Anthropic Partners with SpaceX, Doubles Claude Usage Limits
Summary
- • Anthropic signs compute deal with SpaceX, gaining 300MW and 220,000 NVIDIA GPUs
- • Claude Code five-hour rate limits doubled for Pro, Max, Team, and Enterprise plans today
- • Peak hours limit reduction removed for Claude Code Pro and Max accounts
- • Anthropic now has multi-gigawatt compute agreements with Amazon, Google, Microsoft, and SpaceX
Details
Anthropic secured full use of SpaceX's Colossus 1 data center, adding 300MW and 220,000 NVIDIA GPUs
The agreement gives Anthropic access to this capacity within the month. This marks a significant compute acquisition outside traditional cloud providers, adding SpaceX to Anthropic's growing infrastructure network.
Claude Code five-hour rate limits doubled across Pro, Max, Team, and Enterprise plans, effective immediately
This directly addresses a key pain point for developers using Claude Code for extended sessions. Both the doubling of limits and the removal of peak-hours reductions for Pro and Max accounts take effect May 6, 2026.
Peak hours rate limit reduction on Claude Code eliminated for Pro and Max subscribers
Previously, usage during peak periods was throttled for these tiers. Removing this restriction means developers get consistent throughput regardless of time of day.
API rate limits for Claude Opus models raised considerably
Opus is Anthropic's most capable model tier. Higher API limits make it more viable for enterprise and power users running high-volume or latency-sensitive workloads.
Anthropic has up to a 5GW compute agreement with Amazon, including nearly 1GW of new capacity by end of 2026
Amazon is already Anthropic's primary cloud partner via AWS. This agreement scales that relationship dramatically, with near-term capacity arriving before year-end.
5GW agreement with Google and Broadcom secured, beginning in 2027
Broadcom's inclusion suggests custom silicon or networking infrastructure is part of the deal. Google TPUs are already part of Anthropic's training stack, extending an existing relationship at massive scale.
Microsoft and NVIDIA partnership includes $30 billion of Azure capacity
This adds Microsoft as a major compute provider alongside Amazon and Google, giving Anthropic multi-cloud redundancy and access to NVIDIA GPU supply through Azure.
$50 billion committed to American AI infrastructure through Fluidstack partnership
Fluidstack is a distributed GPU cloud provider. This investment signals Anthropic is diversifying beyond hyperscalers to purpose-built AI infrastructure at massive scale.
Anthropic expressed interest in developing orbital AI compute with SpaceX, potentially spanning multiple gigawatts
This is an early-stage interest, not a signed agreement. Space-based compute at this scale would be novel and could support geographically distributed inference with reduced terrestrial infrastructure dependency.
Enterprise inference capacity expanding in Asia and Europe via Amazon collaboration
Regulated industries — financial services, healthcare, government — require in-region data processing. This expansion directly targets enterprise customers who cannot route data through US-only infrastructure.
Anthropic committed to covering consumer electricity price increases caused by its US data centers
An unusual public commitment addressing community concerns about AI infrastructure's energy cost impact. Anthropic is also exploring extending this commitment internationally as it expands.
Partnership = new business agreement, Product Launch = new or updated capabilities, Infrastructure = compute/data center buildout, Financials = capital commitments, Strategy = forward-looking positioning, Policy = regulatory or public commitments
What This Means
Anthropic is executing a sweeping compute expansion across nearly every major infrastructure provider — hyperscalers, GPU clouds, and now SpaceX — that collectively spans tens of gigawatts of future capacity. The immediate payoff for users is tangible: doubled Claude Code rate limits and removal of peak-hour throttling land today, making the platform meaningfully more usable for developers and enterprises. The scale of these deals positions Anthropic to sustain aggressive model development and inference growth without hitting the capacity ceilings that have constrained AI providers during demand spikes. The SpaceX partnership also includes expressed interest in orbital compute — speculative at this stage, but signaling ambitions that go well beyond conventional cloud infrastructure.
Sentiment
Broadly excited about usability improvements and strategic compute access, some amusement at unlikely Musk-Anthropic alliance
“Smart move by Anthropic, their compute crunch seems to have been starting to take a slight toll in the last days on usage growth. Getting new compute online today is very valuable.”
“Claude Code limits doubled, peak throttling removed, API limits raised — insane partnership with SpaceX's Colossus 1 unlocking 300MW+ capacity. Best news for Claude Code users!!!”
“Anthropic just signaled a major AI reality: Compute is becoming a strategic infrastructure. The partnership with SpaceX boosts capacity, enabling higher limits. In the AI race: Compute + energy + infrastructure may decide scale.”
“Wow. The AI supply crunch is real. Frontier labs are desperate for compute. Musk has compute capacity but a meh model, and Anthropic has a fantastic model with weak capacity. Now I wonder if Elon continues to refer to his new business partner as “woke AI,” “Misanthropic,” etc.”
notably highlighting irony given Elon's past criticisms of Anthropic
“Anthropic just doubled the usage limits on Claude Code — and reversed a recent cut they'd made during peak hours. The compute to make this happen came from a new partnership with SpaceX.”
Split
~90/10 positive/neutral — excitement over fixed limits vs minor notes on past Elon-Anthropic tensions.