Why AI fail will fail at scale — and how to fix it now
Many assume AI will follow the path of ecommerce, search, and social — lose money for years, achieve massive scale, then discover profitability.
This will not work for AI.
Advertising only works when content is free
- Ecommerce doesn’t pay for its content. Sellers provide it for free.
- Search doesn’t pay for its content. Publisher provide if for free.
- Social doesn’t pay for its content. Users provide it for free.
- But AI must pay to generate unique content for each and every user query – at its own expense.
Subscriptions won’t work – even at 100% penetration
If every man, woman and child subscribed to AI at $20/month, AI would barely break even – because the cost of AI infra and compute is many levels of magnitudes greater than ecommerce, search and social.
What will work for AI
Revenue: Billing must be tied to actual cost. Not flat-rate subscriptions nor token counts, because flat-rates are not tied to actual cost and tokens are not an accurate proxy for compute.
Expense: Costs must be cut via optimization, both prior to inference, and within the context window where the cost of inference is incurred.
If this resonates, you’re welcome to reach out.