Météo-France: Google and Apple Benefit from Public Data
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
A Massive Public Investment for Accessible Data
Every year, France invests 415 million euros in its meteorological infrastructure, which includes satellites, radars, measurement stations, and supercomputers. Although this data is public and freely accessible, it is utilized by companies like Apple, Google, and Microsoft. These tech giants enrich the data with their own algorithms to monetize it in a global market valued at $2.6 billion.
Weather forecasts have become a strategic resource for many sectors, including energy, agriculture, defense, wildfire management, evacuation decisions during natural disasters, and maritime logistics. When a user checks the weather on their iPhone or via Google, they access information derived from a massive public investment, transformed into a commercial service by companies that did not contribute to the infrastructure.
The Robust Infrastructure of Météo-France
Météo-France operates 550 measurement stations across the national territory, as well as 25 Doppler radars. These radars measure in real-time the intensity and movement of precipitation in the atmosphere, allowing for the tracking of storm developments every five to fifteen minutes over an area with a 150-kilometer radius. Additionally, weather balloons are launched daily from numerous sites to collect data on temperature, humidity, and pressure, which feed into numerical forecasting models.
To process this data, Météo-France uses two supercomputers, Belenos and Taranis, provided by Eviden, a subsidiary of Atos. Their combined power reaches 21.48 petaflops, which is over 21 million trillion operations per second. As of November 2024, these supercomputers ranked 141st and 152nd respectively in the global TOP500 supercomputers list.
Digital Models and International Collaborations
Météo-France relies on two main numerical models: ARPEGE for large-scale forecasts, and AROME for local phenomena such as storms or fog. ARPEGE is ranked second in the world for forecasting in Europe, just behind the ECMWF, the European Centre for Medium-Range Weather Forecasts. In comparison, the NOAA, the equivalent American agency, had a budget of $6.3 billion in 2024, fifteen times that of Météo-France.
The Copernicus program of the European Commission, which produces ERA5, a global atmospheric reanalysis, complements this infrastructure. ERA5 is freely accessible to everyone, including researchers, startups, and multinational corporations.
The Impact of Weather Forecasts on Various Sectors
Weather forecasts have a direct impact on several key sectors. In the energy sector, for example, RTE, the French electricity grid operator, relies on accurate forecasts to balance electricity production and consumption. In agriculture, cooperatives and agricultural insurers use hyper-local data to decide on pest control treatments and harvest dates.
In the defense sector, weather conditions influence military operations, including drone operation windows and artillery fire. In terms of reinsurance, natural disasters caused $320 billion in global economic losses in 2024, of which $140 billion was insured. The ability to predict extreme events has become a competitive advantage.
From Public Data to Commercial Service
Private companies like Google retrieve data from Météo-France, NOAA, and Copernicus, enrich it with their own machine learning models, and then sell it or integrate it into their products. ERA5 data is available as downloadable files, segmented by atmospheric variable. Major tech companies automate the processing of this data to convert it into training sets compatible with their neural networks.
Apple acquired the Dark Sky app for its hyper-local forecasts, integrating this technology into iOS. In 2024, Google DeepMind released NeuralGCM, an atmospheric forecasting model based on ERA5 reanalyses. Microsoft has also invested in this area, making AccuWeather the exclusive weather data provider for Azure Maps, integrating public data from Météo-France and other sources.
Conclusion
The free access to public weather data raises questions about how this information is used by private companies to create commercial services. While Météo-France continues to provide accurate forecasts, the challenge remains to ensure that this data benefits society as a whole, rather than just a few private players.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.