AceCAST 5: what's new
AceCAST 5 is available now. Here is what changed: WRF 4.8, four physics updates and four new model outputs.
Wind ramps, cloud over solar, temperature-driven load and storm restoration are four different forecasting problems. What each one needs, what TempoQuest offers, and roughly what it costs.

Weather risk for energy and the grid comes down to four questions that an operations or trading desk answers every day. How hard will the wind blow at hub height? When will cloud cross the solar plant? How far will temperature push load? Where will storms put crews to work? This article takes them in turn, then sets out what TempoQuest offers for each and roughly what it costs.
Turbines sit 80 to 160 m up, while standard weather stations measure wind at 10 m. Below a turbine's rated speed, the power in the wind rises with the cube of its speed. A forecast that is 10% high on wind is therefore about 33% high on available power (1.1 × 1.1 × 1.1 = 1.331). The costly events are ramps, when output climbs or collapses within an hour as a front or a storm's cold outflow crosses the farm, and cut-outs, when turbines shut down in very strong wind.
Cloud causes most of the short-term swings in solar output. The questions are local and about timing: when morning fog or low cloud clears, when a cloud band arrives, how fast output falls. A coarse grid blurs those edges.
Heating and cooling drive demand, so peak load follows heat waves and cold snaps. Errors in the forecast high or low become errors in the load forecast, and they matter most on the peak days. Weather also limits the network itself: an overhead line can carry least current in hot, still air.
Thunderstorm wind, lightning, ice and hail bring down lines, trip equipment and damage panels. Restoration is faster when crews and materials are in place before the storm, and that takes a forecast of where and when, not only whether.
| Question | What TempoQuest offers |
|---|---|
| Hub-height wind and ramps | An Aerava wind forecast for a site, or your own AceCAST runs. |
| Cloud over solar | Your own AceCAST runs at the resolution you choose. |
| Temperature and load | Your own AceCAST runs for the forecast, and Custom Weather Data for the history a load model learns from. |
| Storms and crews | Aerava storm and hail forecasts, and its standing watch with email alerts. |
Aerava, formerly AceCAST Cloud, is the weather intelligence platform operated by TempoQuest. You pin a site and choose a forecast. Aerava runs a numerical weather model for that place and checks the result against the storm, wind and turbine cut-out limits you set. For US sites it offers:
Storm forecasts are not yet scored and hail-size output is experimental guidance. See how to score a forecast, including a bounded historical comparison with lidar observations. Site checks use the latest published forecast; expired forecasts are marked. StormScope coverage and update times.
Teams that run their own forecasts, on their own domain and schedule, can use AceCAST. It runs WRF, the Weather Research and Forecasting model maintained by NCAR, on NVIDIA GPUs, and it keeps the WRF workflow: WRF-compatible inputs, namelists and outputs, on Linux systems on premises or in the cloud.
See AceCAST 5, available now, for its physics, outputs and upgrade guidance.
For planning studies, load models and AI training, TempoQuest scopes, generates, packages and delivers project-specific high-resolution model outputs and derived datasets. Typical requests are around 1 km or 3 km, at 15-minute, 30-minute, hourly or daily cadence, including historical datasets, delivered as NetCDF, GRIB, Zarr, CSV, JSON or through an API. See what AI weather training data needs.
Aerava. Forecasts are priced in credits, and the price is shown before you run. There are monthly plans, credit packs and a 14-day free trial. See aerava.ai for current pricing.
AceCAST. Licensed by number of GPUs or metered by usage, with pricing on request. A trial runs for 14 days on one machine. You also pay for the GPUs: our WRF GPU vs CPU cost article works that sum from a published benchmark of earlier AceCAST releases.
Custom Weather Data is priced per project. The price follows the work: domain size, compute, history, cadence, validation and integration.

AceCAST 5 is available now. Here is what changed: WRF 4.8, four physics updates and four new model outputs.

Aerava runs a weather model around the site you pin and checks the result against your own storm, wind and turbine cut-out limits. Here is what you can run today, what it costs and what it says about accuracy.

One published 2023 benchmark and two Azure list prices, worked line by line: what a GPU VM costs against one CPU VM and against sixteen, and what the sums leave out.
Your forecast
Tell us the decision it has to serve. One short form reaches the team, and we review every request.