AI weather forecasting enters the energy market as Google targets grid operators with WeatherNext 3

Google’s latest AI weather forecasting mannequin predicts wind velocity at 100 metres above the floor, roughly the peak of a contemporary wind turbine. It additionally forecasts cloud cowl and the way a lot daylight reaches the floor, and it updates each hour. Energy merchants, grid operators and wind and photo voltaic builders already pay different corporations for that information. The introduction of WeatherSubsequent 3 now places Google of their market.

Google DeepMind and Google Research released the mannequin on September 3. It produces a worldwide forecast each hour at as much as five-kilometre decision for floor variables such as temperature and moisture. The earlier model, WeatherSubsequent 2, labored on a 25-kilometre grid and refreshed each six hours. Google says the new energy variables are supposed to assist grid operators and builders predict how a lot energy their wind and photo voltaic property will generate, then match that in opposition to demand.

The shopper aspect of the launch has had most of the consideration. WeatherSubsequent 3 now powers weather ends in Google Search, the Gemini app, Google Maps and the Google Maps Platform Weather API. Behind it sits an enterprise layer that issues extra commercially. The identical forecast information will be queried in BigQuestion and Earth Engine or downloaded in bulk from Google Cloud Storage, with no mannequin setup required by the buyer.

Why the energy sector is shopping for AI weather forecasting

Grid operators are operating a system that has develop into more durable to foretell at each ends. On the era aspect, renewables now account for many new capability. S&P Global Market Intelligence’s US Grid Outlook 2026 initiatives photo voltaic and energy storage as the main sources of latest capability this 12 months, at 51.2GW and 25.7GW respectively out of greater than 90GW of deliberate additions. 

Solar and wind generate in line with the weather fairly than demand, so every gigawatt added makes a short-term forecast extra correct.

On the consumption aspect, the new load is coming from AI. S&P Global identifies the unfold of knowledge centres throughout North America as a main driver of the current surge in electrical energy demand, forcing utilities to revise their load forecasts upward. Deloitte’s 2026 Power and Utilities Industry Outlook initiatives peak demand rising by roughly 26% by 2035, with information centre demand alone doubtlessly reaching 176GW, 5 occasions its 2024 stage.

The price of getting a forecast unsuitable is easy. If an operator underestimates how a lot wind energy will arrive, it has to purchase substitute electrical energy at quick discover, often from fuel crops saved on costly standby. If it overestimates, wind and photo voltaic farms find yourself being paid to change off as a result of the grid can’t soak up what they’re producing. Both outcomes are costly, and each are forecasting failures.

The market Google is coming into

Selling weather forecasts to the energy sector is a longtime enterprise. Vaisala, Solcast, DNV’s WindGEMINI and IBM’s HyperWatch all compete in it. So does Jua, a Swiss agency that claims its EPT-2 mannequin beats Microsoft Aurora and DeepMind’s earlier GraphCast on accuracy whereas updating 24 occasions a day, in opposition to what it describes as a typical 4 updates a day amongst rivals.

Google’s benefit is attain. The identical forecast seems as a desk in BigQuestion, a layer in Earth Engine, an API in Google Maps Platform and the default reply in Google Search. No specialist vendor has that unfold, and the hourly refresh closes the update-frequency hole these distributors have used to distinguish themselves.

The incumbents have one technical argument left. Jua’s revealed place is that physics-based fashions such as ECMWF’s HRES nonetheless outperform purely data-driven AI fashions throughout record-breaking excessive weather, as a result of physics fashions encode guidelines about how energy and mass transfer by way of the environment, whereas AI fashions study patterns from previous information. 

Jua sells a physics-constrained product, so the declare serves its personal pursuits. It additionally describes the situations grid operators fear about most, when a storm falls outdoors something the mannequin has seen in coaching.

