Agents are TOKEN-HUNGRY
> 1 Agent Request = 20 General LLM Prompts
> ~50% of agentic workloads exceed 128k tokens
> Up from 50k tokens last year
> HUGE tailwind for onchain inference players
And the coolest thing?
Most asymmetric inference opps are investable on Base
The inference cycle is here
> Token demand is going parabolic
> 2x increase in token consumption every single quarter
> Web2 inference players have more than doubled their rev in 3-6 months
> Web3 players are carving out their moats on privacy, uncensored AI, and cutthroat inference pricing
> Forecast: Web3 players hit $1B (~10% of inference API market) by 2028
More on today's article "The Inference API Market" 👇
[Link in bio]