AI Forecasts

Introduction

AI weather forecasting has become one of the clearest examples of scientific acceleration in practice rather than theory. Systems such as GraphCast, GenCast and ECMWF’s Artificial Intelligence Forecasting System can now generate global forecasts in minutes instead of hours while matching or outperforming leading traditional models on many measures. That matters not only because better forecasts help people decide whether to carry an umbrella. Weather prediction is one of the hardest scientific problems routinely solved at planetary scale: a constantly changing physical system with enormous volumes of data, chaotic behaviour and high real-world stakes.

AI Forecasts illustration 1 If AI can help compress that loop between observation, prediction, testing and revision in weather science, it offers a preview of something larger in the AI bloom argument. It suggests that machine learning may become a general tool for turning vast scientific datasets into usable models more quickly, cheaply and iteratively than previous approaches allowed. Weather forecasting is therefore important not only for meteorology, but as evidence that AI may accelerate how civilisation learns about complex systems. [Google DeepMind]youtube.comGoogle Deep Mindraeng.org.uk… [Nature]nature.comNatureProbabilistic weather forecasting with machine learningby I Price · 2025 · Cited by 647 — GenCast generates an ensemble of stochast…

Forecasting as a testable scientific loop

Weather forecasting is unusually valuable as a proving ground for AI because it creates constant, measurable feedback. Every prediction is tested against reality within hours or days. Models cannot hide behind vague claims or cherry-picked examples. Either the hurricane turned north or it did not. Either the rainfall arrived or it did not.

That makes forecasting one of the cleanest scientific environments for evaluating whether AI systems genuinely improve predictive capability. Unlike many AI demonstrations, weather prediction has objective scoring systems, decades of historical data and operational benchmarks maintained by organisations such as the European Centre for Medium-Range Weather Forecasts (ECMWF).

Traditional numerical weather prediction works by solving physical equations describing atmospheric motion on giant supercomputers. These systems remain scientifically foundational and still provide many of the observations and simulations AI models learn from. But they are computationally expensive. Running global forecasts at high resolution can require enormous energy and computing infrastructure.

AI weather systems changed the economics of iteration. GraphCast, developed by Google DeepMind, demonstrated global 10-day forecasts in under a minute while outperforming a leading operational system on most tested targets. [Google DeepMind]youtube.comGoogle Deep Mindraeng.org.uk… GenCast later extended this into probabilistic ensemble forecasting, generating multiple possible future weather trajectories in around eight minutes. [Google DeepMind]youtube.comGoogle Deep Mindraeng.org.uk…

The broader scientific importance is the feedback loop this enables:

more forecasts generated more cheaply [ecmwf.int]ecmwf.ints ai forecasts become operationalECMWF's AI forecasts become operationalFeb 25, 2025 — The AIFS is the first fully operational weather prediction open model using machine…
faster comparison between competing models
quicker adaptation after failures
easier experimentation with new architectures
wider access to advanced forecasting capability

That resembles a broader pattern emerging across AI-assisted science: lower-cost hypothesis generation combined with faster validation cycles.

What machine-learning weather models changed

The most important shift is not that AI “replaced physics”. In practice, the strongest systems combine machine learning with decades of physically informed observational data and traditional forecasting infrastructure.

The real change is that AI models learned to emulate the behaviour of enormously expensive simulations at much lower computational cost. Instead of explicitly calculating every atmospheric interaction from first principles each time, machine-learning systems infer patterns from historical atmospheric states and their evolution over time.

GraphCast used graph neural networks trained on ERA5 reanalysis data — a vast reconstruction of historical global weather observations — to predict hundreds of atmospheric variables globally. Researchers reported that it outperformed the ECMWF High Resolution Forecast on around 90% of evaluated targets. [arXiv]arxiv.orgarXivGraphCast: Learning skillful medium-range global weather…by R Lam · 2022 · Cited by 2370 — We show that GraphCast significantly o…

GenCast pushed further by focusing on uncertainty itself. Weather forecasting is inherently probabilistic because small differences in initial conditions can produce very different outcomes days later. Traditional forecasting centres therefore generate ensembles: many slightly different simulations representing possible futures.

