AI data center financing continues, but investor sentiment turns cautious, favoring short terms and amortization structures (Morgan Stanley)
After communicating with investors, Morgan Stanley found that AI-related debt issuance has surged, but investor sentiment has become cautious.
After communicating with investors, Morgan Stanley found that AI-related debt issuance has surged, but investor sentiment has become cautious. While allocations are still increasing, there is a general preference for shorter maturities and amortization structures to hedge against risks such as overbuilding and technology obsolescence. It is predicted that investment-grade AI-related debt issuance will reach US$350-400 billion in 2026, maintaining a bearish view on the investment-grade technology sector. One-sentence conclusion: The financing craze for AI infrastructure continues unabated, but the capital structure is becoming more conservative, reflecting the market's rational vigilance against the risk of AI bubbles. Positive/negative: Negative for the investment-grade technology sector (due to supply pressure and shifting investor preferences). The impact on high-yield data center builders is neutral, with investors focusing more on construction risks and hedging mechanisms rather than pure long-term growth stories. Catalysts: 1) Capital expenditure guidance from hyperscalers; 2) Reports on AI data center construction progress and potential delays; 3) Rating adjustments on data center bonds by credit rating agencies.