Allen Institute Makes Global Weather Prediction Cheap Enough to Ignore

The claim: The Allen Institute for AI has released HiRO-ACEv2, a high-resolution weather forecasting model designed to simulate extreme weather events globally at a 3 km grid spacing while significantly reducing computational costs compared to traditional numerical weather prediction systems.

The reality: Running a global 3 km model still demands enterprise cluster hardware, meaning standard meteorology departments will continue renting cloud nodes from AWS while translating the same old atmospheric equations.

Why should I care? Could be big
Interesting open-weights atmospheric science, but local forecasters still lack the hardware budgets to run it.

Read the original: HiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost