| Plante, Carrasco, Miranda Pineda and Patel, Federal Reserve Bank of Dallas Working Paper 2606r1 (opens in a new tab) March 2026, revised October 2026 · third party estimate | Wholesale electricity price effect and retail electricity rate pass through | Data center demand accounts for an average 6.5 percent increase in wholesale electricity prices across 23 regional markets in 2024 and 2025, reaching 15.0 percent (+$5.22 per MWh) in SERC Virginia and Carolina North and 15.4 percent (+$5.68 per MWh) in the Northwest Power Pool, with wholesale changes passing into retail prices roughly in proportion to the wholesale share of the retail bill over 12 to 24 months. | 2019 to 2025 historical and 2026 to 2028 scenario projections | Synthetic supply stack model across 23 regional power markets combined with reduced form retail pass through regressions | Federal Reserve Bank of Dallas (government research working paper; views are the authors') |
| Watten, Bistline and Blanford, Electric Power Research Institute (arXiv:2606.19777) (opens in a new tab) | State level retail electricity rate change across all customer classes | Where data centers pay standard industrial rates and improve grid utilization, state average retail rates could fall by roughly 1 percent nationally by 2030 (up to 4 percent in high growth states, or roughly $7 billion per year in fixed cost spreading); outcomes reverse toward higher residential bills if new generation and transmission upgrades outpace load revenues. | 2026 to 2030 | Capacity expansion and utility revenue requirement model allocating fixed transmission and distribution costs across customer classes | Electric Power Research Institute (nonprofit energy research institute funded primarily by member electric utilities) |
| Meeks, Pless, Qi and Wang, MIT Center for Energy and Environmental Policy Research (opens in a new tab) December 2025 · third party estimate | Realized utility revenue and customer bill changes from historical data center openings | In the 6 years following data center entry (1998 to 2018), average electric utility revenue rose 8 percent (+$104 million per utility per year), residential bills rose 2 percent (+$30 per household per year) and commercial bills rose 4 percent (+$262 per customer per year), driven by peak coincident demand charges and infrastructure capital additions. | 1998 to 2018 | Difference in differences event study of historical data center openings matched to utility level EIA Form 861 billing and revenue panels | MIT Center for Energy and Environmental Policy Research (university working paper) |
| Norris, Profeta, Patino Echeverri and Cowie, Duke University Nicholas Institute (NI R 26-04) (opens in a new tab) | Grid headroom and new gigawatts of large load accommodated under minimal peak hour curtailment | Across 22 major US balancing authorities, roughly 76 GW of new data center load can be integrated without building new peaking capacity if facilities curtail load by 0.25 percent of annual operating hours (roughly 22 hours per year), expanding to 98 GW at 0.5 percent annual curtailment and 126 GW at 1.0 percent annual curtailment. | 2025 to 2035 | Hourly balancing authority load duration and reserve margin modeling across 22 US power systems | Duke University Nicholas Institute for Energy, Environment & Sustainability (university research institute) |
| Jeremiah Johnson et al., North Carolina State University (Environmental Research Letters) (opens in a new tab) | Wholesale electricity prices, residential bills by state and power sector carbon emissions through 2030 | Projected data center and crypto load growth raises US electricity demand 6 to 12 percent by 2030 and increases residential electricity bills up to 57 percent in the most heavily affected states (including Virginia at roughly $32 per month or $384 per year for an average household) while adding 49 to 125 million metric tons of CO2 per year. | 2024 to 2030 | National capacity expansion and unit commitment dispatch model (TEMOA) across US grid regions | North Carolina State University researchers in Environmental Research Letters (peer reviewed journal) |
| Energy and Environmental Economics (E3), Understanding the Drivers of Rising Electricity Rates and the Role of Data Centers (opens in a new tab) May 2026 · third party estimate | Historical drivers of US retail electricity rate increases and utility rate design for large loads | Historical US retail electricity rate increases from 2020 to 2025 were driven primarily by distribution capital spending, wildfire mitigation, natural gas price spikes and inflation rather than data centers, and well designed large load tariffs can keep new data center additions net neutral or beneficial for existing ratepayers. | 2020 to 2025 historical and forward rate design case studies | Decomposition of historical utility revenue requirements and illustrative utility tariff cost of service modeling | Energy and Environmental Economics (E3), industry commissioned consulting report |
| Nicholas Z. Muller, National Bureau of Economic Research Working Paper 35100 (opens in a new tab) | Monetized air pollution and greenhouse gas health and climate damages per megawatt hour of data center electricity use | Marginal air pollution and greenhouse gas damages from fossil fuel generation serving US data centers average $49 to $94 per MWh ($0.049 to $0.094 per kWh), rivaling or exceeding the direct wholesale price of electricity in coal and gas heavy interconnects such as PJM, MISO and SERC. | 2022 to 2025 | Integrated assessment marginal emissions and air quality transport model (AP4T) coupled with EPA CEMS unit level generation data | National Bureau of Economic Research (university working paper by Nicholas Z. Muller, Carnegie Mellon University) |
