Sou Hu Cai Jing·2026-01-15 10:22Core Insights - The construction boom of data centers driven by artificial intelligence continues, but concerns about potential industry collapse due to over-speculation and rising investment demands are increasing [2] - Moody's report estimates that at least $3 trillion will be needed in investments by the end of this decade to keep up with expected capacity expansion levels [2] - Major challenges for new data center development include power supply issues, local opposition, skilled labor shortages, and rising costs of construction materials and key equipment [5] Investment Projections - The six major U.S. cloud service providers are projected to have capital expenditures nearing $400 billion in 2025, expected to reach $500 billion in 2026 and $600 billion in 2027 [3] - Global total investment in data centers is expected to peak in 2029 and then decline starting in 2030 [3] Challenges in Development - New data center projects face increasing resistance primarily due to power supply issues and public concerns over resource consumption [3] - Developers are under pressure to shorten construction timelines to meet the demands of large-scale tenants, which conflicts with high demand for skilled labor and materials [3][5] Tenant Behavior - Some tenants are now willing to take on delivery risks they previously avoided, including waiving power and utility availability requirements from completion conditions [4] - OpenAI's increasing presence in the AI ecosystem poses growing credit risks, as its financing relies heavily on long-term lease agreements with major cloud service providers [4]