Why Everybody Is Losing Money On AI
5 days ago
- #Unsustainable Costs
- #Financial Losses
- #Generative AI
- Generative AI companies, including OpenAI and Anthropic, are unprofitable, with significant losses reported.
- The cost of running generative AI services is unsustainable, with companies spending more on compute than they earn in revenue.
- Cursor, an AI coding app, sends all its revenue to Anthropic, which then uses it to build a competing product.
- Perplexity and OpenAI spend a large percentage of their revenue on compute costs, leading to massive financial losses.
- The cost of inference (generating outputs from models) is increasing, not decreasing, contrary to common claims.
- Notion's AI costs consumed 10% of its profit margins, highlighting the financial burden of AI integration.
- OpenAI and Anthropic are expected to continue losing billions, with doubts about their projected financial improvements.
- Generative AI's core economics are flawed, with no clear path to profitability for companies relying on these models.
- Usage-based pricing may not solve profitability issues, as the compute required for useful AI outputs is too expensive.
- Generative AI's inefficiencies and high costs make it fundamentally unprofitable and unsustainable in its current form.