Amazon’s project Rainier features world’s biggest AI supercluster in Indiana

Amazon has recently launched its $11 billion worth Rainier project on October 29, 2005 to power Anthropic’s Claude model by the end of the year.

The project features over 1 million Trainium 2 chips in the 1200 acre rural Indiana complex, converting it into the world’s biggest AI supercluster.

These cutting-edge chips will be used to train Anthropic’s AI model exclusively on Amazon chips.

According to the tech company, currently the project is running on 500,000 chips, but Anthropic has made plans to increase investment in Amazon’s AI infrastructure and cloud computing.

Given the capacity of chips, the project Rainier is considered as the world’s most powerful AI training compute cluster.

In addition to this, Amazon has invested $8 billion and Google poured around $3 billion into Anthropic.

According to Amazon Web Services (AWS) CEO Matt Garman, this project has emerged as an operational reality instead of a far-fetched concept.

“This is not some future project that we’ve talked about that maybe comes alive…This is running and training their models today,” Garman explained.

The facility, which is known for spanning multiple states and consuming 2.2 gigawatts of power, will offer 30-40 percent lower costs than Nvidia GPUs.

The $11 billion facility, spanning multiple states and consuming 2.2 gigawatts of power, offers 30-40% lower costs than Nvidia GPUs which the chipmaker planned to establish in a $100 billion deal with OpenAI.