AI Demand Drives Big Tech's Rising Power Bills

The burgeoning demand for artificial intelligence (AI) is leading to a significant surge in electricity consumption by data centers, prompting states like Indiana and Ohio to implement new policies aimed at ensuring these tech giants contribute more to the power infrastructure costs. While these measures seek to alleviate the burden on residential customers, concerns remain that average citizens may still face higher utility bills.

Data centers, the physical backbone of the digital world and increasingly, AI, are voracious energy consumers. A single large data center can use as much power as tens of thousands to over 100,000 U.S. homes, and the largest centers currently under construction are projected to use 20 times that amount. The nation’s approximately 2,700 data centers consumed over 4% of all U.S. electricity in 2022, a figure projected to more than double to 9% by 2030. Some estimates suggest that by 2029, consumers and small businesses could see their electricity bills rise by as much as 70% due to this escalating demand. The electricity use of AI data centers is growing faster than the grid can accommodate, leading to strain on the power infrastructure and contributing to increased costs for all consumers.

In response to this escalating demand, states are taking action. Ohio’s Public Utilities Commission (PUCO) recently approved a new tariff proposed by American Electric Power (AEP) requiring large new data centers to pay for at least 85% of the energy they subscribe to use each month, even if their actual consumption is lower. This measure, effective in 2025, aims to ensure that data centers contribute more directly to the costs of building and maintaining the grid infrastructure necessary to meet their substantial electricity needs, thereby protecting other ratepayers from disproportionately shouldering these expenses. The ruling also lifts a moratorium on new data center construction in central Ohio that had been in place since March 2023.

Similarly, Indiana is grappling with how to accommodate the energy demands of massive data centers while safeguarding its residents. An $11 billion Amazon Web Services data center campus under construction in Northwest Indiana alone could use as much electricity as 1.5 million households, or up to half the state’s households. Indiana regulators have approved a settlement agreement setting terms for connecting large loads, including data centers, to the grid. This agreement, involving Indiana Michigan Power (an AEP utility), consumer advocates, and tech companies like Amazon, Google, and Microsoft, aims to ensure that the cost of new generation and grid upgrades is not unfairly passed on to regular customers. Under this agreement, data centers in I&M territory must provide collateral during early years of operation, sign contracts of at least 12 years, and agree to pay at least 80% of their expected monthly demand. Despite these agreements, some advocates in Indiana are calling for a moratorium on new hyperscale data centers until further study on their potential impact on power costs and the need for new power plants is completed.

The challenge is not merely about increasing supply, but also about the nature of AI’s energy consumption. AI data centers often require high-density infrastructure, with racks of the latest AI chips consuming at least 10 times as much power as regular web servers and generating significant heat that necessitates advanced cooling solutions, including liquid cooling. The United Nations Secretary-General António Guterres has urged major tech firms to commit to fully powering data centers with renewable energy by 2030, highlighting that a typical AI data center consumes as much power as 100,000 homes.

While many major AI data centers are already exploring or utilizing 100% renewable energy from sources like solar and wind, and even emerging technologies like geothermal and hydrogen fuel cells, the sheer scale of AI-driven demand is overwhelming the current capabilities of existing renewable technologies and the aging power grid. Energy efficiency measures within data centers, such as optimizing cooling systems, using more efficient hardware, and virtualization, are crucial for mitigating this demand. Furthermore, strategies like shifting workloads to different times can help avoid stressing the electric grid and potentially save costs, though the emissions impact depends on the regional energy mix.

The ongoing debate in states like Indiana and Ohio underscores a critical nationwide challenge: balancing the economic benefits of a booming tech sector with the need to maintain affordable and reliable electricity for all consumers. As AI continues its rapid expansion, innovative policies, significant infrastructure investments, and a concerted effort towards greater energy efficiency and renewable integration will be vital to manage the escalating power demands and prevent the “Big Tech power bill” from disproportionately falling on everyday residents.