Anthropic Doubles Claude Code Rate Limits, Signs Compute Deal with SpaceX
Anthropic announced higher usage limits for Claude Code and the Claude API, effective immediately, alongside a compute partnership with SpaceX that brings over 300 MW of new capacity online within a month.
Claude Code Rate Limit Changes
- Five-hour rate limits for Pro, Max, Team, and seat-based Enterprise plans are doubled.
- Peak hours limit reduction for Pro and Max accounts is removed entirely.
API Rate Limit Increases for Claude Opus
Rate limits for Claude Opus models have been raised considerably. See the full table in the source article for exact numbers.
Compute Partnership with SpaceX
Anthropic signed an agreement to use all compute capacity at SpaceX's Colossus 1 data center, providing over 300 MW of new capacity (more than 220,000 NVIDIA GPUs) within the month. This capacity directly improves service for Claude Pro and Max subscribers.
This joins other large compute deals: up to 5 GW with Amazon (nearly 1 GW by end of 2026), 5 GW with Google and Broadcom (coming online 2027), $30 billion Azure capacity with Microsoft and NVIDIA, and a $50 billion investment with Fluidstack.
Anthropic expressed interest in partnering with SpaceX for multiple gigawatts of orbital AI compute capacity.
International Expansion
Enterprise customers in regulated industries need in-region infrastructure. The Amazon collaboration includes additional inference capacity in Asia and Europe. Anthropic committed to covering consumer electricity price increases caused by its US data centers and is exploring extending that commitment to new jurisdictions.
📖 Read the full source: HN AI Agents
👀 See Also
Claude Code 2.1.84 adds general-purpose agent prompt and PowerShell tool, removes redundant prompts
Claude Code 2.1.84 introduces a new general-purpose subagent prompt for codebase operations and a PowerShell tool description with sleep command avoidance guidelines. The update removes nine redundant prompts and simplifies multiple tool descriptions.
Qwen KV Cache Quantization Deep Dive: PPL, KL Divergence, and Asymmetric K/V Results
Second round of benchmarks on Qwen 3.6-35B-A3B with KV cache quantization: perplexity, KL divergence, asymmetric K/V combos, and 64K context depth on Apple M5 Max.
Claude Code v2.1.281: Bedrock assume_role, Guardrails, and a Long List of Resume Fixes
Claude Code v2.1.281 adds assume_role on Bedrock upstreams, per-request Bedrock guardrails, telemetry.resource_attributes, an "attribution": false setting, and fixes a pile of session-resume and proxy stream bugs.
OpenAI's Sam Altman Supports Anthropic's Pentagon Red Lines, Proposes Technical Safeguards
OpenAI CEO Sam Altman has expressed support for Anthropic's ethical stance against Pentagon AI use for mass surveillance and autonomous weapons, while proposing technical safeguards like cloud-only deployment as a resolution.