Rohan Paul (@rohanpaul_ai) · 10/12/2026, 02:08:43
Agent Swarms Drive Token Volume to Cheaper Models, Undermining Per-Seat AI Pricing
Agent swarm architectures route high-volume, verifiable narrow tasks from child agents to cheaper models, causing token consumption to surge (up 14x in six months). This exponential growth invalidates the assumption that human prompting frequency caps inference costs, forcing a shift from per-seat subscriptions to usage-based metering.
SOURCE COVERAGEFull Original Report
Rohan Paul (@rohanpaul_ai) · 2026-10-11 18
Agent swarms send most new token volume to cheaper models, so agent token counts will outrun frontier-model revenue.
In a swarm, a parent agent plans while child agents handle narrow jobs like searching files or running tests. Narrow jobs are usually easy to check, so developers can route them to the cheapest model that passes.
And this massive token-explosion is also the major reason, flat-rate and per-seat pricing for AI software will give way to metering, because it relied on human prompting speed to cap usage. A person sends only so many prompts a day, so vendors could charge per user and still predict inference bills. But that equation will not hold with agent swarms working 24x7.
↳ 引用 @a16z (a16z): Your next users are machines
Agents burn nearly 5x the tokens people do, up 14x in six months
