What We’re Reading (Week Ending 26 October 2025) : The Good Investors %

Market Analysis - The current market appears to be correctly priced with a 10-year Treasury rate of 4% and an S&P 500 index P/E of 25.5x, aligning with historical averages [5] - Historical market bubbles have shown significant overvaluation, with the market being overpriced by 100% before Black Monday and 40% during the internet bubble, but currently, there is no such stretch in valuations [6] AI Investment Landscape - There are contrasting views on AI investments, with concerns about the return on investments in AI infrastructure versus fears of AI displacing entire industries [6][7] - The potential job displacement due to AI could lead to significant cost savings for companies, with estimates suggesting that replacing 10% of U.S. jobs could save $1.6 trillion annually [8] Meta's Data Center Strategy - Meta has entered a joint venture with Blue Owl Capital for a $27 billion Hyperion Data Center campus, owning 20% and signing a short-term operating lease with options for extension [29][30] - The deal includes a "Residual Value Guarantee" (RVG) to cover potential losses if Meta exits the lease early, with an estimated maximum exposure of $22.95 billion [32][33] Debt and Yield Considerations - The bonds issued for the Hyperion project have a yield of 6.58%, which is closer to junk bond yields due to the perceived risks associated with the specialized nature of the data center [34][35] - Investors demand higher yields to compensate for the structural subordination of the debt and the potential obsolescence of the data center [34] AI Infrastructure Buildout - The AI capital expenditure in the U.S. is currently around $400 billion per year, with a need for data centers to generate equivalent revenue over their lifetime to avoid being classified as a bubble [42] - The construction of new data centers is competing for skilled labor, which may become a significant bottleneck in the AI infrastructure buildout [39][40] Global Competitive Landscape - China is positioned to dominate the manufacturing of components necessary for AI infrastructure, producing significantly more solar panels and electricity than the U.S. [46] - The collaboration between Chinese state entities and tech giants could facilitate the rapid buildout of AI capacity, contrasting with the competitive struggles among U.S. hyperscalers [46]