Claude Code 5-hour limits doubled for Pro/Max/Team/Enterprise tiers on May 6 — peak-hour throttling removed via SpaceX and Colossus 1 compute deal
Anthropic doubled Claude Code 5-hour usage limits across Pro, Max, Team, and Enterprise tiers on May 6, and removed peak-hour throttling entirely. The capacity unlock came from Anthropic's compute deal with SpaceX and Colossus 1 — the rumored compute-capacity agreement that lets Anthropic serve Claude Code workloads without the peak-hour constraints that had been the most-complained-about friction point through Q1.
The capacity story is what makes this consequential beyond a routine limit increase. Through Q1 2026, Claude Code's peak-hour throttling was the single most common complaint from professional users — developers reported workflow disruption during the 9am-noon Pacific window when usage volume routinely exceeded Anthropic's then-available capacity. Doubling the 5-hour limits plus removing peak-hour throttling is the operational difference between Claude Code being a sometimes-frustrating tool and a reliably-fast tool. The capacity to do that came from new compute commitments — the SpaceX/Colossus 1 deal is the public-facing piece.
The strategic context is the defense against Cursor and Antigravity. With Cursor's Composer 2.5 offering in-house-model economics at $20/month and Google Antigravity bundling agent tooling with Gemini 3.5 Flash's $1.50/$9 pricing, Anthropic's Claude Code value proposition has to be capability superiority plus reliability. The capability story remains intact (46% senior-developer preference per JetBrains' May survey). The reliability story needed work, which the May 6 capacity unlock provides. The remaining frontier is whether Anthropic can keep capacity ahead of agent-workflow demand growth as enterprise managed-agent deployments scale through Q3-Q4.
Nipralo — Best AI Coding Tools 2026 → · Pasquale Pillitteri — GitHub Copilot Under Pressure 2026 → · SitePoint — Claude Code vs Cursor vs Copilot 2026 Comparison →