How the AI Bubble Will Pop
4 hours ago
- #Capital Expenditure
- #Economic Impact
- #AI Bubble
- The AI infrastructure boom is the most important economic story in the world, but the numbers don't add up.
- Tech companies are projected to spend about $400 billion this year on AI infrastructure, surpassing historical expenditures like the Apollo program.
- AI capital expenditures in the U.S. are projected to exceed $500 billion in 2026 and 2027, while consumer spending on AI services is only $12 billion annually.
- AI startups like Thinking Machines are raising massive funding without clear products or plans, signaling potential bubble behavior.
- AI spending is creating a capital black hole, diverting resources from other sectors and potentially repeating the telecom capital drain of the 1990s.
- Large AI firms are using accounting tricks and special purpose vehicles (SPVs) to disguise their infrastructure spending, inflating profits.
- AI spending is highly concentrated in specific areas like Northern Virginia, affecting GDP growth and local economies.
- The AI build-out is warping the 2025 economy, similar to how telecom capital drained manufacturing in the 1990s.
- Private equity firms prefer large investments in data centers over smaller manufacturing ventures, exacerbating capital allocation issues.
- Energy consumption and NIMBY (Not In My Backyard) opposition may lead to the offshoring of data centers to places like India and the Middle East.