Not everyone supports Elon Musk’s vision for orbital data centers.
SoftBank founder and CEO Masayoshi Son argued at a recent shareholder meeting that building data centers in space won’t do much to reduce costs and would take too long, arguing that “the next few years will be much more important in the AI fight than what might happen in 10 years or so.”
On the latest episode of TechCrunch’s Equity podcast, Kirsten Kolosek, Sean O’Kane, and I discussed Son’s remarks as part of a broader discussion that included OpenAI’s custom chip plans and chipmaker Groq’s new $650 million funding round.
Kirsten said it was “deeply ironic” that Son was playing the skeptic here, given SoftBank’s “long history of wild bets.”
Meanwhile, Schon said that when Musk talks about “creating constellations of satellites (satellites also need to be replaced every few years) to form ‘orbital data centers,'” he is just “guaranteeing that much more business” for SpaceX.
Continue reading for a preview of the conversation, edited for length and clarity.
Sean O’Kane: Listen, neocloud is the new oil and everyone who wants to make money is looking at neocloud. We’re proud to announce that TechCrunch is now Neocloud. Please donate the full amount.
So this is what you do. There seem to be so many compute-constrained players that anyone who has a chance to lease that compute is going to take advantage of it, whether it’s Groq, the company that got half-hollowed out by Nvidia, or Allbirds, which went out of business and rose up from there as a new neo-cloud provider instead of selling shoes. Tim Fernholz has an interview with the new CEO of the new initiative that I recommend you read.
Or even if you’re SpaceX, here’s your idea. We’re going to build an AI platform with an addressable market the size of the US GDP, but before we get there, we’re just going to rent the compute. And we saw this continue to happen with SpaceX. At SpaceX, we just signed another deal, our first post-IPO deal, to lend computing to another smaller player, although it’s not as large as the deals we signed with Google and Anthropic. They continue on that path.
You know, this is going to be Groq’s business in the short term. The question with all of this is how durable it will be in the long run.
Anthony Ha: If we’re talking about SpaceX and its AI business and data center business, we also have to talk about recent comments made by SoftBank CEO Masayoshi Son. He basically says: “What’s the point of a data center in space?” is the question we asked on this show.
And this, again, speaks to the feeling that this industry is really compute-constrained. There’s a need to build as many data centers as possible (and) there are many reasons why that’s proving difficult here on Earth, so maybe space is the answer. But I think Mr. Son makes some pretty fair points. All these things we’re talking about, even if everything worked out, and there would be very serious costs to make it work, this won’t happen for years and years and years, so this is not a solution to the pressing problem as far as current data center needs are concerned.
Kirsten Kolosek: I just want to point out that SoftBank has a long history of making wild bets. I think it’s saying something for Mr. Son to come out and ask the question that so many people have been asking.
So it seems like there are a lot of venture capitalists and founders who are crazy about the idea of orbital data centers, and all of a sudden they’re all on board. If someone had said something like that just a few years ago, I think they would have taken a bit of a beating. So I think it’s an important part of the process for someone who is fairly high profile to ask that question. But it’s very ironic to me that he’s asking that. Because if you look at his pitch deck, they’re putting a lot of money into some pretty bold ideas.
Sean: WeWork! Listen, we’re going to keep saying this over and over again for years to come. The idea of putting these things in space would be an interesting engineering challenge, and certainly an interesting economic challenge.
Anthony, what you said is certainly true to some extent. Elon Musk hates bureaucracy and there are no NIMBYs in the universe, so of course he would try to do it.
For me it looks like this: SpaceX’s current business, especially its launch business, is overwhelmingly dependent on Starlink. The reason they have 80-90% of the global launch market is not only because they do all these things better than almost every launch provider in the world, but also because they have Starlink that’s driving that number. If you take Starlink out of the equation, they’re closer to 20% or 30% or even 40% of the launch market. But it won’t be 90%.
And when you talk about building a constellation of satellites (similarly satellites that need to be replaced every few years) to make up an “orbital data center,” you’re only guaranteeing that much business to launch business. And I can’t stop going back to that point.
Kirsten: By the way, I want to say right off the bat that other big companies (SpaceX) are renting out their computing. Let’s get back to chips. I’ve come full circle.
Anthony: Another theme that might run through this episode is the idea of talking about your books. This is not a new phenomenon. Executives at technology and other companies are predicting a future that will ultimately benefit their businesses.
But I think this is something that’s always worth remembering when we have conversations about big AI companies. Because there’s incredible uncertainty right now, and everyone’s wondering, “What will the job market look like in the future?” What impact does this have on the environment? What skills do we need to learn?
All AI CEOs and AI investors are thinking about it. And while they are not wrong or intentionally misleading, in each case these predictions are marked with an asterisk. In Musk’s case, he’s talking about things that are very good for SpaceX’s business. In Softbank’s case, they are investing very heavily in data center projects here on Earth. Sam Altman is another notable person who’s rolling his eyes a bit at the idea of orbital data centers. And then again, he and Elon Musk clearly have a long and complicated history.
In other words, there is no objective, impartial observer here. They were all people with luggage and huge amounts of money.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
