Why AI Funding Is So Price-Insensitive

Summary

Andrew Sheets explains the economic concept of inelastic demand applied to AI infrastructure spending. He notes that despite soaring costs for components like copper, gas turbines, and memory, large tech companies continue to accelerate investment. This price-insensitive spending supports US growth and equity earnings but also risks fueling inflation and widening corporate bond spreads.

• AI infrastructure investment is estimated at $800 billion in 2025, nearly double 2024 and triple 2023.
• Spending forecasts for 2026 have been revised higher, with $1.1 trillion projected for 2027.
• Component prices like copper (+40%), gas turbines (+50%), and memory (+150-300%) have surged but demand remains robust.
• The inelastic nature of AI spending provides a tailwind for US growth and equity earnings this year.
• Risks include persistent inflation as input costs rise and potential widening of corporate bond spreads from increased borrowing.
• The speaker cites Mike Wilson's view that AI spending supports earnings but questions the ultimate returns on the historic investment.