Core Scientific to Deliver Additional 112 MW of HPC Infrastructure for CoreWeave

Core Scientific announced on August 6 an extension of its partnership with CoreWeave to deliver an additional 112 megawatts (MW) of high-performance computing (HPC) infrastructure.

According to a press release shared with Cryptonews.com, CoreWeave has exercised its option to expand its contract with Core Scientific by 112 MW, bringing the total HPC infrastructure to 382 MW to support CoreWeave’s NVIDIA GPUs .

CoreWeave Exercises Option to Expand Contract

The newly expanded contract with CoreWeave is anticipated to generate an additional $2.0 billion in cumulative revenue over the 12-year term. This addition brings the total projected revenue from CoreWeave’s contracts to $6.7 billion, including the previously announced $4.7 billion.

“We have now contracted with CoreWeave for a total of 382 megawatts of HPC infrastructure, reflecting the strong demand for high-power data center infrastructure and the unique ability of our team to deliver it,” said Core Scientific CEO Adam Sullivan .

“The latest contract also validates that our strategy for developing application-specific data centers aligns with the increasing energy density requirements for high-performance computing that legacy data centers do not typically satisfy,” stated Sullivan.

Core Scientific Wields 1.2 GW of Contracted Power

Based on the agreements announced on June 3 and June 25, 2024, CoreWeave will fund all capital investments needed to upgrade Core Scientific’s infrastructure for high-performance computing. The new agreement also includes options for two additional five-year renewal terms.

“CoreWeave has one remaining option for 118 megawatts of our infrastructure, and we are actively building our pipeline of potential new sites to expand our infrastructure portfolio and business opportunity,” added Sullivan.

The release demonstrated that Core Scientific, with 1.2 gigawatts of contracted power, can now provide nearly 500 MW of HPC hosting infrastructure. This capacity is intended for alternative compute workloads, strategically positioned based on proximity to major cities and fiber lines.