Weatherwise offers a cloud‑based, GPU‑accelerated API that delivers hyper‑local weather forecasts at 200 m resolution with up to 1,800 variables, refreshed in under an hour. The service lets energy traders, utilities, and renewable asset managers integrate near‑real‑time weather data directly into trading, grid‑management, and asset‑optimization workflows without maintaining on‑premise HPC.
Funding
Funding not disclosed
Founders
Product
Problem
Energy traders, utilities, and other asset operators need highly localized, up-to-date weather information to manage risk and optimize operations, but traditional forecasting services are coarse, updated infrequently, and require managing expensive high‑performance computing infrastructure.
Solution
Weatherwise provides a GPU‑accelerated, API‑driven platform that delivers hyper‑local weather forecasts at 200 m horizontal resolution with up to 1,800 vertical and surface variables. The service can generate short‑term, subseasonal, seasonal, and climate forecasts, refreshing data in under an hour to match the 15‑minute granularity required by modern power markets. Users submit location and resolution parameters via a simple REST API, and the platform runs physics‑based numerical weather prediction models on cloud GPUs, eliminating the need for on‑premise HPC. Forecast outputs are returned in standard formats for direct integration into trading, grid‑management, or asset‑optimization workflows.
Target Audience
Primary customers are energy traders, utility grid operators, and renewable asset managers who need high‑resolution, near‑real‑time weather data to reduce imbalance costs and improve grid resilience. Secondary markets include automotive manufacturers seeking road‑level friction data and precision agriculture firms requiring field‑scale forecasts.
Features
- 200 m × 200 m grid spacing enabling meter‑scale forecasts for individual assets
- Access to 1,800 atmospheric variables across vertical and surface levels
- Support for multiple forecast horizons: short‑term (hours), subseasonal (days), seasonal (months), and climate (decades)
- GPU‑based simulation engine that refreshes forecasts in under one hour, meeting 15‑minute market update cycles
- RESTful API with client libraries for Python, Node.js, JavaScript, cURL, and Bash for easy integration
- Ability to run both forward forecasts (GFS, IFS, HRRR) and hindcasts (ERA5, CERRA5) for historical analysis
- No on‑premise HPC required; all compute is managed in the cloud