原文 · 未翻译
Anthropic taps SpaceX's Colossus-1 data center for 220,000 GPUs to power Claude
Anthropic has partnered with SpaceX to take over the full computing capacity of SpaceX's Colossus 1 data center.
That's more than 300 megawatts and over 220,000 NVIDIA GPUs, expected to come online within a month. As part of the deal, Anthropic is bumping up usage limits across the board for Claude: five-hour rate limits for Claude Code will double within a month for Pro, Max, Team, and Enterprise customers, peak-time throttling for Pro and Max accounts goes away entirely, and API limits for Claude Opus models are getting a major increase.
New API rate limits for Claude Opus
Tier Maximum input tokens per minute Maximum output tokens per minute 1 30,000 → 500,000 8,000 → 80,000 2 450,000 → 2,000,000 90,000 → 200,000 3 800,000 → 5,000,000 160,000 → 400,000 4 2,000,000 → 10,000,000 400,000 → 800,000
The SpaceX agreement sits alongside a string of existing compute deals: up to 5 GW with Amazon, 5 GW with Google and Broadcom, $30 billion in Azure capacity with Microsoft and NVIDIA, and a $50 billion investment with Fluidstack.
The two companies are also exploring what they call "orbital AI compute capacity." Separately, Anthropic has committed to covering any electricity price increases for US consumers caused by its data centers and says it's looking into extending that commitment to future locations.
Anthropic is using the announcement to stake out a political position as well: the company says it will only partner with "democratic countries whose legal and regulatory frameworks support investments of this scale, and where the supply chain on which our compute depends—hardware, networking, and facilities—will be secure." That makes this particular deal with SpaceX, a company owned by Elon Musk whose political views are increasingly considered authoritarian, a notable contradiction.
AI News Without the Hype – Curated by Humans