GenCast applied diffusion-model techniques to produce probabilistic ensembles directly. In Nature, researchers reported greater skill than ECMWF’s ENS ensemble system on more than 97% of evaluated targets, including tropical cyclone tracks and wind-power prediction. [Nature]nature.comDeepMind AI accurately forecasts weather — on a desktop…Nov 14, 2023 — The machine-learning model takes less than a minute to predict…

This matters for scientific acceleration because it changes what becomes economically feasible. Faster and cheaper forecasting means:

more scenarios can be tested
uncertainty can be explored more deeply
forecasts can be updated more frequently
smaller institutions can access sophisticated prediction systems
scientists can experiment with coupled Earth-system modelling more rapidly

In other words, AI forecasting increases the effective supply of predictive computation.

Why weather matters beyond weather

Weather prediction sits at the intersection of many other systems civilisation depends on: food, electricity, logistics, insurance, shipping, water management and disaster response.

That makes forecasting a useful example of how scientific acceleration can propagate through the wider economy.

Energy systems become easier to coordinate

Modern electricity grids increasingly depend on weather-sensitive renewable energy. Wind output, solar generation and electricity demand all fluctuate with atmospheric conditions.

More accurate forecasts improve the ability to balance supply and demand ahead of time. GenCast researchers specifically highlighted gains in predicting wind-power production. [Nature]nature.com50 years of weather forecasting at the ECMWFby F Rabier · 2025 — A newly operational model, known as the Artificial Intelligence Forecast…

This may sound narrow, but the implications compound. Better forecasting can reduce wasted reserve capacity, lower balancing costs and make high-renewable grids easier to operate. In the long-term AI bloom frame, that matters because abundant clean energy is one of the main prerequisites for broader material abundance.

Extreme-weather response improves

The most valuable forecast is often not the average one but the rare-event warning.

GraphCast and GenCast both showed strong performance on tropical cyclones and extreme-weather events. [arXiv]arxiv.orgarXivGraphCast: Learning skillful medium-range global weather…by R Lam · 2022 · Cited by 2370 — We show that GraphCast significantly o… Faster forecasting also allows agencies to run many more scenarios around uncertain storm tracks or flood risks.

Even modest gains in warning time can matter enormously:

evacuations can begin earlier
ports and transport systems can shut down safely
hospitals can prepare for heatwaves
farmers can protect crops
emergency services can pre-position resources

The broader scientific lesson is that AI systems become more socially valuable when they model uncertainty rather than simply outputting one “best guess”.

AI Forecasts illustration 2

Planning systems gain richer information

[Weather forecasting is increasingly becoming infrastructure for other AI systems.]articsledge.comai weather forecasting2026: Models, Accuracy & Results3 days ago — Google DeepMind's GraphCast outperformed ECMWF's flagship HRES model on 90% of 1,380 verific…

Agricultural planning, shipping logistics, insurance pricing, flood management and autonomous systems all depend on environmental prediction. Faster forecasting therefore increases the responsiveness of many downstream systems at once.

That is part of why weather forecasting matters for scientific acceleration beyond meteorology itself. It demonstrates how AI prediction tools can become general coordination infrastructure for a more complex civilisation.

Why meteorologists did not simply abandon physics

One common misunderstanding is that AI weather systems made traditional meteorology obsolete overnight. In reality, operational forecasting centres have generally adopted hybrid approaches.

ECMWF’s Artificial Intelligence Forecasting System runs alongside conventional physics-based systems rather than replacing them entirely. [ECMWF]ecmwf.ints ai forecasts become operationalECMWF's AI forecasts become operationalFeb 25, 2025 — The AIFS is the first fully operational weather prediction open model using machine… Forecasting agencies still rely heavily on observational networks, physical understanding and numerical models.

There are several reasons for this caution.

First, machine-learning systems can inherit blind spots from training data. Climate shifts or rare atmospheric states may differ from historical patterns.

Second, AI models can sometimes produce physically inconsistent outputs unless carefully constrained. ECMWF researchers have explicitly worked on adding physical consistency mechanisms into newer AIFS versions. [arXiv]arxiv.orgarXivGraphCast: Learning skillful medium-range global weather…by R Lam · 2022 · Cited by 2370 — We show that GraphCast significantly o…

Third, operational forecasting requires trust, interpretability and resilience. Governments making evacuation decisions need systems that fail predictably and can be audited.

This tension is important for the wider AI bloom debate. Weather forecasting suggests AI can dramatically accelerate prediction, but it also shows that scientific reliability still depends on validation, institutional expertise and physical grounding.

