Masayoshi Son has publicly challenged Elon Musk’s push for orbital data centers, arguing they won’t cut costs and are too slow for the immediate AI race. In a wide-ranging discussion, the skepticism was met with a sharper question of business incentives: Musk’
Masayoshi Son didn’t just disagree with Elon Musk’s latest big idea—he questioned whether the idea is even worth the hype.
At a recent SoftBank shareholder meeting. the company’s founder and CEO argued that building data centers in space won’t do much to reduce costs and will take too long. His timing critique was blunt: in the battle for AI. “the next few years will be far more important than what might happen a decade or so from now.”.
That challenge landed in an environment already crowded with certainty. On TechCrunch’s Equity podcast, Kirsten Korosec and Sean O’Kane debated Son’s remarks alongside other tech industry flashpoints, including OpenAI’s plans for custom chips and chipmaker Groq’s new $650 million funding.
Korosec zeroed in on the contrast between Son’s skepticism and SoftBank’s reputation for taking chances. She called it “very ironic” that Son is playing the skeptic. citing SoftBank’s “long history of wild bets.” Her bigger point was that orbital data centers—once a notion that would likely have been dismissed—have started to draw wider enthusiasm. even from people who. just a couple of years earlier. might not have spoken so confidently.
O’Kane, meanwhile, framed the Musk side of the equation through the business mechanics it would create. When Musk talks about “making a constellation of satellites”—with satellites that need to be replaced every few years—O’Kane said Musk is effectively “guaranteeing that much more business” for SpaceX.
The episode returned to SpaceX’s dependence on its launch business, and specifically on Starlink. O’Kane argued that SpaceX’s dominance in global launches—described as 80% to 90%—is not only tied to better launch performance. but also to Starlink’s role in driving demand. Without Starlink, he said SpaceX would be much closer to 20% to 30% of the launch market, or possibly 40%, rather than 90%.
For O’Kane. that’s exactly why the orbital data center concept doesn’t just sound like an engineering moonshot; it also sounds like a lever for near-term growth. If a constellation is the foundation, and the satellites must be replaced periodically, then the launches don’t stop. “You’re just guaranteeing that much more business for your launch business,” he said.
The thread of compute also wound back in. Kirsten Korosec reminded listeners that SpaceX’s other major business is renting out compute. “We’ve come full circle,” she said, connecting Son’s critique of timelines and costs to the broader industry chase for usable computing power.
Anthony Ha added a different kind of caution to the discussion—less about who benefits. more about whether the timeline matches the problem. He said the point Son makes lands on a question the show had already raised: “What is the point of data centers in space?” Even if the idea works. Ha argued. the costs would be very serious. and the project still wouldn’t arrive for years and years. That would make it hard to treat orbital data centers as a solution to today’s data center needs.
What stayed consistent throughout the conversation was the recognition that the debate isn’t being run from a neutral distance. Ha pointed to the way big predictions in tech often line up with the future that helps the companies making them. He described the pattern as “talking your own book. ” and said it’s not necessarily deliberate—just naturally shaped by what each player has to win.
He also highlighted the uncomfortable mismatch of roles: Son’s SoftBank is heavily invested in data center projects here on Earth. while Musk’s vision overlaps with a business that benefits from satellite deployments. Ha also referenced Sam Altman as another notable figure who has “rolled his eyes” at the orbital data center idea. noting Altman and Musk have a long and complicated history.
By the end, the central tension wasn’t just whether orbital data centers will work. It was whether anyone in the room is truly removed from the incentives—when money, timelines, and markets are all on the line.
SoftBank Masayoshi Son Elon Musk orbital data centers SpaceX Starlink satellites AI compute Groq OpenAI custom chips podcast