Transcript:
Caroline WoodsThe AI revolution is about to consume the equivalent of the entire power grid of New York City, and the stock market is looking for the next infrastructure play. Forget chips. We’re talking about the power for the chips. Baiju Bhatt is co-founder of Robinhood and the founder of Aetherflux He joins us now. Baiju, thanks so much for being here.
Baiju BhattThanks for having me.
Caroline WoodsSo first you disrupted Wall Street with Robinhood and now you’re disrupting the sun. Explain what you’re doing with Arthur Flax in plain English for us. What are you actually trying to pull off here?
Baiju BhattYeah. So started the company, a little over a year and a half plus two years ago. At this point, now, with the goal of building an energy grid and space. And the idea here is that the power of the sun is limitless. And we if we had better ways of harnessing it, we could use it to power more of the resources on Earth.
And this idea of space solar power is actually a pretty old one from NASA. The Department of Energy in the 70s to collect solar power in orbit. And the benefits of this are that you get to get access to the sunlight nearly continuously. Number one. And number two, the actual density of sunlight that you get in space, averages out to be considerably higher than on the ground.
And so we kind of have this untapped natural resource of energy and space. How can we harness that to make it useful for humanity? And so we started out with building power beaming that actually captures energy in space. And that beams it down to locations where it’s either contested environments for the Department of War, or whether there’s no existing energy infrastructure from the ground.
There’s actually hardware that’s going to space, and over my shoulder that’s being worked on right now, that’s going to demonstrate that, ideally, starting the summer. And additionally, we announced, a month and a half to two months ago, some of that time frame that we’re building actual infrastructure in space that lets us do that, lets us serve one of the biggest users of energy today, which obviously you alluded to, which is artificial intelligence.
By actually taking the chipsets that require the power and putting them proverbially next to the sunlight. So building AI powered, sorry, solar powered AI satellites in orbit.
Caroline WoodsSo is that the galactic brain then?
Baiju BhattYes it is. So that’s our name for this initiative. And the idea is, is that starting, as early as next year, we’ll have GPUs in orbit that will be able to be powered by solar power from the sun, and we’ll be able to do real inference jobs, which we think is one of the first use cases for this.
And so just a little more, a little more clarity on what that is. So once you’ve got these artificial intelligence models trained and built the actual hardware that you run them on to interact with, I’ve been using the, it’s one of the places where the majority of the energy use is projected to go in years to come.
And that actually is, a use case that we think is well suited for the space environment.
Caroline WoodsBut you tell me a bit more about the why, though. Are we just running out of power here on Earth?
Baiju BhattYeah, I think actually one of the more acute problems is how long it takes to bring that power online. So when you look at bringing online new grid scale installations for, AI data centers, the time between when you want to get started and when those are operational, some of those timelines are pushing out 5 to 8 years, where this presents an alternative that lets you go sort of more directly to bringing things online.
So whereas the terrestrial route is you build the building, you dig the dirt and you connect it to the power grid. And potentially you also may need to build a dedicated power plant for that data center. Then you have to get the water the other resources in to operate it. The alternative here is that once this technology is up and running, once we get the chipsets, all we need to do is integrate them into satellites.
And as soon as they’re launched into orbit, they can begin operating pretty much right away. So it presents an alternative to bringing online more chipsets and more power that’s, you know, largely disconnected from terrestrial limitations.
Caroline WoodsSo what is your timeline? How soon until these will be in orbit and the galactic brains will be will be at work?
Baiju BhattYeah. So we’re we’re hard at work here on making, this stuff come to life. We’re targeting, first half of next year for first, demonstrations of this technology. And what that’s going to look like is not just a single GPU that’s operating in orbit, but an interconnected sort of, cluster of these things. So think like roughly ten GPUs somewhere in that range where together they’ll be able to do inference workloads for AI.
So these are like some of like the bread and butter use cases for, power going into artificial intelligence.
Caroline WoodsSo what’s the goal here? Are you looking to be like the Nvidia of energy and really power as much as possible, or just an alternative.
Baiju BhattTo actually not quite. Our vision for the company is to build an energy company in space. And the the vision here is, is broad in that regard, in that we think that the future of humanity is commercialization of space. And as I look at the commercial opportunities in space and what we need to do to make space a real commercial environment, we think that energy and energy infrastructure is one of the building blocks of building that future economy.
And we view this work that we’re doing both power beaming and with data centers as actually building an energy grid in space. And so artificial intelligence is this massive consumer of energy. And I think is going to be one of the things that, sort of is one of the the landmark use cases for energy and space.