The likely future is not “AI replaces science”, but increasingly automated scientific systems embedded inside human institutions and experimental feedback loops.

Weather forecasting as a model for machine-speed science

[The deeper significance of AI weather forecasting may be methodological.]articsledge.comai weather forecasting2026: Models, Accuracy & Results3 days ago — Google DeepMind's GraphCast outperformed ECMWF's flagship HRES model on 90% of 1,380 verific…

Forecasting compresses many ingredients now appearing across AI-enabled science:

Scientific ingredientWeather forecasting exampleVast historical datasetsDecades of atmospheric observationsComplex interacting systemsGlobal atmosphere and oceansExpensive simulationsNumerical weather predictionContinuous feedbackForecasts checked dailyHigh economic stakesEnergy, agriculture, disastersProbabilistic reasoningEnsemble forecastingHuman-machine collaborationMeteorologists interpreting outputs

That combination makes weather science an unusually clear demonstration of how AI may accelerate discovery elsewhere.

Researchers are already extending these methods toward coupled Earth-system models integrating atmosphere, oceans, waves and sea ice. ECMWF researchers reported that newer AIFS work improved many marine forecast variables by roughly a day at medium-range lead times. [arXiv]arxiv.orgarXivGraphCast: Learning skillful medium-range global weather…by R Lam · 2022 · Cited by 2370 — We show that GraphCast significantly o…

The broader implication is that machine learning may increasingly help scientists model systems too large, fast or interconnected for purely hand-crafted approaches.

That possibility extends beyond weather:

climate modelling
materials science
epidemiology
plasma physics
ecosystems
fusion research
biological systems

In each case, the key promise is similar: faster iteration between data, prediction and testing.

AI Forecasts illustration 3

Where speed, uncertainty and trust matter most

AI forecasting also exposes some of the hardest questions in the optimistic case for scientific acceleration.

Faster prediction is not automatically better governance

Better forecasts do not guarantee wiser decisions. Governments may ignore warnings. Infrastructure may remain underfunded. Insurance systems may fail to adapt. Forecasting capacity can still be distributed unevenly between countries and institutions.

Scientific acceleration only becomes broad human flourishing if societies can translate prediction into coordination.

AI models still depend on shared scientific infrastructure

Machine-learning weather systems rely heavily on decades of publicly funded observation networks, satellites, meteorological archives and international scientific collaboration.

That matters politically. If advanced forecasting becomes dominated by a few private firms controlling critical infrastructure, some benefits of scientific acceleration could become concentrated rather than broadly shared.

Climate change creates moving targets

Historical weather data may become less reliable as climate patterns shift. Extreme events outside previous training distributions could challenge AI systems.

For that reason, many researchers argue that physics-based models remain essential anchors even as AI forecasting improves. The strongest systems may combine physical simulation, observational science and machine learning rather than treating them as rivals.

A preview of broader scientific acceleration

AI weather forecasting matters because it demonstrates something civilisation rarely sees clearly: a major scientific capability improving rapidly in both speed and usefulness at the same time.

Forecasting once demanded enormous specialised computing resources and long runtimes. Increasingly, advanced models can produce competitive global predictions on dramatically smaller computational budgets. [Google DeepMind]youtube.comGoogle Deep Mindraeng.org.uk… [Nature That does not prove an intelligence explosion is inevitable. It does not guarantee post-scarcity abundance or fully automated science. But it]nature.comNatureProbabilistic weather forecasting with machine learningby I Price · 2025 · Cited by 647 — GenCast generates an ensemble of stochast… does provide concrete evidence for a narrower and more defensible claim at the centre of the AI bloom thesis: machine learning can sometimes compress the cycle between data, prediction and practical action across extremely complex domains.

Weather forecasting matters because it is not a toy problem. It is planetary-scale science under constant real-world evaluation. When AI systems improve there, even cautiously and imperfectly, it suggests that scientific discovery itself may become faster, more iterative and more widely deployable over time.

That possibility — civilisation learning faster about the systems it depends on — may ultimately be one of the most important pathways through which AI reshapes humanity’s long future.

