Supercomputer Unveiled for Wildfire Prediction
Technosylva has launched what it describes as the world’s largest supercomputer dedicated to wildfire modeling, designed to simulate more than one billion fire scenarios each day. The system applies artificial intelligence to decades of fire weather data and provides forecasts up to five days in advance, with results delivered in seconds. The tool is intended to give utilities and fire agencies additional lead time to protect communities, electric infrastructure, and other assets.
The system was developed in partnership with PSSC Labs and is built around 11,500 cores purpose-designed for wildfire behavior forecasting. It incorporates a wide range of fire-related data beyond standard weather and fuel models, including dead fuel moisture, vegetation conditions, terrain-driven winds, and 30 years of historical fire weather. Forecasts are generated at a two-kilometer resolution, updated every few hours, and processed in parallel using multiple fire behavior models, including proprietary fire spread, crown spotting, and fuel moisture simulations.
Technosylva said the expanded computing power allows utilities and agencies to make more precise public safety power shutoff decisions, accelerate evacuation alerts, and better anticipate wildfire risk. The company noted that the same infrastructure can also be used to support emergency planning for hurricanes, floods, and other extreme weather events.
Joaquin Ramirez, Technosylva’s chief technology officer and founder, described the platform as combining “high-performance computing with advanced AI, running ensemble forecasts against decades of meteorological and fire-behavior data.” Brent Shaw, senior numerical weather prediction architect at Technosylva, compared the results to moving “from an X-ray to a real-time MRI of wildfire behavior across the entire country.”
The new supercomputer also supports Technosylva’s research and development by enabling faster integration of artificial intelligence and machine learning into its wildfire forecasting products. According to Ramirez, the system provides the scale and speed needed to train and validate models with larger datasets and then translate those insights into operational tools.
PSSC Labs has worked with Technosylva since its early wildfire modeling projects and supported the transition to nationwide coverage. The new high-performance computing system is now operational and available for use by utilities and agencies across the United States.