But over time, we actually want to view ourselves much more as an energy company.
Caroline WoodsI’m curious for investors watching this, how does this space based energy change the valuation of traditional utility stocks?
Baiju BhattYou know, I think this is still early technology, and I think that’s still going to play out in the years to come. But this is an emerging technology that I think has a degree of an inevitability to it. So I think as we go down the path that we are and there are other companies that are doing similar things, namely space X, I think it’s going to start to become clear that this is one of the next major commercial use cases for space.
And I think on the whole, the demand for artificial intelligence compute doesn’t have any real signs of slowing down. So I think it’s actually going to sort of grow the, the total number of ways of serving this basic need.
Caroline WoodsIn theory, on paper space sounds great. Obviously there’s free land, there’s the nonstop sun. But I can’t imagine that launching things into orbit is cheap. So is cost a barrier for you? And then what happens when you have to make any sort of, you know, technical changes? You can’t really send a repair person up into space, right?
Baiju BhattYeah. So two parts to that. The first one is, is that bringing the cost down and making this cost competitive with terrestrial options? That’s the name of the game right now for us. Like finding the paths to how do we down cost this technology so that we can be competitive with terrestrial options. Is the primary focus. And I think that there’s been lots of examples of or rather 1 or 2 significant examples of this happening.
And specifically, I’m pointing to both, reduction in launch cost and that enabling things like Starlink and internet communications and space. So the question is, is how do we get the cost of what we’re doing down, really substantially. And what that looks like is looking at each component and saying, why is this cost so much in the past?
Is there a path to doing it much, much cheaper? And we think that there there is for many of the core component technologies. And so the other part of this also is, is that reducing costs in a well-established industry is what I spent the last, you know, decade, a decade and a half of my life doing via Robinhood, right.
Which was a dramatic sort of cost structure change for access to capital markets and investing in the stock market. And while these are not the same thing, there’s some degree of it doesn’t always have to cost. What its cost in the past is the mindset that of bringing to do this.
Caroline WoodsI’m curious what farming power from space means for the everyday person, because there are a lot of people now who are looking at their electric bills saying, why the heck is this so high? And I’m subsidizing data centers. Like what? Even if we are using AI, what could this mean for mainstream assets?
Baiju BhattYeah, I mean, I think the the broad vision here is that the demand for energy for AI, is going to be a real problem for people living on Earth, right? Like if you have to build new data centers and for example, in Virginia, where I know it’s a significant chunk of the energy use already that’s plugging into the same energy grid that people that live in that state are.
Right. And when you see energy costs and electrical bills going up like that’s for for a country that, you know, has sort of high cost of living and sort of people’s day to day budgets being strained, like that’s a no go. And so the way that this has the potential to really be trivialized for people on Earth is that it provides an off ramp to some of these, like very, very energy intensive use cases that would otherwise be being built in their backyards.
And so a couple of years before that really materializes, but already we’re seeing the timelines that people are planning. AI infrastructure is years and years into the future.
Caroline WoodsI’m curious because you left a big public company to do this. What did you see in Energy Day that that made you say, I have to leave Robinhood and build these space lasers and galactic brains instead?
Baiju BhattIt’s actually been kind of a passion of mine for a really long time. So I grew up around the space industry and studied physics and math at Stanford, and I kind of had this idea in the back of my head where if I ever got a chance to, I wanted to be a part of the space race and sort of the next generation of it.
So the the thought process was, was definitely one that was leading me towards space. I think in many ways from, from childhood. But look, I mean, we’re on the doorstep of, of a major change in the way that humanity interacts with space with these heavy launch vehicles and the mass to orbit constraint really fundamentally being lifted as these bigger vehicles come online.
And, you know, we have the potential right now, as a country to be building what are going to be new industries in space that last 50 years, possibly longer. And, you know, I’ve wanted to be a part of that since I was a kid. And here I am.
Caroline WoodsAnd you’re not alone there. There’s a lot of interest in space. We see it with you mentioned SpaceX, obviously, Elon Musk, Jeff Bezos. Are you worried that it’s getting crowded? And do you see it as you competing against those other companies are working together?
Baiju BhattI think much more working together. And I think that a world in which I guess, a galaxy in which space is being crowded because there’s so much commercial activity would be an awesome one. I think we’re still pretty far from that as we look at and we look at even the launch availability of space. You know, there’s a huge demand for this.
