AI Pulse · Frontier Labs Watch

Anthropic has been giving Claude Code users a 50% boost to weekly usage limits since May 13. That boost expires tomorrow, August 19. Pro, Max, Team, and legacy Enterprise plans all go back to standard limits. There’s no mention of any change to what you’re paying, so it’s not a price cut or a price hike. It’s a quiet reduction in how much capacity you get for the same money.

If you or your team lean on Claude Code for real development work, this matters. Anyone running close to the old ceiling is going to hit walls sooner next week than they did last week. That’s especially awkward timing, because this is landing right in the middle of a pricing fight between OpenAI, Cerebras, and xAI’s Grok 4.6. When your competitors are cutting prices or expanding access, and your AI vendor is quietly pulling back capacity, it’s worth paying attention to which way the wind is blowing before you build more workflows on top of one tool.

The developer reaction backs this up. The announcement pulled 216 points and 163 comments on Hacker News, which is a lot of noise for what sounds like a routine expiration notice. People built habits and workloads around the extra headroom, and now they’re recalculating. You can read the original notice from Claude if you want the exact plan-by-plan breakdown.

The bigger lesson for owners is simple. If AI tools are becoming core to how your team ships work, you don’t want to be surprised by a vendor’s usage terms changing overnight. This is the kind of thing we build into an AI command centre, so you can see usage, costs, and vendor shifts in one place instead of finding out the hard way when a limit hits mid project.