Musk has been on a tear lately, posting 56 tweets yesterday alone. He also changed his Twitter bio. He's been talking a lot about the new valuations of SPCX and Tesla recently. 1. In a recent podcast with Walter Isaacson, he was very direct: After completing his DOGE project, it wasn't personnel issues that made him give up; he realized that saving money wouldn't solve the US debt problem. The annual interest on US debt is more than the annual military spending, so it can't be saved. The only way out is for AI and robots to boost productivity. Here, Musk asserts that inflation is due to insufficient production capacity, not overcapacity. He believes that as long as the production rate exceeds the money supply rate, deflation will occur, and prices will fall. He didn't receive his maximum salary in the performance-based agreement, and none of the milestones for Tesla's trillion-dollar incentive plan have been triggered yet. SpaceX and Tesla are already in merger talks. 2. In the past few days, he changed his Twitter bio to focus on SpaceX and Tesla's latest business, Terafab. Terafab's narrative is that it aims to become a vertically integrated chip manufacturer, encompassing logic, storage, and packaging, and its collaboration with Intel aims to boost AI computing power to the terawatt level, serving Tesla cars and robots. If chip iterations can truly be reduced from one year to one month, the cost structure of all subsequent applications will change. 3. Starmind and Terafab are upstream and downstream entities in the same endeavor. Terafab is responsible for manufacturing the chips, while Starmind is responsible for launching a series of computing satellites into sun-synchronous orbits—generating power facing the sun and dissipating heat away from it. Connecting them with lasers to form a space-based supercomputing center, powered entirely by solar energy, would reduce the energy cost of computing power to zero. A few days ago, Professor Jan Broersen of Utrecht University in the Netherlands posted that a vacuum is the best insulator, as it cannot dissipate heat at all. Musk directly retorted, posting design diagrams of his computing satellite on his official website, showcasing his space-based thermal radiation cooling solution. (This is also the first time I've learned that because there is no air or water in space, heat cannot be convection or conduction, and can only be dissipated.) Musk's space-based computing center solutions, from Mars to the Moon, and now to Earth orbit, are becoming increasingly plausible. Terafab, Starmind, robots, and Starlink are all placed in the same narrative, shifting the valuation anchor from cars and rockets to super-intelligent infrastructure.