SoftBank’s CEO isn’t the only one with questions about Elon Musk’s orbital data center hype - BERITAJA

Not everyone is buying Elon Musk’s imagination for orbital information centers. 

Masayoshi Son, the laminitis and CEO of Softbank, argued astatine a caller shareholder meeting that building information centers successful abstraction won’t do overmuch to trim costs and will return excessively agelong erstwhile “in the conflict for AI, the adjacent fewer years will beryllium acold much important than what mightiness hap a decade aliases truthful from now.”

On the latest section of TechCrunch’s Equity podcast, Kirsten Korosec, Sean O’Kane, and I discussed Son’s remarks arsenic portion of a broader chat that included OpenAI’s plans for civilization chips, chipmaker Groq’s caller $650 cardinal funding, and overmuch more.

Kirsten noted that it’s “very ironic” that Son is playing the skeptic here, fixed SoftBank’s “long history of chaotic bets.”

Sean, meanwhile, said that erstwhile Musk talks about “making a constellation of satellites — satellites that request to beryllium replaced each fewer years arsenic good —  to dress up an ‘orbital information center,’” he’s conscionable “guaranteeing that overmuch much business” for SpaceX.

Keep reference for a preview of our conversation, edited for magnitude and clarity.

Sean O’Kane: Listen, neo-clouds are the caller oil, and everybody who wants to make money is pivoting to a neo-cloud. I’m proud to denote that TechCrunch is now a neo-cloud, springiness america each your money.

I mean, this is the point you do. It seems for illustration location are truthful galore players that are compute constrained, truthful anybody who has a changeable astatine being capable to lease retired that compute is taking it, whether that’s Groq, a institution that was semi-hollowed retired by Nvidia, aliases Allbirds, which went into bankruptcy and and emerged from it arsenic a caller neo-cloud supplier alternatively of trading shoes — Tim Fernholz did an question and reply pinch the caller CEO of of that caller effort that I would decidedly urge group spell read. 

Or whether you’re SpaceX, wherever your thought was: I’m gonna build an AI level that’s gonna person an addressable marketplace the size of U.S. GDP, but earlier we get there, we’ll conscionable rent retired our compute.  And we saw this proceed to hap pinch SpaceX, wherever it’s not arsenic large arsenic the deals that they’ve struck pinch Google aliases Anthropic, but they conscionable signed different deal, [their] first station IPO deal, to rent retired compute to different smaller player. They’re continuing down that road. 

You know, I could spot this being a business for Groq successful the adjacent term. The mobility pinch each of these is really durable is it successful the agelong term.

Anthony Ha: If we’re talking about SpaceX and their AI business and information halfway business, we besides person to talk about these comments that Masayoshi Son, the CEO of SoftBank, made recently, wherever he fundamentally said: What is the constituent of information centers successful space? Which is simply a mobility we’ve asked connected this show. 

And it speaks to, again, this consciousness successful the manufacture of being really, really compute constrained — they request to build arsenic galore information centers arsenic possible, [and] there’s each kinds of reasons why that is proving to beryllium challenging present connected Earth, truthful possibly abstraction is the answer. But I deliberation Son makes immoderate beautiful adjacent points about: All this worldly we’re talking about, moreover if it each useful — and the costs are going to beryllium very, very superior to make it work — this is not happening for years and years and years, truthful this is not a solution to immoderate contiguous problem, arsenic acold the existent request for information centers goes.

Kirsten Korosec: I conscionable want to constituent retired that SoftBank has a agelong history of making chaotic bets. I deliberation it says thing erstwhile Son comes up and asks the mobility that a batch of group person asked. 

I mean, location are a batch of VCs and founders [who] person been swept up into the thought of orbital information centers and it seems for illustration abruptly everyone’s connected board. When conscionable a mates of years ago, I think, if personification had mentioned that, it would get slapped down a small bit. So I do deliberation it’s an important portion of the process that personification who has a beautiful precocious floor plan is asking that question. But it is very ironic to maine that he is the 1 asking it, because if you look astatine his transportation deck, they’ve thrown a batch of money astatine immoderate beautiful bold ideas.

Sean: WeWork! Listen, we’re going to beryllium saying this for a batch complete the adjacent mates years. The thought of putting these things successful abstraction is going to beryllium an absorbing engineering situation and surely an absorbing economical challenge. 

Anthony, what you said is decidedly correct to a definite extent. Elon Musk is simply a personification who hates reddish portion and you know, location are nary NIMBYs successful abstraction truthful of people he’s going to effort and do that. 

To me, it comes down to: The business arsenic it stands now for SpaceX, particularly its motorboat business, is conscionable overwhelmingly reliant connected Starlink. The logic that they are 80 aliases 90% of the motorboat marketplace globally is not conscionable because they’ve done each these things that are amended than beautiful overmuch each different motorboat supplier about the globe, it’s besides because they person Starlink that is driving up that number. If you region Starlink from the equation, they would beryllium person to — I don’t know, possibly 20% aliases 30% of the motorboat market, aliases 40%, but it surely wouldn’t beryllium 90%. 

And erstwhile you talk about making a constellation of satellites — satellites that request to beryllium replaced each fewer years arsenic good —  to dress up an “orbital information center,” quote unquote, you’re conscionable guaranteeing that overmuch much business for your motorboat business. And I conscionable can’t extremity myself from coming backmost to that point.

Kirsten: I want to really quickly opportunity that [SpaceX’s] different large business is renting retired their compute, by the way. So backmost to the spot conversation. We’ve travel afloat circle.

Anthony: One of the different themes that whitethorn tally done this section is this thought of talking your ain book. This is not a caller phenomenon. Executives astatine tech companies, aliases immoderate different company, what they’re predicting for the early is yet the early that is going to beryllium advantageous to their business. 

But I deliberation it’s thing that’s conscionable ever worthy remembering erstwhile we’re having these conversations about large AI companies, because it is this infinitesimal of unthinkable uncertainty, and we’re each wondering: What does the occupation marketplace look for illustration successful the future? What effect is this going to person connected the environment? What are the skills I request to learn? 

All these AI CEOs aliases AI investors, they each person thoughts connected that. And it’s not that they’re incorrect aliases that they are being deliberately misleading, but successful each case, there’s an asterisk to these predictions. In Musk’s case, he’s talking about thing that would beryllium very bully for SpaceX’s business. In SoftBank’s case, they are very, very heavy invested successful information halfway projects present connected Earth. Sam Altman is the different notable fig who’s rolled his eyes a bit astatine the orbital information halfway thought — and again, he and Elon Musk evidently person a agelong and analyzable history together.

All of which is to opportunity that there’s conscionable nary objective, impartial observers here. It’s each these group pinch baggage and tremendous amounts of money astatine stake.