- Google Cloud will serve as the primary high-performance computing provider
- NOAA is expecting the shift to ease infrastructure constraints
- The Weather and Climate Operational Supercomputing System currently runs on twin HPE Cray systems in Virginia and Arizona
The National Oceanic and Atmospheric Administration is shifting its Weather and Climate Operational Supercomputing System, or WCOSS, onto commercial cloud infrastructure. Google Cloud will act as the primary provider of high-performance computing for the program, NOAA said Monday.
Operations moving to the new environment include the Global Forecast System and the Global Ensemble Forecast System. That work runs through early 2027, with the full transition scheduled to finish by December 2027. NOAA divided the effort into two phases: standing up the supercomputing capability in the cloud, then migrating individual forecast models for operational use.
Data, cloud and compute infrastructure are among the topics on the agenda for the Potomac Officers Club’s 2026 FedCiv Summit on Oct. 29. Department of Transportation Chief Digital and Information Officer Pavan Pidugu, a Wash100 awardee, will deliver a keynote address. Register now to join the conversation!
What Will Change in NOAA’s Weather Forecast Modeling?
NOAA said the move removes infrastructure constraints that slow the pace of model revisions. The agency expects fewer and shorter outages, along with the ability to add computing capacity when demand spikes.
NOAA Administrator Neil Jacobs said cloud-based high-performance computing will speed the handoff of research into operations by “eliminating traditional bottlenecks of on-premise systems.” He added that the approach gives the agency access to newer chips.
The agency described itself as one of the first operational numerical weather prediction centers worldwide to move to secure commercial cloud infrastructure.
How Will Google Cloud Support NOAA’s Weather Models?
Google said its H4D virtual machines running fifth-generation AMD EPYC processors will serve as the primary computing backbone for NOAA’s workloads.
The two organizations are also collaborating on AI forecast models. NOAA is using Google DeepMind as the framework for the Artificial Intelligence Global Forecast System, which the agency described as its first suite of weather forecast models driven by machine learning.
In July 2025, Google and NOAA’s National Hurricane Center signed a cooperative research and development agreement to evaluate AI-based tropical cyclone forecast models. Google said NOAA later used its WeatherNext model to forecast Hurricane Melissa’s Category 5 landfall five days in advance.
What Is NOAA’s Existing WCOSS Supercomputing Setup?
WCOSS currently runs on twin Hewlett Packard Enterprise Cray systems: Dogwood in Manassas, Virginia, and Cactus in Phoenix, Arizona. Each operates at 14.5 petaflops following a 2023 capacity expansion.
General Dynamics Information Technology architected both machines under a potential 10-year, $505 million contract awarded in February 2020.
The cloud migration follows contracts NOAA announced earlier this year to build cloud-based technology for forecasters. The agency said the work covers 122 local weather forecast offices and other National Weather Service centers.
