The Turbine Whisperers Tackling AI's Electricity Habit

Baker Hughes teams with Google Cloud to deploy AI-driven power optimization across US data centers, announced at CERAWeek 2026

15 May 2026

Baker Hughes and Google Cloud have formed a collaboration to apply artificial intelligence to power management inside data centers, the Houston-based oilfield services company announced at the CERAWeek energy conference on March 24. The partnership will draw on Baker Hughes's industrial energy expertise alongside Google Cloud's data analytics tools to reduce electricity waste in facilities where demand is rising faster than infrastructure was built to absorb.

The initiative centres on operational data generated within data centers, much of it historically unused, to identify new efficiencies in power consumption. Baker Hughes contributes knowledge from its turbomachinery and energy systems businesses; Google Cloud supplies the computing and analytical layer. Lorenzo Simonelli, chief executive of Baker Hughes, said the partnership was part of efforts to rewrite what he called "The Energy Equation," the relationship between industrial operations and energy supply.

Data centers account for roughly 2 per cent of global electricity consumption, a share rising quickly as AI infrastructure expands. That pressure has drawn industrial companies into adjacent markets, with turbomachinery specialists increasingly repositioning themselves as energy technology providers across generation, management, and digital optimisation.

The move reflects a broader strategic shift at Baker Hughes, which has sought to extend its role beyond upstream oil and gas services. Cross-sector arrangements of this kind have gathered momentum as power demand from digital infrastructure creates new commercial openings for companies with established energy management capabilities.

Concrete efficiency targets and a deployment timeline have not been disclosed. The practical impact of the collaboration will depend on how quickly the two companies move from testing to full-scale implementation, and whether operational data from data centers proves rich enough to yield measurable gains.