NOAA Selects Google Cloud as HPC Backbone for Weather Supercomputing System

The U.S. National Oceanic and Atmospheric Administration (NOAA) has chosen Google Cloud as its primary high-performance computing provider for the Weather and Climate Operational Supercomputing System (WCOSS), marking a significant step in the agency’s transition from traditional on-premises infrastructure to a cloud-first approach. The move is expected to accelerate numerical weather prediction and enable more agile, scalable modeling capabilities.

Under the new arrangement, Google Cloud’s H4D virtual machines powered by fifth-generation AMD Epyc processors will serve as the core computing backbone for NOAA, effectively functioning as a cloud-based supercomputer. This replaces the agency’s previous reliance on dedicated on-premises HPC systems, which were regularly upgraded but constrained by fixed capacity and long procurement cycles.

NOAA and Google have a long-standing partnership that previously included sharing environmental data, real-time wildfire tracking, and marine conservation efforts. In a notable earlier collaboration, NOAA used Google DeepMind and the company’s WeatherNext model to predict the landfall of Hurricane Melissa five days before it arrived, demonstrating the potential of cloud-based AI in operational forecasting.

“Cloud-based high-performance computing will accelerate the transition of research into operations by eliminating traditional bottlenecks of on-premise systems,” said Neil Jacobs, Ph.D., NOAA administrator. “Not only will this new approach provide access to the latest chips, it will advance weather forecasting by aligning operational and research compute and provide capabilities for elastic on-demand model development.”

Ken Graham, director of NOAA’s National Weather Service, added: “We are charting a course to global leadership in cloud-based numerical weather prediction. This is a significant achievement for NOAA; one that will ensure our critical modeling operations can seamlessly adapt to emerging science, and strengthen our mission to protect life and property.”

WCOSS encompasses the supercomputing, storage, wide-area networking, and developmental computing resources that NOAA uses to run daily weather, ocean, and climate forecast models. The agency last upgraded its on-premises WCOSS in summer 2023, deploying twin supercomputers each with 14.5 petaflops of computing capacity — a 20 percent increase over the prior system. In September 2024, NOAA awarded a $100 million grant from the Bipartisan Infrastructure Law and Inflation Reduction Act to acquire a GPU-equipped supercomputer named Rhea, designed to strengthen the agency’s exploration of artificial intelligence and machine learning. More recently, in March 2026, NOAA awarded two contracts to move National Weather Service data and application services to the cloud as part of a broader modernization initiative.

NOAA is not the first major meteorological service to embrace cloud-based HPC. The UK’s Met Office signed an agreement with Microsoft in 2021, resulting in a dedicated supercomputer housed in a specially designed facility within a Microsoft cloud data center in the UK. That system officially entered service in May 2025. Unlike a fully public cloud deployment, Microsoft hosts the Met Office’s multiple supercomputers in dedicated halls with custom power, cooling, and networking configurations optimized for energy efficiency and operational resilience. The solution also offers access to Microsoft’s broader cloud storage services.

The NOAA-Google Cloud partnership signals a broader industry shift toward cloud-based numerical weather prediction, where elastic compute resources can support on-demand model development and faster integration of cutting-edge chips and AI capabilities. By aligning operational and research computing on a common cloud platform, NOAA aims to shorten the cycle from scientific discovery to operational forecast, ultimately improving the accuracy and timeliness of weather warnings that protect life and property.