Urgent.News

What's breaking now, across thousands of outlets.

AI

Anthropic agrees to $45bn cloud computing deal with Nscale - report

Anthropic PBC has struck a $45 billion agreement to lease AI cloud computing power from Nscale's development in West Virginia, according to Bloomberg. This follows the AI company's efforts to secure capacity ahead of its anticipated initial public offering. Nscale will utilize Nvidia Corp.'s Vera Rubin chips to fulfill Anthropic's requirements, with the chips set to commence availability late next year.

The commitment extends over six years and encompasses approximately 460 megawatts of power, capable of supplying around 345,000 US homes simultaneously. Anthropic, the creator of the Claude AI assistant, has recently inked several computing deals. These include a $50 billion arrangement with Fluidstack, a $10 billion pact with Volta Infra Holdings Ltd., and a $45 billion agreement with SpaceX. In May, Anthropic was valued at $965 billion.

The deal pertains to the first of three structures at Nscale's Monarch campus in West Virginia. Upon completion, the entire campus will generate about 1.35 gigawatts of power. The project's total cost will amount to roughly $71 billion, with $47 billion allocated towards AI chips. Servers in the subsequent buildings are projected to begin delivering power in 2028.

Microsoft Corp. had previously signed a letter of intent for the West Virginia site in March, but withdrew earlier in the summer, allowing Anthropic to assume the tenant role. The agreement holds significant importance for Nscale, as it anticipates an initial public offering later this month. Nscale is part of a group of firms known as neoclouds, which lease AI-focused computing power.

Written by urgent.news from Investing.com's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at investing.com →

More in AI

Every Layer of Your AI Stack Is an Attack Vector. Count Them.

From the team at Backboard.io . We build AI infrastructure, so we have a position here. We state it at the end, clearly labeled. Everything before that is just counting.

  • AI stack comprises 6-9 vendors, each adding 5+ attack surface factors
  • Attack surface includes model, data, and integration layers
  • Vector databases often overlooked, but vulnerable to embedding inversion attacks

More from Wednesday 26 August →