Google is strengthening its push into artificial intelligence and cloud computing by expanding its partnership with NextEra Energy to power massive, AI-scale data centres across the United States.
Google Cloud and NextEra Energy have announced plans to jointly develop multiple gigawatt-scale data centre campuses, aimed at meeting the rapidly growing energy needs of hyperscale computing and AI infrastructure.
Google Cloud and NextEra Energy partnership
The new agreement builds on an existing partnership between the two companies, which already includes around 3.5 gigawatts of power capacity that is either operational or under contract.
Amidst the rising demand for data centres, the increased collaboration is significant because it reflects a growing convergence of energy planning and technological development on an unprecedented scale.
The companies reported that recent highlights include the restart of the Duane Arnold Energy Centre in Iowa, in addition to two long-term power contracts that are set to add 600 megawatts of clean energy to the Oklahoma power grid, with this energy specifically meant to power Google’s infrastructure.
The development of the first three campuses for the data centres has already begun, based on the new agreement. The two firms are targeting mid-2026 for the launch of the first commercial tech product, which is under collaboration on the Google Cloud Marketplace.
An important area that the collaboration is aiming to address is the acceleration of the implementation of a data centre, with common infrastructure issues such as land, grid, and load management being considered.
Nuclear energy is a very core part of this plan. NextEra has called on regulators in the USA to reinstate the grid interconnect rights on the Duane Arnold nuclear power plant, which indicates a return to nuclear base-load power in order to serve the continuous energy requirement for AI tasks.
Beyond infrastructure, Google Cloud will also help modernise NextEra’s operations using AI tools like TimesFM 2.5 for forecasting and WeatherNext 2 for weather prediction.