The backlash against data-centre expansion and the worsening shortage of warehouse workers may be two sides of the same economic problem, according to an analysis based on hourly-workforce data from a staffing company.
The analysis argues that workers facing rising household costs are being asked to support an artificial intelligence and data-centre boom while struggling to keep pace financially. At the same time, the warehouses supplying those facilities are finding it harder to fill shifts.
Hourly workers face a sharper rise in costs
Official headline inflation stood at 3.7% in September, but the analysis argues that this does not reflect the experience of many hourly workers. A survey of 3,083 workers carried out during the summer found that 69.4% earned less than $30,000 a year across all their jobs, while only about half had a savings account.
The analysis created an alternative measure, called “FeelFlation”, by weighting price rises according to the proportion of take-home pay spent on different essentials.
Rent or mortgage costs accounted for 28% of take-home pay and had risen by 3% over the previous year. Interest on debt represented 16% and was estimated to have increased by 21%, while transportation accounted for 15% and rose by 11%.
Discretionary spending made up 13% of take-home pay and rose by 2.6%. Food accounted for 10.9% and increased by 3%, insurance premiums represented 10% and rose by 12%, and healthcare made up 7% and increased by 2.6%.
On that basis, the weighted increase was calculated at 7.86% – more than twice the headline inflation rate. The analysis said the figure could understate the pressure because newer data put insurance premium increases at 16%.
Interest, insurance and transport were identified as the main contributors, together taking 41% of workers’ pay and accounting for nearly 80% of the calculated increase. The analysis noted that official consumer price measures include the cost of goods and services but exclude interest paid on credit cards and car loans used to finance them.
Hourly pay on the staffing platform rose by 3.1% over the same 12 months, leaving the cost of the goods and services tracked by the analysis 4.76 percentage points higher than pay growth.
The survey found that 67.6% of respondents would take an additional shift rather than borrow money when faced with an unexpected bill.
Data-centre expansion adds to warehouse staffing pressure
In Columbus, Ohio, the proportion of warehouse shifts filled through the platform fell from 87.9% in 2024 to 82.8% this year. More than one in six shifts is now going unfilled in the city.
Fill rates also declined in each of the data-centre locations tracked. Northern Virginia fell from 96.2% to 92.6%, Dallas-Fort Worth dropped from 93.5% to 91.4%, and Phoenix declined from 91.5% to 90.1%.
Across the four metropolitan areas, the combined fill rate fell by 1.8 percentage points to 90.3%. Columbus experienced a decline almost three times as large, with the figures coming as warehouses approach the peak season.
The staffing company’s warehouse volume in the areas has more than doubled since 2024, rising from 7,258 shifts to 17,342. The analysis said the fall in fill rates partly reflected demand growing faster than the available workforce.
It cautioned that the platform could not establish whether individual warehouse workers had moved to jobs at data-centre sites. However, it argued that the expansion was likely to attract the most skilled hourly workers first, leaving warehouses competing for a smaller pool of available staff.
The analysis also pointed to a fall in net immigration, from about 3.5 million in 2023 to roughly 410,000 in 2025, citing Congressional Budget Office figures.
Calls for workers to share in the data-centre boom
The analysis said the economic pressure on workers was helping to fuel opposition to new data centres. People who hear promises of vast gains from artificial intelligence alongside warnings that technology could threaten jobs may be less willing to support a power-intensive facility in their area, it argued.
It also said that data centres which are built could worsen labour shortages in the warehouses that supply them, creating a link between public opposition and the staffing problems facing the wider buildout.
The author argued that the US should align legal immigration for skilled trades with the labour requirements of the expansion. Two further proposals were offered as ways to improve workers’ finances more quickly.
One was to provide the benefit of an overtime tax break in weekly pay packets rather than through an annual tax return. Under existing arrangements, workers can deduct the premium portion of overtime pay – the additional half in time-and-a-half – from federal taxable income through 2028, up to $12,500 a year. Social Security and Medicare taxes still apply, and the benefit is received later through a tax refund.
The analysis proposed a two- to three-year overtime tax holiday that would deliver the benefit at each payday. It said this could support workers who already rely on extra shifts rather than borrowing, while helping warehouses fill vacant shifts.
It also called for workers who build, supply and staff data-centre facilities to receive a share of the financial gains promised by artificial intelligence companies and data-centre owners, including through equity.
The central argument is that public support for the data-centre boom will depend on whether hourly workers and the communities hosting the facilities see a tangible improvement in their pay and prospects.
