AI data center boom sparks fears of glut amid lending frenzy

Core Insights - The artificial intelligence data center boom is raising concerns about potential financial bubbles and unsustainable debt levels in the industry [2][3][5] Group 1: Market Dynamics - The overall infrastructure rollout cost for AI could reach $10 trillion, with lenders eager to finance these projects, raising fears of a bubble [2][6] - OpenAI plans to invest $1.4 trillion in AI infrastructure, highlighting the immense financial commitments being made in the sector [7] - At least $175 billion in data-center related credit deals have been struck in the US this year, indicating a significant surge in financing activity [15] Group 2: Risks and Concerns - There is a risk of a glut in data center construction, which could lead to some facilities becoming uneconomic and owners facing bankruptcy [3] - Concerns are growing about the level of leverage in the industry, particularly if technology underperforms expectations, making refinancing more difficult [4] - The Bank of England is reviewing lending practices to data centers due to rising concerns about spending and financing levels [16] Group 3: Financial Engineering - Companies are increasingly using financial engineering techniques, such as synthetic leases, to keep liabilities off their balance sheets while still benefiting from tax relief [12][13] - Borrowers are requesting loans exceeding 100% of construction costs, with some asking for as much as 150%, justifying this based on anticipated increases in property value [9] - The complexity of debt structures, including securitization and tranching, is creating opacity in the risk profile of these investments [8]