In a significant advance for improving weather forecast accuracy, Google DeepMind scientists have developed an AI-based weather model that largely beats the world's most accurate modeling system.

Why it matters: It's the first machine learning model to reliably conduct the vital task of "ensemble-based forecasting."

• That means the same model is run with many different initial conditions to generate probability-based projections.
• Previous AI models only resulted in a take-it-or-leave-it projection (also known as a deterministic model), rather than a range of outcomes.

Driving the news: Google DeepMind researchers told reporters on Monday that the GenCast model shows skill in anticipating extreme events outside the bounds of what occurred during the training period.

• That indicates it may accurately predict unprecedented events that climate change makes more likely and severe.

The big picture: AI-based weather forecasting is about to be further integrated into government weather and climate agencies along with private companies, experts told Axios.

• However, it won't replace human forecasters or negate the need for people with scientific expertise.
• University of Oklahoma meteorologist Aaron Hill, who wasn't part of the new research, said human forecasters have "unparalleled" abilities "to parse through complex weather prediction model output and observations."

Models like GenCast are also unlikely to fully displace the current generation of weather models.

• Rather, they will be used as another forecasting tool in meteorologists' toolbox when predicting day-to-day weather, including extreme events.

Zoom in: The new model is detailed in a study published in the journal Nature yesterday.

• It finds the model has greater skill than the extremely reliable ensemble run by the European Center for Medium-Range Weather Forecasts (ECMWF) on 97.2% of 1,320 forecast metrics they evaluated.
• This includes predicting extreme weather events, the tracks of tropical cyclones and wind power output.
• GenCast's machine learning approach used weather data from 1979 to 2018 for training purposes.

What they're saying: Andrea Lopez Lang, a meteorology professor at the University of Wisconsin-Madison, said the machine learning ensemble is a big advance.

• However, even GenCast still depends on physics-based models to generate the weather data such models train on.
• "We are at an interesting time for weather prediction, AI/ML models are rapidly improving," she told Axios via email. "But we (and these AI/ML models) still require a better understanding of the physical systems that produce the weather, climate, and extreme conditions in the atmosphere for improved forecasts."

Go deeper

President-elect Donald Trump's pick of billionaire and commercial astronaut Jared Isaacman to head NASA will put a largely unknown figure in climate and Earth science circles atop one of the government's top climate research agencies.

Why it matters: Investigating our home planet is one of NASA's core missions, and to date it has been relatively free from partisan interference.

Zoom in: Isaacman is a pioneer in private sector spaceflight, and if confirmed, would be the second consecutive NASA leader to have ventured into space.

• That may provide a clue to his views on climate change, given that space travelers have tended to view the protection of Earth and its environment in a new light after seeing the planet from space.
• This is known as the "overview effect."

The intrigue: Isaacman is a close associate of SpaceX CEO Elon Musk, who has emerged as one of Trump's biggest backers.

• "Jared will drive NASA's mission of discovery and inspiration, paving the way for groundbreaking achievements in Space science, technology, and exploration," Trump wrote in a Truth Social post yesterday.

Zoom out: NASA's satellite programs gather crucial data on the planet's climate and weather systems, and the agency runs labs that track global temperatures, the melting of glaciers and sea ice, and other climate-related functions.

Between the lines: Isaacman is focused most closely on pushing the boundaries of space exploration.

• "I can confidently say this second space age has only just begun," he stated on X after Trump made the nomination announcement.
• "Space holds unparalleled potential for breakthroughs in manufacturing, biotechnology, mining, and perhaps even pathways to new sources of energy," he said.

Worthy of your time: How it Happened — The Next Astronauts

Meta and utility giant Entergy are exploring nuclear power as a potential energy source for data centers.

Why it matters: It's the latest in a breakneck series of big tech plans to tap nuclear sources as AI boosts data center power needs.

• It follows Meta's separate move Tuesday to seek partners for up to four gigawatts of nuclear projects.

Driving the news: The companies are "exploring nuclear energy as a future power supply option alongside renewable sources like solar and wind," Entergy said.

• "This effort includes researching conventional nuclear technologies, supporting small modular reactors and potential upgrades to enhance the output of existing nuclear plants in Southeast Louisiana," it said.

👟 Catch up quick: The news comes alongside yesterday's formal unveiling of Meta's plans for a very large data center in Richland Parish, La., that Entergy will power.

• Axios' Katie Fehrenbacher first reported that plan last month.

What we're watching: More details on what's a 30,000-foot-level Meta-Entergy nuclear announcement for now.

Shell's decision not to lead new offshore wind projects highlights a wider dynamic: the company and peer BP taking a more selective approach to renewables.

Catch up quick: Shell yesterday said it's focused on "maximising the value of our existing renewable generation platforms," and added:

• "While we will not lead new offshore wind developments, we remain interested in offtakes where commercial terms are acceptable and are cautiously open to equity positions, if there is a compelling investment case."
• Reuters first reported the decision about offshore wind, a sector that faces a number of cost pressures. The FT also has more.

Why it matters: As we've noted, it's part of Shell and BP's wider recalibration as fossil fuel demand remains resilient and they look to boost valuations that lag U.S. giants.

• They're getting more choosy about where they should compete in power and renewables.

What's next: Shell is also restructuring its power business, splitting Shell Energy into separate units.

• One will handle renewable generation, and another that deals with B2B retail, customer solutions, marketing and trading, Shell said.
• The company operates around 3.4 gigawatts of renewables capacity worldwide.

The bottom line: "We believe that selective investments in renewable power generation and storage systems, combined with our deep power trading and B2B sales expertise, will enable us to create more value with less emissions," Shell said.

Despite constant and often tragic geopolitical friction, oil prices in 2024 have been historically very stable.

The big picture: "Brent crude oil prices exhibited a minimal average monthly change and a monthly range-bound movement between US$74 and US$90 per barrel in 2024, making 2024 one of the most stable years in the past 25 years," Deloitte, citing spot prices, said in a new report.

That's a tidbit from their wider new 2025 oil and gas industry outlook.