Endnotes

Source: deepmind.google
Title: graphcast ai model for faster and more accurate global weather forecasting
Link: https://deepmind.google/blog/graphcast-ai-model-for-faster-and-more-accurate-global-weather-forecasting/
Source snippet
Google DeepMindGraphCast: AI model for faster and more accurate global...Nov 14, 2023 — Our state-of-the-art model delivers 10-day weath...
Source: nature.com
Link: https://www.nature.com/articles/s41586-024-08252-9
Source snippet
NatureProbabilistic weather forecasting with machine learningby I Price · 2025 · Cited by 647 — GenCast generates an ensemble of stochast...
Source: ecmwf.int
Title: s ai forecasts become operational
Link: https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ai-forecasts-become-operational
Source snippet
ECMWF's AI forecasts become operationalFeb 25, 2025 — The AIFS is the first fully operational weather prediction open model using machine...
Source: arxiv.org
Link: https://arxiv.org/abs/2212.12794
Source snippet
arXivGraphCast: Learning skillful medium-range global weather...by R Lam · 2022 · Cited by 2370 — We show that GraphCast significantly o...
Source: deepmind.google
Title: gencast predicts weather and the risks of extreme conditions with sota accuracy
Link: https://deepmind.google/blog/gencast-predicts-weather-and-the-risks-of-extreme-conditions-with-sota-accuracy/
Source snippet
Google DeepMindGenCast predicts weather and the risks of extreme...Dec 4, 2024 — New AI model advances the prediction of weather uncerta...
Source: arxiv.org
Title: arXiv Gen Cast: Diffusion-based ensemble forecasting for medium-range weather
Link: https://arxiv.org/abs/2312.15796
Source snippet
GenCast: Diffusion-based ensemble forecasting for...by I Price · 2023 · Cited by 197 — We introduce GenCast, a probabilistic weather mod...
Source: ecmwf.int
Title: s ensemble ai forecasts become operational
Link: https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ensemble-ai-forecasts-become-operational
Source snippet
ECMWFECMWF's ensemble AI forecasts become operationalJul 1, 2025 — ECMWF has taken the ensemble version of the Artificial Intelligence Fo...
Source: ecmwf.int
Link: https://www.ecmwf.int/en/forecasts/documentation-and-support
Source snippet
About our forecastsEnsemble forecasts looking from days to seasons ahead · Our Integrated Forecasting System (IFS) · Artificial Intellige...
Source: arxiv.org
Link: https://arxiv.org/abs/2509.18994
Source snippet
arXivAn update to ECMWF's machine-learned weather forecast model AIFSSeptember 23, 2025...

Published: September 23, 2025
Source: ecmwf.int
Link: https://www.ecmwf.int/en/about/media-centre/news/2026/ifs-cycle-50r1-aifsv2-live
Source snippet
generation of AI forecasting into operation, building on ECMWF's technical...Read more...
Source: arxiv.org
Link: https://arxiv.org/abs/2604.25559
Source snippet
arXivRepresenting the Surface Ocean in ECMWF's data-driven forecasting system AIFSApril 28, 2026...

