Data centers are gobbling up kilowatts
of ConsumerAffairsJuly 23, 2025
-
Rising demand from AI data centers is putting pressure on power grids across the globe.
-
Electricity costs are starting to climb in areas with high AI infrastructure growth.
-
Regulators and utilities are grappling with how to balance innovation with sustainability and affordability.
As artificial intelligence rapidly embeds itself into everything from search engines to personal assistants to medical diagnostics, the demand for electricity is growing, placing mounting stress on the worlds electric grids. Those costs are beginning to show up on consumers electric bills.
The Consumer Price Index for June showed electricity costs rose a full percentage point from May and are up 5.8% year-over-year.
Modern AI models, especially large language models and generative AI systems, require a lot of power. These systems are hosted in data centers equipped with specialized hardware, such as GPUs and TPUs, which run at high intensities day and night.
According to the International Energy Agency (IEA), electricity consumption by data centers could double by 2026, fueled primarily by AI and cryptocurrency mining.
OpenAIs GPT-4 model, for instance, is estimated to consume several megawatt-hours of electricity daily when operating at peak capacity. And thats just one model. Multiply that by thousands of instances across companies like Microsoft, Google, Meta, and Amazon, and the aggregate demand starts rivaling that of some small countries.
Costs to consumers
While many of these AI services are hosted in remote data centers, the electricity they use can affect regional grids and prices. In the U.S., states like Georgia, Virginia, and Oregon hubs for AI and cloud infrastructure have seen significant upticks in commercial power usage. Utilities in those areas have begun to file rate hike requests to offset the infrastructure investments needed to accommodate AI-driven load increases.
In Northern Virginia, Dominion Energy is seeking approval to spend over $800 million on grid upgrades linked to new data center demand. Some critics warn that residential customers will shoulder part of the cost, even though they arent the direct beneficiaries.
Similarly, in Ireland where a boom in AI data center development is underway national regulators warned earlier this year that surging electricity demand may necessitate rationing or higher consumer tariffs to maintain supply security.
Environmental tradeoffs
Beyond electric bills, AIs energy appetite is also raising environmental concerns. While tech companies claim to offset their power consumption with renewable energy purchases, the reality is a bit more complex. Renewable generation isnt always available when data centers need it most, leading to greater reliance on backup fossil fuel sources.
A recent MIT study concluded that training a single large AI model can emit as much carbon as five cars over their full lifetimes. The carbon footprint grows significantly when models are continually retrained or fine-tuned for specific tasks, as is increasingly the norm in enterprise applications.
#electric #bill #rising #reason