The useful question is what changes for users, developers or buyers, and whether the announcement stays industry context or becomes something people can actually use.

Low-level marine clouds reflect sunlight and cool the planet, but their response to climate change has been uncertain. Researchers at Caltech and Google analyzed thousands of high-resolution simulations to assess how these clouds react to warmer seas and higher CO2. The findings, published July 24 in Science Advances, suggest low clouds thin significantly under elevated CO2, even when temperatures are held constant, intensifying warming through a feedback loop.

The team ran more than 7,000 simulations using NOAA global climate data and Google’s tensor processing units to model cloud behavior across 500 tropical Pacific locations during different seasons. Four climate scenarios were tested: a 4°C sea-surface temperature rise, quadrupled CO2, and combinations of both. The results indicate that cloud thinning accelerates as CO2 levels increase, challenging assumptions in some climate models that underestimate this sensitivity.

The study resolves fine-scale atmospheric turbulence that drives cloud formation, a challenge for traditional models. By leveraging AI-optimized hardware, the researchers achieved unprecedented computational scale, enabling simulations previously infeasible. The dataset, now public, allows scientists to refine turbulence and convection schemes in global climate models and explore past climates like the Eocene Epoch.

The findings support growing evidence that low clouds amplify warming, narrowing uncertainties in climate projections. The team, including members of the Climate Modeling Alliance, plans to integrate the dataset into a new AI-driven climate model to improve turbulence and convection representations. While widespread adoption may take time, the work offers a pathway to reduce long-standing uncertainties in climate predictions.