Published: April 28, 2026
Source: ecmwf.int
Title: simplifying ai weather forecasting european weather cloud
Link: https://www.ecmwf.int/en/about/media-centre/science-blog/2025/simplifying-ai-weather-forecasting-european-weather-cloud
Source snippet
Simplifying AI for weather forecasting with the European...24 Nov 2025 — At ECMWF, this shift is embodied by the Artificial Intelligence...
Source: deepmind.google
Link: https://deepmind.google/
Source snippet
Google DeepMindArtificial intelligence could be one of humanity's most useful inventions. We research and build safe artificial intellige...
Source: deepmind.google
Link: https://deepmind.google/science/weathernext/
Source snippet
WeatherNext 2WeatherNext 2 is eight times faster. That means it can analyze a higher number of possible scenarios per forecast – making i...
Source: deepmind.google
Link: https://deepmind.google/research/publications/22598/
Source snippet
GraphCast: Learned Global Weather Forecasting14 Nov 2023 — GraphCast is orders of magnitude faster than ECMWF's operational systems, and...
Source: ecmwf.int
Link: https://www.ecmwf.int/en/newsletter/178/news/aifs-new-ecmwf-forecasting-system
Source snippet
AIFS: a new ECMWF forecasting systemThe first incarnation of the AIFS shows very promising results, replicating the rapid progress that h...
Source: ecmwf.int
Link: https://www.ecmwf.int/en/about/media-centre/aifs-blog/2026/farewell-external-ai-models
Source snippet
ompting a discontinuation of their real-time use...
Source: ecmwf.int
Title: AIF S forecast
Link: https://www.ecmwf.int/en/forecasts/datasets/set-ix
Source snippet
AIFS forecast - AIFS SingleSingle forecast runs produced by the ECMWF Artificial Intelligence Forecasting System (AIFS) deterministic mod...
Source: ecmwf.int
Title: aifs machine learning data
Link: https://www.ecmwf.int/en/forecasts/dataset/aifs-machine-learning-data
Source snippet
Jan 1, 2023 — ECMWF is now running its own Artificial Intelligence Forecasting System (AIFS). The AIFS consists of a deterministic model...
Source: ecmwf.int
Link: https://www.ecmwf.int/en/forecasts/datasets/open-data
Source snippet
Open dataA subset of ECMWF real-time forecast data from the IFS and AIFS models is made available to the public free of charge. Their use...
Source: ecmwf.int
Link: https://www.ecmwf.int/
Source snippet
We are both a research institute and a 24/7 operational service, producing global numerical...
Source: ecmwf.int
Link: https://www.ecmwf.int/en/forecasts/documentation-and-support/changes-ecmwf-model
Source snippet
, followed by an ensemble version in July. Evolution...Read more...
Source: aiweatherquest.ecmwf.int
Title: int A I Weather Quest • Overview
Link: https://aiweatherquest.ecmwf.int/
Source snippet
Weather Quest • Overview - ECMWFThe AI Weather Quest serves as a unique platform to benchmark state-of-the-art AI/ML-based sub-seasonal p...
Source: nature.com
Link: https://www.nature.com/articles/d41586-023-03552-y
Source snippet
DeepMind AI accurately forecasts weather — on a desktop...Nov 14, 2023 — The machine-learning model takes less than a minute to predict...
Source: nature.com
Link: https://www.nature.com/articles/s41467-025-65837-2
Source snippet
50 years of weather forecasting at the ECMWFby F Rabier · 2025 — A newly operational model, known as the Artificial Intelligence Forecast...
Source: nature.com
Link: https://www.nature.com/articles/d41586-024-03957-3
Source snippet
DeepMind AI weather forecaster beats world-class systemby A Soliman · 2024 · Cited by 7 — Artificial-intelligence model provides forecast...
Source: arxiv.org
Title: Report issue for preceding element. 1
Link: https://arxiv.org/html/2509.18994v1
Source snippet
An update to ECMWF's machine-learned weather forecast...Sep 23, 2025 — The AIFS has been fully operational at ECMWF since the 25th of Fe...
Source: medium.com
Link: https://medium.com/%40malintha1996/how-ai-is-revolutionizing-weather-forecasting-cf69a70dbb67
Source snippet
Google DeepMind's GraphCast for Weather ForecastingGraphCast is a cutting-edge AI model that delivers the world's most accurate 10-day gl...
Source: weather.us
Title: ECMW F AIFS
Link: https://weather.us/model-charts/euro-ai
Source snippet
ECMWF AIFS - Model charts for USA (Temperature)ECMWF AIFS - Current model charts of parameter "Temperature" for map "USA"...
Source: youtube.com
Title: AI and the Future of Weather Forecasting
Link: https://www.youtube.com/watch?v=I794OrppOVc
Source snippet
Google DeepMind - winner of the 2024 MacRobert Award...
Source: youtube.com
Title: Google Deep Mind
Link: https://www.youtube.com/watch?v=fdgA1Cyrk0g
Source snippet
raeng.org.uk...
Source: github.com
Link: https://github.com/google-deepmind/graphcast
Source snippet
Google DeepMind GraphCast and GenCastGraphCast: Learning skillful medium-range global weather forecasting. This package provides three pr...
Source: Wikipedia
Title: Google Deep Mind
Link: https://en.wikipedia.org/wiki/Google_DeepMind
Source snippet
Google DeepMindFounded in the UK in 2010, it was acquired by Google in 2014 and merged with Google AI's Google Brain division to becom...
Source: theguardian.com
Title: google deepmind predicts weather more accurately than leading system
Link: https://www.theguardian.com/science/2024/dec/04/google-deepmind-predicts-weather-more-accurately-than-leading-system
Source snippet
Google DeepMind predicts weather more accurately than...4 Dec 2024 — AI program GenCast performed better than ENS forecast at predicting...
Source: linkedin.com
Title: google deepmind touts ai weatherman 5819812
Link: https://www.linkedin.com/news/story/google-deepmind-touts-ai-weatherman-5819812/
Source snippet
Google DeepMind touts AI weatherman15 Nov 2023 — Scientists at Google DeepMind have developed an AI system that uses decades of historica...
Source: sbs.edu
Link: https://www.sbs.edu/foresight-news/google-deepminds-gencast-revolutionises-weather-forecasting-with-ai-driven-predictions/
Source snippet
Google DeepMind's GenCast Revolutionises Weather...9 Dec 2024 — Google DeepMind has introduced GenCast, an AI weather model outperformin...
Source: blog.geogarage.com
Title: google deepminds new ai model is best
Link: https://blog.geogarage.com/2024/12/google-deepminds-new-ai-model-is-best.html
Source snippet
This is the second AI weather model that Google has launched in just the past few months.Read more...
Source: huggingface.co
Title: aifs single 1.0
Link: https://huggingface.co/ecmwf/aifs-single-1.0
Source snippet
ecmwf/aifs-single-1.0Jun 3, 2024 — AIFS produces highly skilled forecasts for upper-air variables, surface weather parameters and tropica...
Source: articsledge.com
Title: ai weather forecasting
Link: https://www.articsledge.com/post/ai-weather-forecasting
Source snippet
2026: Models, Accuracy & Results3 days ago — Google DeepMind's GraphCast outperformed ECMWF's flagship HRES model on 90% of 1,380 verific...

