AM-IA brings together news, research papers and applications relating to artificial intelligence published around the world into a single feed — and makes them available in your European language at the click of a button.
Scroll to browse · Click to filter
Whilst the UN Secretary-General is calling for all data centres to be powered by renewable energy by 2030, investment in gas and coal is on the rise again.
A semi-permeable membrane liquid cooling technology enables high-density data centres to operate in tropical climates whilst consuming almost no water.
According to Bank of America, US power companies will add only 93 GW by 2030, whereas 230 GW is needed: the bottleneck is no longer demand, but the supply of energy.
An analysis by Allianz Trade estimates data center emissions at 286 million tons of CO2 in 2025, well above the International Energy Agency's projections.
A report by UNU-INWEH argues that a carbon-only assessment misses the point: water, land and e-waste weigh just as much.
A study published in Science Advances puts GraphCast, Pangu-Weather and Fuxi in difficulty on the rarest events: trained on the past, they underestimate the intensity of records.
USC researchers are coupling VIIRS and GOES data with physical simulations to move from reactive observation to fire front prediction.
The study combines dynamic modelling, photovoltaic simulation and learning optimisation to arbitrate between billing, profitability and avoided emissions.
Researchers at the University of Illinois at Chicago are combining AI, materials physics and supercomputers to speed up the discovery of new CO₂ capture materials.
Global electricity demand will grow by more than 3.5% per year until 2030, driven in particular by data centers, whose consumption is expected to more than double.
NeuralGCM combines traditional atmospheric physics with machine learning to simulate the climate more quickly, without compromising on accuracy.
Whilst 2025 was characterised by a rush for ever-greater computing power, 2026 now requires AI stakeholders to reconcile this growth with their climate commitments.
Google DeepMind’s GenCast model now predicts the weather and the risk of extreme events with unprecedented accuracy, and its code has been made publicly available.
Google Research’s flood forecasting system now provides up to 24 hours’ advance warning of flash floods in urban areas – a crucial timeframe for alerting the public.
Researchers at Cornell have shown that AI-driven hydrological models provide more reliable flood projections than physical models alone, as the climate changes.
A new AI-driven system designed to better predict forest fires and their impact on the global carbon cycle was presented at the annual conference of the American Meteorological Society.