19 hours ago
- The AI industry requires $3 trillion or more in revenue by 2030 to sustain its infrastructure, with planned data centers costing $9.5–15 trillion.
- Anthropic and OpenAI have made massive compute commitments ($770 billion+ and $330 billion+ respectively) that demand exponential revenue growth—$174 billion/year by 2029 for Anthropic and $184 billion for OpenAI.
- Token-based billing has exposed AI's lack of ROI: companies are capping spend (e.g., Uber at $1,500/month per user) and struggling to measure value.
- The AI economy is circular: labs pay hyperscalers, who invest back into labs, but the con requires ever-increasing demand—needing two more OpenAI-scale companies to justify existing compute.
- AI is slowing down as customers realize costs are unsustainable; many firms have no visibility into spend until after billing.
- Workers in tech are miserable, forced to use AI tools without clear benefit, and the industry is run by 'Business Idiots' disconnected from production.
- The author claims an upcoming story will burst the AI bubble, sourced from tech workers who are 'in agony'.