Anthropic has revised Claude Code’s weekly usage limits so that, after a September 14 policy change, the new permanent cap will sit about 17% below today’s temporarily boosted allowance for most paid plans. The change affects Pro, Max, Team and seat-based Enterprise customers who have enjoyed a 50% promotional increase in weekly limits since May 13, 2026.
This article aggregates reporting from 3 news sources. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
Anthropic’s quiet reset of Claude Code’s weekly limits is a small but telling move in how frontier labs are balancing growth, compute supply and customer expectations. After months of a 50% promotional boost to weekly quotas, reverting to a level that is only 25% above the original baseline but 17% below what users have grown used to is effectively a price increase in usage terms, even if list prices stay the same. It signals that sustained demand for coding agents is eating into Anthropic’s capacity budget enough that it needs to normalise consumption, not just keep throwing extra quota at power users. ([en.softonic.com](https://en.softonic.com/articles/claude-code-weekly-limits-updated-less-for-most-users))
For the race to AGI, this matters less as a billing tweak and more as a data point on how constrained even well‑funded labs remain on GPU time. Claude Code is one of Anthropic’s heaviest workloads, and trimming its headroom while promising better visibility into usage suggests the company wants to steer usage toward higher‑margin or strategically valuable activity rather than brute‑force volume. Expect other labs to follow similar patterns: generous launch promos to seed workflows, followed by more conservative, metered regimes as they discover how much their agent products really cost to run at scale. That dynamic will shape which use cases get commoditised and which stay premium in the run‑up to more capable agentic systems.