We’ve agreed to a partnership with SpaceX that may considerably improve our compute capability. This, together with our different latest compute offers, implies that we’ve been capable of improve our utilization limits for Claude Code and the Claude API.
Beneath, we describe these modifications and the progress we’re making on compute.
Greater utilization limits
The next three modifications—all efficient at present—are geared toward enhancing the expertise of utilizing Claude for our most devoted prospects.
First, we’re doubling Claude Code’s five-hour charge limits for Professional, Max, Crew, and seat-based Enterprise plans.
Second, we’re eradicating the height hours restrict discount on Claude Code for Professional and Max accounts.
Third, we’re elevating our API rate limits significantly for Claude Opus fashions, as proven within the desk under:
New compute partnership with SpaceX
We’ve signed an settlement with SpaceX to make use of the entire compute capability at their Colossus 1 knowledge heart. This offers us entry to greater than 300 megawatts of recent capability (over 220,000 NVIDIA GPUs) throughout the month. This extra capability will immediately enhance capability for Claude Professional and Claude Max subscribers.
This joins our different important compute bulletins:
- An up to 5 gigawatt (GW) agreement with Amazon, which incorporates almost 1 GW of recent capability by the top of 2026;
- A 5 GW agreement with Google and Broadcom, which can start coming on-line in 2027;
- A strategic partnership with Microsoft and NVIDIA that features $30 billion of Azure capability;
- Our $50 billion investment in American AI infrastructure with Fluidstack.
We practice and run Claude on a spread of AI {hardware}—AWS Trainium, Google TPUs, and NVIDIA GPUs—and proceed to discover alternatives to convey further capability on-line.
As a part of this settlement, we’ve additionally expressed curiosity in partnering with SpaceX to develop a number of gigawatts of orbital AI compute capability.
Increasing internationally
Our enterprise prospects—significantly these in regulated industries like monetary companies, healthcare, and authorities—more and more want in-region infrastructure to fulfill compliance and knowledge residency necessities. Accordingly, a few of our capability enlargement will likely be worldwide: our just lately introduced collaboration with Amazon contains further inference in Asia and Europe.
We’re very intentional about the place we’ll add capability—partnering with democratic nations whose authorized and regulatory frameworks help investments of this scale, and the place the provision chain on which our compute relies upon—{hardware}, networking, and amenities—will likely be safe.
Lastly, we just lately made a commitment to cowl any shopper electrical energy value will increase brought on by our knowledge facilities within the US. As a part of our worldwide enlargement, we’re exploring methods to increase that dedication to new jurisdictions, in addition to partnering with native leaders to take a position again into the communities that host our amenities.