Google grants Ai2 $5M to turn AI weather model into a climate forecaster
What's the deal? Google is giving Seattle's Allen Institute for AI (Ai2)Dealroom has a profile for this one. Try Dealroom → about $5 million, plus $1.8 million in cloud credits, to adapt a Google weather model into an open-source tool for projecting climate change. The three-year grant, announced in November 2026, comes from Google.orgDealroom has a profile for this one. Try Dealroom →, the company's philanthropic arm, with technical support from Google Research.
What's the endgame? Ai2 plans to build NeuralGCM+, an open-source model that gives scientists, water agencies, and planners a faster look at how rainfall, temperatures, and water supplies could shift over the next five to 50 years. It will build on Google's NeuralGCM, which mixes physics equations with AI.
Why now? Google's model looks ahead a year at most and has less interest in decades-long projections. "Climate modeling is a bit more of a public good," said Oliver Watt-Meyer, Ai2's director of research for AI climate modeling. "It is information that people need to know, but it's often more government agencies that are thinking about this."
What could go wrong? Weather models learn from observed patterns, while climate projections must stay credible under future conditions that may fall outside them. When researchers simulated ocean temperatures about 7°F warmer than today, NeuralGCM's atmospheric predictions broke down. Ai2 will connect the model to an ocean and sea ice model and build in more physics to fix that.
Who's doing the work? Ai2 will add four people to its eight-person climate modeling team, which is made up largely of atmospheric scientists backed by the institute's engineers and computing power. Alex Diaz, head of applied science at Google.org, said Ai2 is "uniquely positioned to extend Google's NeuralGCM model into the next generation of long-term climate projections."
The signal: At $5 million, this ranks among the larger AI grants on record, in the 95th percentile of US artificial intelligence grant rounds. It contrasts with commercial weather capital — Tomorrow.io raised $175 million in February and another $35 million in May — underscoring that long-term climate projection remains more public good than business. The funding alone does not establish that the resulting projections will be reliable.
Read more: geekwire.com
Image credit: Allen Institute for AI (Ai2)
