openaienmodel: gpt-5-mini-2025-08-07
Higher usage limits for Claude and a compute deal with SpaceX
Key Points
- Claude Code 5-hour limits doubled
- SpaceX deal adds 300+ MW (~220k GPUs)
- Claude Opus API rate limits increased
Summary
On 2026-05-06 the company announced a new compute partnership with SpaceX and immediate increases to Claude usage limits. The SpaceX deal provides access to Colossus 1 capacity—>300 MW (over ~220,000 NVIDIA GPUs) arriving within the month—which, together with other supplier agreements, expands capacity for Claude Pro and Max customers. Effective immediately, Claude Code and Claude Opus API limits have been raised for paid tiers.
Key Points
- Usage limits
- Claude Code: five-hour rate limits doubled for Pro, Max, Team, and seat-based Enterprise plans.
- Claude Code: removal of peak-hours limit reduction for Pro and Max accounts.
- Claude Opus models: API rate limits significantly increased (see API docs for exact per-model values).
- New compute capacity
- SpaceX Colossus 1: >300 MW capacity (~220k NVIDIA GPUs) coming online within a month.
- Other agreements: up to 5 GW with Amazon (nearly 1 GW by end of 2026), 5 GW with Google/Broadcom (starting 2027), $30B Azure capacity with Microsoft/NVIDIA, and $50B investment with Fluidstack.
- Hardware mix: AWS Trainium, Google TPUs, NVIDIA GPUs; exploring orbital AI compute.
- International expansion & compliance
- In-region capacity additions to meet regulated customers (financial, healthcare, government).
- Amazon collaboration includes additional inference capacity in Asia and Europe.
- Commitment to cover US consumer electricity price increases from data centers; exploring similar commitments and local investments internationally.
What engineers should do
- Review updated Claude Opus API docs and adjust client rate-limit handling and backoff logic.
- Re-evaluate capacity planning for Pro/Max deployments and CI/perf testing given higher quotas.
- Update monitoring/alerts for quota exhaustion and billing exposure.
- Coordinate with security/compliance teams for in-region deployment requirements and data residency.
Impact
These changes reduce throttling and increase capacity for heavy Claude users immediately and add multi-year capacity growth through several large supplier partnerships, improving throughput and regional availability for regulated workloads.