Technology thesis · Artificial Intelligence
high conviction growthAI weather and climate prediction
AI medium-range forecasting is settled: ECMWF runs AIFS operationally and its v2 now beats the physics-based IFS. The frontier has moved past parity to extremes, sub-seasonal skill and climate digital twins.
Position maintained continuously · last reviewed Jun 24, 2026
The thesis
State of the art (2026)
By mid-2026 AI medium-range forecasting is settled science, not a bet. ECMWF runs AIFS operationally — the deterministic AIFS Single since February 2025 and the AIFS ENS ensemble since July 2025, both upgraded to v2 on 12 May 2026 and now beating the physics-based IFS across most lead times and variables. In a telling consolidation, ECMWF stopped running external models (GraphCast, Pangu-Weather, Aurora, FourCastNet) the same day. Google DeepMind’s GenCast bettered ECMWF’s ENS on 97.2% of targets, and NVIDIA’s open Earth-2 stack pushes kilometre-scale downscaling via CorrDiff. The frontier has moved past parity to the hard problems: calibrated extremes, sub-seasonal skill and climate-scale digital twins.
Everything below is live inside CanaryIQ
The full analysis behind the verdict — the structure is real; the content unlocks when you log in.
Signal stack
Evidence stacked leading → lagging
Technology-native KPIs
Metrics that predict trajectory, tracked over time
Landscape map
Who builds what — and who depends on whom
Catalyst calendar
Dated events that will move the position
Technology roadmap
Milestones on the path to maturity
Watchlists
Companies, people and papers — each with a remove-by condition
Decision frameworks
The same call, framed for your desk
Thesis changelog
When our view changed, and why
Change our mind
2 disconfirming conditions
The rest is inside
You've read the verdict. The file is much deeper.
The full signal stack, technology-native KPIs tracked over time, the landscape of who depends on whom, the dated catalyst calendar, decision frameworks for every desk, live watchlists and the changelog of every time our call on AI weather and climate prediction has changed — all live inside CanaryIQ.