Google: AI buildout lifted electricity use 37% in 2025
Google’s data centers used more than 42 million megawatt-hours in 2025.
TL;DR
- 01Google’s data centers used more than 42 million megawatt-hours in 2025.
- 02Google’s annual electricity consumption rose 37 percent in 2025, the largest single-year increase in the company’s history as its AI data center buildout continued.
- 03Data centers alone consumed more than 42 million megawatt-hours in 2025, up from 30.6 million megawatt-hours in 2024.
Google’s annual electricity consumption rose 37 percent in 2025, the largest single-year increase in the company’s history as its AI data center buildout continued. Data centers alone consumed more than 42 million megawatt-hours in 2025, up from 30.6 million megawatt-hours in 2024.
How much did Google's electricity use grow in 2025?
Google’s total electricity usage increased 37 percent year over year in 2025, and has climbed by more than 250 percent since 2019. The company’s electricity demand also grew 27 percent in 2024, making the 2025 rise part of a sustained, multi-year acceleration tied to growth in Google Cloud, YouTube streaming, and data center construction and operations supporting AI products.
The jump translated into data center consumption rising from 30.6 million megawatt-hours in 2024 to more than 42 million megawatt-hours in 2025. Google compared this scale to the electricity usage of entire countries, citing New Zealand, Denmark, and Nigeria as rough peers.
How did Google keep emissions lower despite higher electricity use?
Google reported that operational carbon emissions fell 2 percent in 2025 even as electricity use surged, a result the company links to large-scale clean energy purchases and its longer-term procurement strategies. Google has matched 100 percent of its electricity consumption with renewable energy purchases for nine years running and announced purchase agreements encompassing 12 gigawatts of "net-new clean energy" in 2025, the largest annual total in its history.
The sustainability report also emphasized a move toward more granular accounting under a 24/7 carbon-free energy ambition and highlighted investments of more than $3.8 billion between 2010 and 2025 expected to bring 7.5 gigawatts of clean energy online. At the same time, Google noted that "our AI infrastructure buildout is currently accelerating faster than the grid is decarbonizing," a constraint the company says will require stepped-up clean energy investments and closer local partnerships.
Why it matters
The energy footprint of Google’s AI infrastructure now rivals that of small countries, which raises questions about grid capacity and how quickly local electricity systems can decarbonize. Supply-chain emissions are moving the other way: emissions from contracted manufacturers and suppliers grew 25 percent in 2025, driving total ambition-based emissions up 18 percent year over year and leaving Google’s overall carbon footprint at about 14.5 million metric tons of carbon dioxide equivalent, roughly between Ivory Coast and Panama in country rankings.
The sustainability picture also contains trade-offs. Michael Thomas of the Cleanview data platform flagged a potential Texas campus tied to a 933-megawatt natural gas plant that could emit 4.5 million tons of carbon dioxide annually if run without carbon capture. A Google spokesperson told Thomas that the company had not yet signed an agreement for how much power the data center might draw from that plant.
What to watch
Track Google’s progress on its 24/7 carbon-free energy accounting and whether its announced 12 gigawatts of net-new clean energy deals translate into local, hourly matches on grids that remain fossil-fuel heavy. Watch for any finalized agreements that set how much power Google will draw from the 933-megawatt natural gas plant in Texas, and for changes in supply-chain emissions in the next reporting cycle, which would show whether procurement and supplier engagement are starting to reduce contracted emissions.
Written by The Brieftide · Source: Ars Technica
The Brieftide Daily · 06:00
Briefs like this one, in your inbox every morning.
Continue reading
More in AI InfrastructureAI power use strains grids, data centers and AWS demand
Volatile power draw from AI workloads, including at AWS facilities, is increasing demand patterns that stress the electrical grid.
IEEE launches virtual training course on large language models
IEEE is offering a virtual training course that teaches engineers to use large language models as reasoning engines in development.
AI4SE and SE4AI: A decade review of AI in systems engineering
H. Sinan Bank, Daniel R. Herber and Thomas Bradley map three research phases and assess 1.
Hyperscalers AI spending to outpace cash flow by Q3 2026
Epoch AI data shows infrastructure spending growing ~70% annually versus operating cash flow at ~23%, with a crossover around Q3 2026.