Elon Musk said Monday that future superintelligence compute will be “overwhelmingly in space,” clarifying that by “overwhelmingly” he means “it will round up to 100%.” The posts came in reply to Sequoia Capital partner Shaun Maguire, who has been circulating a bottoms-up SpaceX financial model and said he will publish “a systems level analysis of orbital compute over the next week or two,” adding that “Almost all SI compute will be in space.” Musk answered Maguire’s post with a one-word reply: “Absolutely.” The exchange builds on Maguire’s argument that SpaceX is the low-cost producer for compute that starts on the ground and moves into orbit. Musk’s Monday posts go further than his Friday note that monthly orbital compute deployment will probably pass terrestrial capacity around the end of 2028. SpaceX has already sketched the hardware path through its planned Star Mind AI satellites, which are meant to fly in large formations with more than 250 kilowatts of solar power each, and a longer-term Starmind constellation concept. Google is also pursuing orbital AI hardware, with its first Project Suncatcher satellite slated for a SpaceX Transporter-18 rideshare after earlier talks about putting data centers into orbit. Musk’s 100 percent line is a forecast, not a deployment schedule, and Maguire’s systems-level write-up has not been published yet.