SoftBank CEO questions Musk's orbital data center hype: too slow, too costly

SoftBank CEO Masayoshi Son poured cold water on Elon Musk's orbital data center concept during a recent shareholder meeting. Son argued that building data centers in space won't significantly reduce costs and will take far too long given the immediate urgency of the AI race. 'In the battle for AI, the next few years will be far more important than what might happen a decade or so from now,' he said. His remarks were dissected on TechCrunch's Equity podcast, where hosts noted the irony: SoftBank has a long history of placing wild bets, including the ill-fated WeWork investment. Kirsten Korosec observed that it's 'very ironic' for Son to play the skeptic, but also valuable to have a high-profile voice asking tough questions about space-based infrastructure.

The podcast discussion extended beyond orbital data centers to the broader compute-constrained landscape. Sean O'Kane highlighted how SpaceX is monetizing its satellite constellation by renting compute to smaller players, including a post-IPO deal—a strategy that effectively guarantees more business for SpaceX as satellites need frequent replacement. The episode also covered Groq's new $650 million funding and OpenAI's plans for custom chips. The consensus was that while orbital data centers pose interesting engineering and economic challenges, they are not a near-term solution to Earth's AI infrastructure bottlenecks. As O'Kane quipped, 'Neo-clouds are the new oil,' and every company with spare compute—from chip startups to shoe companies—is pivoting to lease it out, but durability remains an open question.