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
The 345 MW Natrium reactor stores its heat in liquid sodium, allowing it to keep up with sudden changes in the load of GPU clusters without giant batteries.
A £5 million British trial, over two winters, will test AI-driven drag avoidance in Shanwick Ocean Space, with independent verification by Imperial College.
Faced with the explosion in demand for electricity from AI, data center operators are building their own natural gas-fired power plants, at the risk of causing American CO2 emissions to soar.
SpaceX has applied to the FCC for permission to operate up to one million satellites serving as data centers. The direct thermal effect is negligible; atmospheric fallout is much less so.
This is the first decline in three years in this market, while investments in AI increased the group's emissions by 25% last year.
A report published on 13 August suggests that the actual uses of AI generate 3.3 to 13.3 times higher emissions than the data centers themselves, including through the fossil industry.
John Sterman and Jennifer Turliuk estimate the direct effect of AI at 0.1°C by 2100. The rebound effect would increase the projected warming from 3.3°C to 3.6°C.
Researchers at Koç University are combining molecular simulation and machine learning to sort out hybrid metal-organic/polymer membranes to separate carbon dioxide.
The FERC has unanimously passed an order requiring the largest network operators in the United States to explain, within 30 days, how they intend to clear the queue.
Published in npj Climate Action, a general equilibrium model estimates the net balance of AI at between 0.47 and 1.8 gigatonnes of additional CO₂ per year, due to a lack of sufficient gains from renewables.
Data centers currently account for just over 1% of global electricity demand and 0.5% of CO2 emissions. Projections place the shift around 2030.
A system combining physics and machine learning reduces forecasting errors by between 20 and 60 per cent compared with the European reference centre, over the challenging two- to six-week forecast window.
With micro-reactors that can be deployed within eighteen months and conventional reactors that take more than ten years to build, the timescales for nuclear power are poorly aligned with those for AI.
Based on more than 7 million ice thickness measurements, the model estimates the volume of ice outside the ice caps at 150,000 km³, or 32.3 cm of potential sea level rise.
Three operational systems, including a hybrid physical-AI system, producing 16-day forecasts in 40 minutes using only a fraction of the usual computing resources.
More than 70 per cent of Americans oppose the establishment of an AI data centre near their homes; public protests have already blocked or delayed 75 projects worth 130 billion dollars.
A hyperscale facility can consume up to 5 million gallons of water per day; nearly 150 protests were held in 42 U.S. states in July 2026.
Amazon's planned site in Texas would include a gas-fired power plant that is allowed to emit up to 33 million tons of CO2 per year, more than any other U.S. power plant.
The $3.2 billion project comes two days after the Texas governor suspended data center connections.
Published in Nature at the beginning of August, the model provides an accuracy of two days that was previously achieved at three days, and its code and weights are released under the Apache 2.0 license.