Anthropic just gave developers more AI coding power, and SpaceX is a big reason why.
Anthropic has doubled Claude Code usage limits for paid users after signing a major compute partnership with SpaceX.
The upgrade affects Claude Pro, Max, Team, and Enterprise plans, giving users roughly twice the coding capacity during five hour usage windows. Anthropic also removed peak-hour restrictions for some users and increased API limits for its Claude Opus models.
Claude Code is Anthropic’s AI coding assistant that helps developers write code, edit files, run commands, and handle programming tasks directly from natural language prompts.
And demand for it has been growing fast.
The company said the improvements are possible because of a new agreement giving Anthropic access to SpaceX’s massive Colossus 1 AI data center, which reportedly contains more than 220,000 Nvidia GPUs and over 300 megawatts of computing capacity.
That is huge infrastructure.
And it shows how expensive the AI race is becoming behind the scenes.
Companies are no longer competing only on models or features. They are competing on who controls the most computing power.
What makes this partnership surprising is the history behind it.
Elon Musk previously criticized Anthropic publicly, especially around AI safety debates. But according to reports, Musk recently changed tone after meetings with Anthropic staff and said he was impressed with the company’s focus on aligning AI with human values.
Now both sides are working together.
And the benefits are immediate.
Developers using Claude Code have been complaining for months about strict usage caps and interruptions during heavy workloads. Anthropic had even experimented with reducing access limits earlier this year because demand was outpacing available compute.
This deal helps relieve that pressure.
But it also reveals something bigger about the industry.
AI companies are becoming heavily dependent on giant infrastructure partnerships just to keep their services running smoothly. The smarter these systems become, the more expensive they are to operate.
And that means access to chips, data centers, and electricity is quietly becoming one of the most important competitive advantages in AI.
So the real question is not whether AI coding tools will keep improving.
It is whether only companies with massive infrastructure backing will be able to keep up with the growing demand.
