The Hidden Transfer: What the April Real Wage Decline Is Actually Measuring

The Bureau of Labor Statistics reported on May 12 that real average hourly earnings for U.S. workers fell 0.3 percent from April 2025 to April 2026. Nominal wages grew 3.6 percent; consumer prices grew 3.8 percent. The math left workers behind by three-tenths of a percent.

This is the headline number. It is accurate. It is also one of the more misleading summaries produced by the national accounts in recent memory — not because the BLS got the math wrong, but because an average computed across the full wage distribution, deflated by a national price index, is precisely the wrong tool for seeing what actually happened in April 2026.

What actually happened is more interesting, more inequitable, and more directly traceable to specific decisions made by specific companies in the AI industry.

The Number Behind the Number

The April CPI rose 0.6 percent on a monthly basis, accelerating from March's 0.9 percent. Year over year, the all-items index is up 3.8 percent.

Energy accounts for an unusually large portion of that. The energy index rose 3.8 percent in April alone — more than six times the monthly rate of core services inflation — and is up 17.9 percent over the prior twelve months. Within energy: gasoline up 28.4 percent year over year; electricity up 6.1 percent.

Energy at the headline level accounts for over 40 percent of the month's inflation. The standard analysis stops there and notes that energy prices are volatile, driven by the Hormuz closure and the broader Iran conflict, and will eventually normalize.

That analysis is partly correct and mostly incomplete.

The problem is distribution. Energy costs are not a fixed percentage of consumption across income levels — they are a regressive tax, landing with systematically more force on the households least able to absorb them. The Energy Information Administration estimates that extremely low-income households spend approximately 14 percent of their income on energy, and roughly 11 percent on electricity alone. For low-income households more broadly, the average energy burden is projected at 8.6 percent of income in 2026, compared to 3.0 percent for other families.

When the BLS reports that "real wages fell 0.3 percent," it is computing the average over everyone. For a household spending 8–14 percent of income on energy, a 17.9 percent increase in energy costs doesn't produce a 0.3 percent decline in purchasing power. It produces something considerably worse.

The Diverging Wage Story

The wage side compounds the distribution problem.

Data from December 2025 — the most recent detailed breakdown available — showed nominal wage growth running at approximately 3 percent for high-income households, 1.5 percent for middle-income households, and 1.1 percent for low-income households. That is three different wage growth rates for three different populations being deflated by the same CPI — which already understates the inflation experienced by the bottom of the distribution.

The Cleveland Federal Reserve's February 2026 analysis of wage growth from 2020 through late 2025 identifies the underlying structure. Workers at the 10th percentile saw nominal wages grow 9.7 percent faster, in percentage terms, than workers at the 90th percentile over that period. The headlines during that period described it as an era of exceptional gains for low-wage workers. The Cleveland Fed's finding on the dollar amounts is more clarifying: the 10th percentile worker gained $1.34 per hour in real terms over five years. The 90th percentile worker gained $3.09.

The percentage gain was better at the bottom. The dollar gain was more than twice as large at the top. In an economy where essential goods — rent, food, energy — have fixed dollar costs, not fixed percentage costs, the dollar gain is the relevant number.

Economist Arindra Dube's analysis applying distributional price indices to wage data finds that post-pandemic inflation was explicitly regressive: cumulative inflation between 2019 and 2024 was 23.8 percent for the bottom income decile versus 20.4 percent for the top. The key drivers were food, shelter, and energy — precisely the categories that weigh more heavily in lower-income consumption baskets. The same nominal raise buys less at the bottom than at the top, because the bottom is spending a larger share of income on the things that got more expensive fastest.

April 2026 is not a reversal of this pattern. It is its continuation.

Where the Electricity Bill Goes

The energy component of April's inflation is substantially explained by the Iran conflict and the partial Hormuz closure — geopolitical forces outside anyone's control. This is true and should be stated plainly.

But it is not the complete explanation for electricity prices specifically. Electricity rose 6.1 percent year over year in April, and the structural driver behind that increase — separate from the fuel price shock affecting gasoline — runs directly through the AI industry's infrastructure buildout.

Goldman Sachs projects U.S. data center power demand rising from 31 gigawatts in 2025 to 41 gigawatts in 2026 and 66 gigawatts by 2027 — more than double in two years. The Lawrence Berkeley National Laboratory estimates U.S. data center electricity consumption will nearly double by 2028 from 2023 levels. Data center electricity consumption surged 17 percent in 2025 alone.

This demand is appearing in rate cases. Dominion Energy proposed a monthly household rate increase of $8.51 in 2026 and an additional $2.00 in 2027 for a typical customer, attributable substantially to data center load growth. In the PJM interconnection — the grid serving parts of 13 states and the District of Columbia — data centers have been responsible for approximately 63 percent of wholesale price increases over the past year.

A March 2025 Harvard Electricity Law Initiative paper documented the cost-shifting mechanism directly: utilities serving hyperscale data centers are building out transmission infrastructure whose costs are socialized across all ratepayers. In seven PJM states, at least $4.4 billion in local transmission upgrades were paid for by regular customers to facilitate faster connections for data center operators. The customers who paid for these upgrades are not the customers extracting the benefit.

American households paid nearly 10 percent more for electricity in 2025 than in 2024, representing over $60 billion in additional costs nationwide. That is not simply a consequence of the Iran conflict, which affects oil prices primarily. It is partly a consequence of grid infrastructure investment for which the primary beneficiary is not the residential ratepayer.

