thank you both for joining me on stage. Thank you very much.