Core Scientific, Inc. has expanded its digital infrastructure agreement with CoreWeave, allocating additional infrastructure for high-performance computing (HPC) operations under a 200-megawatt, 12-year contract initiated on June 3, 2024.
As per the announcement, Core Scientific will adapt 100 MW of its infrastructure to provide about 70 MW for hosting CoreWeave’s NVIDIA GPUs for HPC operations. Site renovations are scheduled to commence in the second half of 2024, with operations set to begin in the second half of 2025.
This new 12-year HPC hosting agreement increases Core Scientific’s exposure to long-term, dollar-denominated revenue. This extension will raise Core Scientific’s committed, multi-year, dollar-denominated revenue by an estimated $1.225 billion over 12 years, augmenting the $3.5 billion from current CoreWeave agreements.
Core Scientific’s Chief Executive Officer, Adam Sullivan, said, “We are excited to build on our momentum and expand the scope of our HPC hosting business with significant additional infrastructure.”
Sullivan emphasizes that Core Scientific is poised to design and construct application-specific data centers (ASDCs) for high rack energy densities across expansive footprints. These ASDCs will support clients like CoreWeave deploying advanced chips at scale, leveraging their flexible, high-power infrastructure and expert team for efficient execution.
Sullivan emphasizes that Core Scientific will accelerate its hosting business by delivering projects early, such as the Austin data center for CoreWeave. They aim to bolster AI computing capabilities and maximize opportunities in bitcoin mining amid industry transformations.
The agreement states that CoreWeave will finance all capital investments to modernize Core Scientific’s dense HPC infrastructure, totaling about $105 million. These charges will be deducted from hosting fees, with a ceiling of 50% of monthly payments until fully reimbursed. The deal also allows for two five-year renewal terms.
Core Scientific intends to extend its HPC infrastructure for CoreWeave to 270 MW by late 2025, adding to its current 200 MW. CoreWeave also has the option to add 230 MW to other Core Scientific locations, possibly making Core Scientific one of the largest data center operators in the United States.
Overall, Core Scientific’s contracted power of 1.2 gigawatts enables it to deliver nearly 500 MW for HPC hosting, catering to diverse compute workloads based on geographic proximity.
Also Read: Core Scientific Declines CoreWeave’s $1 Billion Buyout Offer
