Your Claude Code Limit Just Doubled: The SpaceX Deal Behind It

Elon Musk just handed Anthropic 220,000 more GPUs.

Yeah, the same Musk who once called Anthropic "misanthropic and evil."

And you're feeling the effect right now: your Claude Code usage limits just doubled.

On May 6, 2026, Anthropic officially announced a deal with SpaceX — full access to Colossus 1, a data center in Memphis, Tennessee. 300+ megawatts of capacity. 220,000 NVIDIA GPUs. Per Anthropic Official Blog, this isn't a gradual expansion. It's effective now.

But why would Musk — Anthropic's most vocal critic — supply compute to a direct competitor?

And what other changes are coming after this?

The answer is more pragmatic — and more surprising — than you'd expect.

Developers Have Been Hitting the Same Wall for a While

Teams deploying Claude Code report 2-3x productivity gains on large-scale refactoring tasks, according to Developers Digest 2026. As dependencies run deeper, every limit hit hurts more — and Uber proved it: they burned through their entire 2026 AI budget in just four months, largely credited to Claude Code, per Humai Blog.

But before we get to the deal:

If you use Claude Code seriously, you know exactly what this feels like. You're mid-refactor. 400-something files, three PRs open, CI/CD running. Then you get the message you hate most:

"Usage limit reached."

Not a skill problem. Not a bad strategy. Anthropic's servers being overloaded.

Uber isn't an outlier — they're just the most transparent. Imagine how many other teams are going through the same thing without anyone writing articles about them.

That's why the SpaceX deal isn't just infrastructure news. It's a direct response to demand that's been blowing past old capacity limits for a while.

The Unlikely Alliance: Why SpaceX Got Involved and What It Means for Your Claude Code Usage Limits

Anthropic got 220,000 NVIDIA GPUs from SpaceX via Colossus 1 in Memphis, Tennessee — a deal that shocked people because Elon Musk is one of Anthropic's most vocal critics. Per Anthropic Official Blog, 100% of Colossus 1's capacity (300+ megawatts) is scheduled to be fully utilized before the end of May 2026.

Before we get to the numbers, let's acknowledge how weird this is:

Musk publicly called Anthropic "misanthropic and evil." Not a slip of the tongue. It's a position he's held pretty consistently.

But it turns out:

SpaceX has Colossus 1 — and needs revenue from its capacity. Anthropic needs GPUs to handle the demand surge. Both interests met at one point, regardless of anything anyone's ever tweeted.

This isn't a friendship story. It's a story of capacity meeting demand.

And Anthropic isn't stopping at one deal:

According to Axios and Bloomberg, they're running several infrastructure pipelines in parallel. Amazon: 5GW agreement, with 1GW active before end of 2026. Google + Broadcom: an additional 5GW, targeting 2027. Microsoft Azure: $30 billion worth of capacity. Fluidstack: $50 billion infrastructure investment.

SpaceX isn't the only one. But this deal is the fastest to produce real capacity — and the one that's most directly changing your Claude Code usage limits in concrete terms.

4 Things That Changed in Your Plan Right Now

Anthropic doubled the five-hour Claude Code limit for Pro, Max, Team, and Enterprise plans, effective immediately, per the official May 2026 announcement. Peak-hour restrictions for Pro and Max have been removed entirely, giving consistent 24/7 access. Opus API rate limits were also raised for developers accessing the model directly via API.

Here's what that looks like in practice, per Engadget, 9to5Google, and Dataconomy:

1. The five-hour Claude Code limit doubled — for all paid plans

Pro, Max, Team, and seat-based Enterprise all get 2x the previous limit. No upgrade needed, no extra cost. Effective now, no waiting list. Long coding sprints won't get cut off mid-session anymore — and this is what teams who've made Claude Code part of their daily workflow will feel most directly.

2. Peak-hour restrictions are gone — Pro and Max

Before, during US peak hours (East Coast morning to midday), rate limits were cut automatically without warning. Now Pro and Max get consistent access all day. Developers outside US peak hours — who've noticed certain "slow periods" — will feel the difference immediately.

3. Opus API rate limits are up

For teams building autonomous agents or pipelines with intensive calls to Claude Opus directly via API — not through the Claude Code CLI — per-token rate limits have been raised. This matters for production systems that depend on Anthropic's most powerful model.

4. Effective immediately, no phased rollout

All changes are active now for all eligible plans. No waiting list, no pilot phase. Open Claude Code today and your new limits are already in effect.

What to Watch After This

SpaceX is actively exploring orbital compute — server farms outside Earth — while Amazon is pushing toward 1GW active before end of 2026 and Google+Broadcom are on track for 5GW in 2027. Anthropic itself is reportedly in IPO talks with a timeline around June 2026, according to CoinDesk.

Here's the closing insight you need to take with you:

The Musk-Anthropic rivalry hasn't gone away. But in the AI infrastructure arms race, pragmatism beats ideology. SpaceX has the GPUs. Anthropic has the demand. Money talks louder than any tweet.

And there's more coming:

SpaceX orbital compute — if it materializes — will redefine what low latency means for AI inference. Amazon is sprinting to 1GW active before year-end. Google and Broadcom are on track for 5GW in 2027. Anthropic will most likely IPO in mid-2026.

The takeaway is simple:

Compute is the new oil. Whoever controls the infrastructure controls who gets to scale AI to production. The doubled Claude Code usage limits today are a side effect of an infrastructure race that's way bigger than any single coding tool.

Common Questions About the Claude Code Usage Limit Changes

What are the Claude Code usage limit changes from the SpaceX deal?

Anthropic doubled the five-hour Claude Code limit for all paid plans (Pro, Max, Team, Enterprise) and removed peak-hour restrictions for Pro and Max, effective immediately as of May 2026. Opus API rate limits were also raised for developers accessing directly via API. All changes are active with no waitlist or extra cost.

Why would SpaceX deal with Anthropic when Elon Musk constantly criticizes them?

SpaceX owns Colossus 1 — a data center in Memphis with 220,000 NVIDIA GPUs and 300+ megawatts of capacity — and needs revenue from that capacity. Anthropic needs compute to handle the surge in demand. Even though Musk has called Anthropic misanthropic, business pragmatism beat personal rivalry in this May 2026 deal.

Upgrade Now or Save for Your Next Sprint

Your Claude Code limit has doubled. Peak-hour throttling is gone. And behind it all are 220,000 GPUs from a data center built by Anthropic's own rival's partner.

The best irony of 2026 — and the one that benefits you most as a developer.

Right now: Upgrade to Claude Max and start using your new limits today — link in bio.

Save for later: Save this article before your next sprint planning session — it explains why your team's AI coding budget just got a lot more headroom.