What is new, and what’s being oversold

The architectural declare behind WeatherSubsequent 3 is that it learns from actual observations as an alternative of from simulations. Most AI weather fashions, WeatherSubsequent 2 included, are skilled on output from numerical weather prediction fashions, that are supercomputer-driven physics simulations that carry a six-hour information lag. That lag can introduce bias in fast-changing variables such as rain and floor temperature. WeatherSubsequent 3 ingests dwell geostationary satellite tv for pc imagery and trains straight on readings from particular person weather stations.

The shift is actual, although narrower than a lot of the protection has advised. Google’s personal system diagram exhibits the mannequin taking in one-hour satellite tv for pc mosaics alongside conventional historic evaluation. DeepMind senior analysis scientist Ilan Price instructed Bloomberg the achieve comes from not ready for the subsequent evaluation and utilizing the most up-to-date data out there as an alternative. 

Reporting places the remaining information lag at three to 4 hours, down from about seven. Dependence on numerical weather prediction has been diminished, not eliminated.

The accuracy figures want comparable care. Google stories enhancements of as much as 60% in opposition to NASA’s IMERG satellite tv for pc product, 30% in opposition to MRMS radar and 10% in opposition to rain gauge readings at early lead occasions, measured utilizing a normal scoring methodology for likelihood forecasts. Those are three separate baselines, and the percentages don’t add collectively. The broadly repeated declare of fifty% higher precipitation forecasting applies particularly to forecasts a day or extra forward. Every determine carries an “as much as” qualifier, which makes every one a greatest case fairly than a typical consequence.

Google revealed no impartial third-party validation alongside the launch. It factors as an alternative to dwell evaluations by Brightband, whose leaderboard it cites in claiming WeatherSubsequent 3 is the most correct world weather mannequin to this point. A utility contemplating a change away from a paid specialist will care extra about efficiency in its personal service territory, by itself property, than a couple of world leaderboard place.

Google’s personal stake in the downside

Google is promoting forecasting instruments right into a grid downside its personal trade helped create. The information centre build-out driving the load progress utilities are struggling to serve is led by the hyperscalers, Google amongst them, and Google has signed multi-gigawatt renewable procurement agreements to provide its personal amenities.

Accurate prediction of wind and photo voltaic output is straight helpful to an organization matching giant volumes of unpolluted energy in opposition to a load that’s each rising and variable. That is business logic, and it goes some solution to explaining why the energy variables shipped on this launch.

Google has not revealed pricing for enterprise entry to WeatherSubsequent 3, or mentioned whether or not the BigQuestion and Earth Engine information carries normal Cloud question fees or a separate licence. Utilities weighing a transfer away from a paid specialist will need that determine earlier than they weigh any accuracy declare.

2025, greater than 65,000 staff in its Corporate and Investment Bank had been actively utilizing the platform, whereas greater than 90% of its engineers had been utilizing AI coding assistants.

The financial institution additionally mentioned AI-based transaction screening allowed it to assessment greater than twice the earlier transaction quantity whereas lowering guide operator checks by half.

Bank of America is utilizing a generative AI-enabled system known as EricaAssist with greater than 18,000 customer support staff. The device summarises why a buyer is asking, retrieves related data, and recommends doable subsequent steps whereas holding the worker accountable for the interplay.

Bank of America mentioned in July 2026 that EricaAssist can ship contextual steering in underneath three seconds and has diminished common name occasions by almost one minute. The financial institution plans to increase the system to extra servicing eventualities and enterprise traces later in 2026.

(Photo by Google)

See additionally: MIT AI forecasts extreme weather without historical data

Want to study extra about AI and massive information from trade leaders? Check out AI & Big Data Expo going down in Amsterdam, California, and London. The complete occasion is a part of TechEx and is co-located with different main know-how occasions, click on here for extra data.

AI News is powered by TechForge Media. Explore different upcoming enterprise know-how occasions and webinars here.

The publish AI weather forecasting enters the energy market as Google targets grid operators with WeatherNext 3 appeared first on AI News.