In this episode of Numbers Scream, Tom Ellsworth dives into the massive capital expenditure boom driven by AI and data centers, highlighting a staggering $1.4 trillion investment by utilities projected by 2030. This figure represents a 27% increase from previous estimates, fueled largely by hyperscalers and AI data center owners who are spending $355 billion this year alone. Ellsworth compares this investment to historic infrastructure projects like the Apollo program and the interstate highway system, emphasizing that this is the largest infrastructure buildout in American history, with significant impacts on both public and private sectors.
Ellsworth then addresses the economic impact of these investments, revealing a concerning ratio of just one permanent job created for every $13 million spent on data center construction. He critiques the tax incentives offered by states like Virginia, where billions in tax breaks have been granted to attract data centers under the promise of job creation. However, the reality shows minimal job growth relative to the massive expenditures, suggesting that these tax cuts may not be justified and that data center operators might have built in these locations regardless of incentives.
Energy consumption by data centers is another critical issue discussed, with projections estimating that by 2028, data centers could consume between 6.7% and 12% of the nation’s electricity, averaging around 9% by 2030. Northern Virginia, a major data center hub, is expected to have 40% of its grid power dedicated to data centers, raising concerns about grid stability and national security. Ellsworth advocates for innovative solutions such as small nuclear reactors co-located with data centers in rural areas to mitigate grid strain and reduce community impact.
The episode also explores the financial burden on consumers, noting that $31 billion in utility rate hikes were requested in the first six weeks of 2026 alone, with 64% of these hikes approved. These increases are driven by the dual costs of power generation and the extensive transmission infrastructure required to support new data centers. Ellsworth highlights how utilities are using the data center buildout as an opportunity to upgrade aging infrastructure, costs that ultimately get passed on to consumers, fueling public frustration and political debate.
Finally, Ellsworth discusses the legislative response to the data center boom, reporting over 300 bills and 100 moratoriums introduced in state legislatures in early 2026 aimed at regulating or halting data center expansion. This surge in legislative activity reflects growing public concern over electricity costs, environmental impact, and community disruption. Despite strong lobbying efforts by the power and AI industries, many of these bills have stalled. Ellsworth calls for stronger leadership and legislation to ensure that AI data center growth is balanced with consumer protection and sustainable energy practices, urging for a formalized ratepayer pledge to manage the costs fairly.