And, it’s it’s only growing. So I think as more people are building SpaceX infrastructure, more people are building ways to get to space. These are all good things. And I think, a world in which this is commercial in nature. I’ll add also, we think is really beneficial to the United States. You know, it’s national interest because the flywheel of capitalism is the thing that accelerates US innovation past the rest of the world, and has done so for a very long time.
I think that the US is uniquely positioned to drive the commercialization of space, and in turn, lead as it has from an innovation perspective. So I think more people doing this is strictly better than less. And I think we have a chance to really pioneer a commercially driven space race in the 2020s where in the 1960s, it was, you know, largely nation states.
Caroline WoodsYeah, I’m curious about that. Obviously, we talk a lot about trying to get the edge ahead of China when it comes to AI. Where are we in the space race when it comes to any sort of edge or lead?
Baiju BhattI think the US is dominant in the space race. And what one of the fundamental things that this effort, has the potential to do is it has the potential to translate the US’s dominant position in space to a position of dominance and energy, which we know that the Chinese have systematically made investments in energy infrastructure, you know, for the last two decades.
And in many ways, they’re they’re build of power. Their ability to bring you power online outpaces what we could do in the US. On the other hand, we do lead in the space race. So to the extent that we can translate that advantage in space to an advantage in energy is we think something that’s pretty critical to do.
And so it’s also exciting that there is a degree of competitiveness to this, because we think it actually just pushes things further along faster.
Caroline WoodsWhat’s the takeaway for investors watching right now? They might say, oh, it’s really cool, but obviously you’re not public yet. I will ask you if you have plans to go public, but what’s the takeaway for them as they think about finding the next big opportunity, especially as it relates to AI?
Baiju BhattYeah. So I think as it relates to I think the commercial activity there has been accelerating for years, and I think we’ll continue to do so. I think the exciting thing here, that we’re seeing, at least still first with startups and companies that are, you know, getting bigger and becoming more substantially sized, is that the commercial applications for space have the potential to be much more real, in the coming decade.
And it’s driven by the advancements that a lot of the component technology and a lot of the enabling technology, specifically with the ability to get more stuff into space, more mass, into space. It really enables the ability for more commercial activity to have at space. And that’s something that we want to be at the forefront of pioneering.
And I think there’s going to be more activity. I think it’s, the artificial intelligence component, because there is such a demand for it has the potential to accelerate the development of space hardware. And the confluence of those two things we think is pretty interesting.
Caroline WoodsIf I gave you $1,000 to invest in an AI or space publicly traded company today, where would you put it?
Baiju BhattI’m not sure if I’ve got a good answer to to which what I would, I would do that in, you know, it’s it’s a really quickly moving field. But I think like the fact that there is so much activity here is what makes it interesting. Right? And we look at the demand for the chips. We look at the demand for the energy.
We look at the consumer demand for compute. Right. Like, this is a wave that has the potential to be as big as the internet was. And that’s really exciting.
Caroline WoodsAny areas in this space that you think retail investors are overhyping right now?
Baiju BhattAreas that are overhyped. You know, I, I’m not sure that I’ve got one that’s specific for that either. I think, you know, kind of coming back to the the Robinhood perspective, right. The exciting part is that there is so much retail participation in capital markets today, and that that is fundamentally different than it was 10 to 15 years ago.
Obviously, that’s something I spent a lot of time the majority of my professional career before this working on. So I think that’s personally a bit of a point of pride. Right? It’s like there are more people that care about where the economy is going and that they have, a conduit to be a part of that. And Robinhood is a big part of that.
Caroline WoodsAnd knowing that retail traders and investors are so much more active, due in large part to you and Robinhood. Does that make you want to to have going public with Aetherflux? The main goal here is that on the horizon.
Baiju BhattThat would be that would be an awesome accomplishment. I think it’s still too early in our company’s timeline to to talk about that, but stay tuned. And it would be an exciting thing if it did happen some day.
Caroline WoodsOkay. And just finally, what happens if this doesn’t work out? If you can’t do it cost efficiently, what happens in terms of powering AI, which just seems like it’s inevitably just growing and growing and becoming a bigger part of so many of our lives at this point.
Baiju BhattAnd oh, we’re not giving up. We’re going to push on this thing and make it up. Yeah. Okay. Leave it there. Look, I think there’s to there’s going to be challenges along the way, but, you know, got to work through.
Caroline WoodsBaiju Bhatt, co-founder of Robinhood, the founder of Aetherflux Thank you so much.
Baiju BhattThank you.