[Morgan Stanley: SpaceX is passing the peak of the new cloud and moving towards robotics and space infrastructure; maintaining an "overweight" rating and a target price of $300] On November 10, Morgan Stanley issued a research report stating that SpaceX is passing the peak of the new cloud (neocloud) and distinguishes the winners with its intelligence density advantages per watt, per dollar, and per second. Looking forward to 2027, the bank expects the company to make a profound shift towards robotics, mesh networks and space infrastructure. The bank predicts that SpaceX will have slightly more than 8 gigawatts of computing power by the end of 2027 (compared to Musk's expectation of more than 13 gigawatts), driven by the availability of chips, electricity and financing, and estimates that the issuance size in fiscal 2027 will be about US$120 billion. The bank pointed out that the proportion of the new cloud model in total revenue and profit is expected to decline steadily/rapidly; SpaceX completed the construction of Minihard (more than 400 megawatts) in the green field in about five months, and has a leading power-on and online time advantage. The bank believes that the total addressable market (TAM) of digital AI is about 30 trillion US dollars, and that of physical AI is measured in hundreds of billions of US dollars (equivalent to several times the total global GDP); the cost of extra-atmospheric computing (USD per watt) will be the same as that of terrestrial computing in the next ten years, which is within the reach of the final discounted cash flow value of most land power shells. The bank reminded attention to Terafab's next milestone, including the amount of investment and required human employment, and maintained an "overweight" rating and a target price of US$300.