被特朗普关税政策拖累，软银-OpenAI“星际之门”项目融资遇阻

Core Viewpoint - SoftBank's $100 billion "Interstellar Gateway" project in the U.S. faces significant challenges due to economic risks stemming from the Trump administration's tariff policies, leading to stalled financing negotiations [1][2] Group 1: Project Financing and Economic Factors - As of three months after the project's announcement, SoftBank has not finalized its financing plan or engaged in substantial discussions with banks and private equity firms [2] - Initial talks with various lending institutions and asset management companies were held, but economic volatility and declining AI service costs hindered progress [2] - Higher capital costs and concerns over potential global recession affecting data center demand have made lenders cautious about high-risk investments [2][3] Group 2: Market Conditions and Investor Sentiment - Analysts indicate that tariffs could increase data center construction costs by 5% to 15%, raising concerns among potential investors about industry overcapacity [3] - Major tech companies have invested heavily in AI data centers, yet recent adjustments in strategies by firms like Microsoft and Amazon have led to a slowdown in data center projects [3] Group 3: OpenAI and Project Developments - OpenAI is experiencing internal turmoil, which adds uncertainty to the project, although SoftBank plans to proceed with its $30 billion investment in OpenAI [5] - Despite challenges, construction of the first data center in Abilene, Texas, is ongoing, which is expected to become the world's largest AI training facility [5][6] - SoftBank is evaluating additional data center locations, considering factors like power supply and investment returns to ensure project viability [5] Group 4: Long-term Outlook and Returns - SoftBank's founder believes current market fluctuations are temporary and that the investment will yield substantial returns, with estimates suggesting a 15% to 20% return rate for the "Interstellar Gateway" project [6] - Achieving expected returns will depend on ensuring high utilization rates of the data centers, which remains uncertain at this stage [6]