Additional References

Source: science.org
Title: It predicts hundreds of weather variables for the next 10 days.Read mor
Link: https://www.science.org/doi/10.1126/science.adi2336
Source snippet
Learning skillful medium-range global weather forecastingby R Lam · 2023 · Cited by 2239 — We introduce GraphCast, a machine learning–bas...
Source: reddit.com
Link: https://www.reddit.com/r/machinelearningnews/comments/1h7v850/google_deepmind_opensources_gencast_a_machine/
Source snippet
A Machine Learning-based Weather Model that can Predict...Google DeepMind Open-Sources GenCast: A Machine Learning-based Weather Model t...
Source: linkedin.com
Link: https://www.linkedin.com/pulse/deepmind-says-its-new-ai-system-worlds-most-accurate-10-day-greg-pitz-pcdlf
Source snippet
DeepMind says its new AI system is the world's most...DeepMind's GraphCast promises medium-range weather forecasts of “unprecedented acc...
Source: blog.google
Link: https://blog.google/feed/gencast-weather-prediction/
Source snippet
GenCast from Google DeepMind provides better weather...In a paper published in Nature, Google DeepMind introduced its newest AI model, G...
Source: reddit.com
Link: https://www.reddit.com/r/singularity/comments/17v561g/graphcast_ai_model_for_faster_and_more_accurate/
Source snippet
GraphCast: AI model for faster and more accurate global...GraphCast predicts weather conditions up to 10 days in advance more accurately...
Source: linkedin.com
Link: https://www.linkedin.com/pulse/gencast-google-deepminds-ai-model-precise-efficient-weather-2ot5c
Source snippet
GenCast: Google DeepMind's AI Model for Precise...Probabilistic Ensemble Forecasting - Unlike deterministic models offering a single out...
Source: techmonitor.ai
Title: deepmind develops gencast ai model for extended 15 day weather forecasting
Link: https://www.techmonitor.ai/ai-and-automation/deepmind-develops-gencast-ai-model-for-extended-15-day-weather-forecasting/
Source snippet
DeepMind develops GenCast AI model for extended 15...Dec 5, 2024 — Google DeepMind introduced a new AI model, dubbed GenCast, designed t...
Source: pubmed.ncbi.nlm.nih.gov
Link: https://pubmed.ncbi.nlm.nih.gov/39633054/
Source snippet
weather forecasting with machine learningby I Price · 2025 · Cited by 647 — GenCast is an ML weather prediction method, trained on decade...
Source: egusphere.copernicus.org
Link: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-4716/
Source snippet
1.1.0: An update to ECMWF's machine-learned weather...by G Moldovan · 2025 · Cited by 3 — The AIFS has been fully operational at ECMWF s...
Source: oecd.ai
Title: artificial intelligence forecasting system aifs
Link: https://oecd.ai/en/catalogue/tools/artificial-intelligence-forecasting-system-aifs
Source snippet
Artificial Intelligence Forecasting System (AIFS)19 May 2025 — The AIFS is the first fully operational weather prediction open model usin...

Published: May 2025