Microsoft's ML System Forecasts Grid Risks from Space Weather for 66,935 US Substations
Microsoft Research · rss · 2026-10-01
A Microsoft Research ML pipeline predicts geomagnetically induced current risk for all 66,935 US substations, detecting 80% of major space-weather events with 30-60 minutes of advance warning. It combines AE/Dst index forecasts, local geology and grid data; full inference takes 333ms.
More from Infra
- netkit paper: container network namespaces cost up to 31% throughput on Linux — tianyin_xu · 2026-10-01
- Dev sorts PCIe issues, spins up 7-GPU local rig with 6000 Pro and 6000 ADA cards — TheZachMueller · 2026-10-01
- Ben Bajarin: HPE and others use 'scale-out' vs 'scale-up' in conflicting ways — BenBajarin · 2026-10-01
- NVIDIA and Nous Research detail agent tracing with NeMo Relay across 108-run eval — NVIDIAAI · 2026-10-01
- stuntd 0.1.2: local heads answer confident LLM calls, weakest field caps the hit rate — Inevitable-Log5414 · 2026-10-01
- Framework Desktop with 192GB RAM opens pre-orders at $6,799, ships November — gnukeith · 2026-10-01