Premium: AI Is Getting Way Too Expensive

Source: Where’s Your Ed At Date: 2026-07-31 URL: https://www.wheresyoured.at/premium-ai-is-getting-way-too-expensive/

Summary

Ed Zitron argues the AI industry’s funding vastly outstrips its revenue: roughly $110B in trailing-twelve-month revenue across the entire industry against OpenAI alone raising $122B in March 2026 and AI startups collectively raising $145B in Q1 2026. His thesis is that rising infrastructure costs will force vendors to raise prices at exactly the moment their only viable customers (other AI companies and enterprises with thin margins) can least afford it, turning “hundreds of billions” in compute commitments into an unsolvable economic bind rather than a temporary funding gap.

Implications

• Token-cost-as-operating-cost, adversarial read. This is the bear case running directly against the same week’s OpenAI GPT-5.6 price cuts (Luna -80%, Terra -20%) — Zitron would read those cuts not as an efficiency win but as evidence vendors are subsidizing usage to defend market share while the underlying unit economics stay upside-down. Worth holding both claims in tension: falling sticker prices don’t by themselves resolve the funding/revenue gap he’s pointing at.
• No tracked-dependency angle; this is landscape/critical-press signal rather than a product or capability event.