Anthropic gains access to SpaceX supercomputer with 220,000 GPUs
EUR/MDL - 20.19 0.0885
USD/MDL - 17.16 0.5325
VMS_91 - 3.03%
VMS_364 - 9.54%
BONDS_2Y - 7.40%
GOLD - 4,715.03 0.01%
EURUSD - 1.18 0%
BRENT - 117.29 13.73%
SP500 - 737.62 0.83%
SILVER - 80.34 0%
GAS - 2.77 8.88%

Market shocked: Anthropic gained access to Musk’s 220k GPU supercomputer

AI company Anthropic has closed one of the largest infrastructure deals in the history of the industry, gaining access to the entire computing power of SpaceX's Colossus 1 data center. This includes more than 220,000 Nvidia GPUs and over 300 megawatts of computing power, a resource that was recently used to develop Ilon Musk's rival xAI project.
Дмитрий Калак Reading time: 2 minutes
Link copied
Anthropic

“On the morning of May 6, Anthropic’s Chief Product Officer Ami Vora took the stage at the Code with Claude conference in San Francisco and confirmed a deal that shouldn’t be,” Let’s Data Science writes in this regard.

“We are partnering with SpaceX to utilize the full capacity of their Colossus One data center,” Vora told the crowd. Colossus One is in Memphis, Tennessee. – It is a supercomputer built by Ilon Musk for Grok training. It contains more than 220,000 Nvidia GPUs, including the H100, H200 and the latest GB200 gas pedals. Within a month, Anthropic will have access to all of this capacity: roughly 300 megawatts of computing power that until recently powered a model designed to compete with Claude.”

A few hours later, Musk published his version of events at X. He wrote that he had spent time with Anthropic executives asking how the company was ensuring that Claude was in the best interest of humanity. He came away convinced that a deal was possible. “No one activated my evil detector,” he philosophically stated at the end of the meeting.

From competition to infrastructure alliance

The agreement between Anthropic and SpaceX has become one of the most discussed events of the AI market in recent days. This will allow the company to dramatically increase the capabilities of its Claude AI platform and remove restrictions for corporate users.

What makes the situation particularly intriguing is the fact that Colossus 1 was originally created for Ilon Musk’s xAI project, a direct competitor of Anthropic.

The deal looks especially unexpected against the background of previous criticism of Anthropic by Ilon Musk. The billionaire has previously publicly accused the company of creating “dangerous” AI and criticized its approaches to model safety.

The alliance between Anthropic and SpaceX shows that the main deficit of the AI industry is no longer models, but infrastructure – GPUs, data centers, energy and cooling systems.

Against this background, Anthropic has actually obtained one of the largest computing resources in the industry in just a few weeks, which could significantly change the balance of power in the generative AI market.



Реклама недоступна
Must Read*

We always appreciate your feedback!

Read also