Core42 announced a 42MW expansion of its U.S. AI infrastructure at the Lake Mariner site in New York, increasing total site capacity from 18MW to 60MW of high-performance AI production infrastructure.
The expansion reflects G42’s continued capital investment in scaling AI infrastructure across the United States.
Lake Mariner serves as a North American AI infrastructure hub within Core42’s globally distributed network. The facility previously demonstrated architectural performance with the AMD Instinct MI300-based Maximus cluster securing a Top-20 ranking on the global TOP500 supercomputing list. The integration of additional AMD and NVIDIA infrastructure strengthens the site’s heterogeneous design, enabling workload optimisation across multiple accelerator platforms.
Beyond Lake Mariner in Buffalo, New York, Core42’s U.S. footprint includes deployments in Dallas, Texas, Sunnyvale and Stockton, California, and Minneapolis, Minnesota, including the Condor Galaxy supercomputers, developed in collaboration with Cerebras. Together, these sites extend Core42’s heterogeneous AI architecture across the United States, enabling workload-optimised deployment for both frontier training and high-speed inference at production scale.
“We are scaling our U.S. infrastructure in line with long-term deployment programmes,” said Talal M. Al Kaissi, Chief Executive Officer, Core42. “Increasing our U.S. capacity at Lake Mariner strengthens our ability to serve hyperscale, AI-native and large enterprise workloads, and further extends the build out of our AI infrastructure globally.”
The Buffalo expansion forms part of Core42’s broader international infrastructure buildout, spanning the United States, Europe and the Middle East. In 2025, the company established its European headquarters in Dublin and expanded AI compute deployments across Italy and France, alongside continued U.S. capacity growth.
Core42’s AI Cloud platform, which was introduced in October 2025, is the access layer to the distributed AI infrastructure, allowing customers to provision compute across jurisdictions under a consistent operating model. The platform supports the full AI lifecycle, from large-scale training and fine-tuning to real-time inference, leveraging heterogeneous infrastructure with multiple options to help drive optimal price-performance. With 10 operational sites globally and additional deployments planned for 2026, Core42 continues to scale the infrastructure foundation required to operate AI at national and enterprise scale.
Discussion about this post