The AI industry's capital expenditure announcements — $125–145 billion from Meta this year, $180–190 billion from Alphabet, hundreds of billions more from Microsoft, Amazon, and the hyperscaler ecosystem — are correctly described as private investment decisions. What is less often noted is that the transmission and distribution infrastructure required to power those investments has a socialized cost structure. The investment is private. The infrastructure bill is shared, disproportionately by the people whose bills now consume 8 to 14 percent of their income.

The Safety Net, Such As It Is

The federal government runs the Low Income Home Energy Assistance Program (LIHEAP) to provide energy cost support to households below 150 percent of the federal poverty level. The Trump administration's initial FY2026 budget proposed eliminating the program's $4 billion in funding entirely, on grounds that increased domestic energy production had made it unnecessary.

Congress disagreed: the program survived the appropriations process at $4.045 billion, a $20 million increase over the prior year.

LIHEAP exists. But it serves approximately 18 percent of income-eligible households. Roughly 82 percent of households eligible for energy assistance receive none. In a year when the energy index rose 17.9 percent and forced utility disconnections are projected to reach 4 million households in 2025, the program's reach is materially inadequate to the distributional problem it nominally addresses.

What the Number Actually Measures

The -0.3 percent figure is not wrong. It is a description of the average experience of workers in April 2026, measured using a single national price index.

The distribution it conceals is this: higher-income workers received higher nominal wage increases, face lower energy burdens as a share of income, and are therefore more likely to have experienced positive real wage growth in April. Lower-income workers received lower nominal wage increases, face energy burdens two to five times higher as a share of income, and are therefore more likely to have experienced a meaningful decline in purchasing power — larger in magnitude and more consequential to household survival than the aggregate number suggests.

This is inflation inequality. It is not new. But April 2026 adds a structural dimension that goes beyond the Hormuz closure: the AI infrastructure build, financed by private capital, is now a documented driver of residential electricity prices through the rate-case mechanism. The companies booking record profits from AI are partly financing their computational infrastructure on electricity bills that consume a disproportionate share of income from households who had no say in the investment decision and receive none of the return.

The national accounts will record this as aggregate investment driving aggregate growth. The real wage release will record this as a -0.3 percent average decline in purchasing power. Neither number makes the transfer visible.

That is what the number is actually measuring.

Sources

• Bureau of Labor Statistics, Real Earnings Summary, May 12, 2026. URL: https://www.bls.gov/news.release/realer.nr0.htm
• Bureau of Labor Statistics / Department of Labor, Real Average Hourly Earnings, April 2026, May 12, 2026. URL: https://www.dol.gov/newsroom/economicdata/realer_05122026.pdf
• Bureau of Labor Statistics, Consumer Price Index — April 2026, May 12, 2026. URL: https://www.bls.gov/news.release/cpi.nr0.htm
• Bureau of Labor Statistics, Real Average Hourly Earnings Decreased 0.3 Percent from April 2025 to April 2026, TED (The Economics Daily), May 2026. URL: https://www.bls.gov/opub/ted/2026/real-average-hourly-earnings-decreased-0-3-percent-from-april-2025-to-april-2026.htm
• Federal Reserve Bank of Cleveland, Dollars and Cents: Real Hourly Wage Growth across the Lower Half of the Wage Distribution, Community Development Report, February 18, 2026. URL: https://www.clevelandfed.org/publications/cd-reports/2026/20260218-real-hourly-wage-growth-across-lower-half-of-wage-distribution
• Arindra Dube, Inflation Inequality Meets Wage Compression, Substack. URL: https://arindube.substack.com/p/inflation-inequality-meets-wage-compression
• Rocky Mountain Institute, By the Numbers: Low-Income Energy Assistance. URL: https://rmi.org/by-the-numbers-low-income-energy-assistance/
• U.S. Department of Energy, LEAD Tool, Low-Income Energy Affordability Data. URL: https://www.energy.gov/cmei/scep/low-income-energy-affordability-data-lead-tool
• Goldman Sachs Research, US Data Center Power Demand Projected to Double by 2027. URL: https://www.goldmansachs.com/insights/articles/us-data-center-power-demand-projected-to-double-by-2027
• Belfer Center, Harvard Kennedy School, AI Data Centers and the US Electric Grid. URL: https://www.belfercenter.org/research-analysis/ai-data-centers-us-electric-grid
• Environmental and Energy Study Institute (EESI), Data Center Power Demands Are Contributing to Higher Energy Bills. URL: https://www.eesi.org/articles/view/data-center-power-demands-are-contributing-to-higher-energy-bills
• Rhodium Group, Data Centers and Electricity Demand. URL: https://rhg.com/research/data-centers-electricity-demand/
• Peskoe, Ari (Harvard Electricity Law Initiative), via Utility Dive, How Utilities Subsidize Data Center Growth at Ratepayer Expense, March 2025. URL: https://www.utilitydive.com/news/utilities-subsidize-data-center-growth-ratepayer-cost-shif-harvard-peskoe/742001/
• ElectricityPlans.com / multiple utility data, How Data Centers Impact Your Electricity Bill. URL: https://electricityplans.com/data-centers-impact-electricity-bill/
• National Energy Assistance Directors Association (NEADA), Energy Affordability Project. URL: https://neada.org/energy-affordability-project/
• Utility Dive, Federal Energy Assistance Programs Survive Budget Gauntlet, 2026. URL: https://www.utilitydive.com/news/federal-energy-assistance-programs-survive-budget-gauntlet/811389/
• U.S. Administration for Children and Families (ACF), LIHEAP Household Energy Trends, FY 2020. URL: https://acf.gov/sites/default/files/documents/ocs/RPT_LIHEAP_HEN02HETrends_FY2020.pdf