How much revenue is needed to justify the current AI spend?
12 hours ago
- #economic-risk
- #capital-misallocation
- #AI-bubble
- The author questioned the financial viability of AI investments, sparking widespread industry acknowledgment of unworkable economics.
- Industry insiders, including datacenter owners and investors, share concerns but remain silent, fearing isolation in their skepticism.
- Revised depreciation estimates for AI datacenters (3-5 years) suggest breakeven revenues need to be $320B-$480B, far exceeding current returns.
- Current AI revenue (~$15B-$20B) is vastly insufficient against $400B annual capex, hinting at a looming financial crisis.
- Comparisons to historical bubbles (railroads, fiber-optic) highlight risks of capital misallocation and potential economic collapse.
- AI's economic impact may already account for 2% of US GDP, with multiplier effects possibly doubling its footprint.
- Wealth effect from AI-driven stock gains could artificially inflate GDP growth, masking underlying economic fragility.
- The AI bubble's burst could mirror past crises, with rapid obsolescence of datacenters exacerbating capital destruction.
- Industry desperation is emerging, reminiscent of pre-collapse behaviors in prior tech bubbles (e.g., Lucent, Nortel).
- Timing of the AI reckoning remains uncertain, but parallels to historical financial panics suggest inevitable fallout.