Smorster

Claude Just Solved Session Limits

Nate Herk | AI Automation · 10m 22s · Watch on YouTube · 11 sources

Decision Card

Effort: Fifteen minutes — open Anthropic’s usage-limits announcement, check which tier you’re on, and re-run one Opus workflow or /opus-heavy Claude Code session you previously rationed to see if you still hit the wall.

Honest take: The three usage changes are real and well-sourced, but the title oversells them as “solving” session limits — weekly caps were not raised, only the 5-hour ones, and the video glosses over that the “SpaceX” compute is actually access to xAI’s Colossus 1 data center (which SpaceX rented out), not net-new SpaceX hardware. The host also garbles the API numbers (the real headline is tier-1 input going 30K→500K tokens/min, not the muddled “8,000→80,000” he recites).

Concrete next steps:

  • Read the primary source — anthropic.com/news/higher-limits-spacex — and the official rate-limit docs to see your exact Opus tier numbers (~10 min).
  • If you previously downshifted to Haiku/Sonnet purely to conserve session budget, re-test one real task on Opus now that 5-hour limits doubled (~15 min).
  • Skip if you’re a Free-tier user or only use Claude in the chat app casually — the meaningful changes target Claude Code subscribers (Pro/Max/Team) and API/Opus builders, not light consumer use.

TL;DR

Anthropic doubled Claude Code’s 5-hour rate limits, removed peak-hours throttling for Pro/Max, and sharply raised API rate limits for Opus models — all backed by a new compute deal giving it access to SpaceX/xAI’s Colossus 1 data center (300+ MW, 220,000+ NVIDIA GPUs). The video walks through what each change unblocks for builders and flags a speculative future interest in “orbital” AI compute.

Key Points

  • Anthropic announced a compute deal tied to SpaceX that increases capacity and lets it raise usage limits for Claude Code and the API 00:01
  • The past quarter was plagued by outages, which the host attributes mainly to insufficient compute for surging demand 00:33
  • Claude Code’s 5-hour rate limits are doubled for Pro, Max, and Team plans, effective immediately 01:11
  • The peak-hours limit reduction on Claude Code for Pro and Max accounts has been removed 01:25
  • API rate limits were raised considerably for Opus models, with the lowest tiers getting the biggest multiples (the host cites ~16x and ~10x jumps) 02:46
  • The compute came via 300 megawatts of capacity and over 220,000 NVIDIA GPUs, acquired very quickly 05:34
  • Anthropic and SpaceX expressed interest in developing multiple gigawatts of orbital AI compute — GPUs in space — as a future direction, not a current deployment 06:36
  • The deal sits alongside Anthropic’s other compute partnerships (Amazon, Google/Broadcom, Microsoft, NVIDIA, Fluidstack) 03:17
  • Practical advice for builders: retest Opus workflows that broke on old limits, use Opus more freely, and treat the 1M context window and multi-agent/parallel workflows as more viable now 07:29

Notable Quotes

“This past couple months, this past quarter, has been awful with outages. There’s been so many times where Claude has just died while you’re trying to use it.” 00:33

“Anthropic and SpaceX have expressed interest in developing multiple gigawatts of orbital AI compute capacity. Which means GPUs in space.” 06:39

“If you tried building an Opus agent six months ago and you gave up because of rate limits, the wall might not actually exist anymore and it’s worth a reattempt.” 07:32

Verified Claims

Claim: Anthropic doubled Claude Code’s 5-hour rate limits for Pro, Max, and Team plans. 01:11

Claim: The peak-hours limit reduction on Claude Code for Pro and Max was removed. 01:25

Claim: API rate limits for Opus were raised considerably; tier-1 input went from 30K to ~500K tokens/min. 02:53

Claim: The deal gives access to 300+ MW of capacity and over 220,000 NVIDIA GPUs, available within the month. 05:37

Claim: Anthropic and SpaceX expressed interest in developing multiple gigawatts of orbital AI compute. 06:39

Claim: The compute is “a partnership with SpaceX, Elon Musk’s company.” 00:01

Claim: Code with Claude 2026 was Anthropic’s developer conference in San Francisco, London, and Tokyo. 00:12

Tools, Papers & Standards Mentioned

Follow-up Questions

  1. Did the doubled 5-hour limits meaningfully reduce real-world throttling for heavy Claude Code/Max users, or did weekly caps (unchanged) become the new binding constraint?
  2. What are the actual published per-tier Opus API rate-limit numbers post-change, and how do input vs. output ceilings compare to the video’s loosely-recalled figures?
  3. How real is “orbital AI compute” — what are the credible timelines, power/cooling/latency tradeoffs, and which (if any) partner has committed capital beyond stated “interest”?

Sources