As artificial intelligence steadily and rapidly works its way into all aspects of our daily lives and businesses, the need for massive data centers to process the enormous amount of information generated will only continue to increase.
According to the Kennedy School of Government at Harvard University1, data centers are driving a massive surge in electricity demand, with U.S. data center power consumption expected to reach 6.7–12% of total electricity by 2028, up from 4% in 2023. Primarily fueled by AI, data center demand could rise to 134.4 GW by 2030, with roughly half of all new U.S. electricity use driven by these facilities. According to conservative estimates from the Data Center Coalition2, electricity demand on the Pennsylvania-New Jersey-Maryland (PJM) grid is expected to increase by around 2-3% per year into the early 2030s.
To place the situation into context, a proposed 565,000-square-foot hyperscale AI data center in suburban Pittsburgh will require 180 megawatts of power3, roughly equivalent to the amount of electricity supplied by two average U.S.-based generating units.
AI servers’ electricity usage is expected to rise nearly fivefold between 2025 and 2030, according to the Pew Research Center4, adding that a single large data center can consume as much electricity as 50,000 households. Pew also reports that global data center electricity consumption is projected to grow from 460 TWh in 2024 to over 1,000 TWh by 2030. A terawatt-hour (TWh) is a unit of energy representing one trillion watt-hours, primarily used to measure large-scale electricity generation or consumption denoting annual production or usage on a national scale.
The nation has not seen this type of electricity load growth for the last 20 years – that’s one full generation that has been alive during a time of fairly flat increase in demand. The AI data center boom, while a major contributor to that increase, is not the only one. Other factors also tied to higher demand for power and economic development include a deepening digital infrastructure, advanced manufacturing, electrification of vehicles, homes, buildings, and industry, and emerging hydrogen production. Obviously, the growing need of more data centers and these additional demands mean that the available electrical capacity needs to be addressed, starting now.
Reliable, affordable, and abundant energy is non-negotiable if the U.S. wants to lead in AI. Wouldn’t it be great if the U.S. only could leverage an abundant, affordable, easily accessible domestic resource to help fuel that demand? Fortunately, we can. It’s the resource that powers the world. The most reliable, dependable resource this country has – coal.
Roughly 56% of the electricity used to power data centers nationwide currently comes from generating facilities that use fossil fuels5. Utilities wisely are extending the life of some coal plants to meet this surging demand. While renewables are expected to meet a portion of the additional demand over the next few years, fossil fuels like coal remain crucial for near-term needs. How interesting that the incredible expansion of knowledge and productivity possible through AI will rely in large part to coal, the most traditional and widespread source of energy in the world.
Sources:
- Belfer Center for Science and International Affairs. AI, Data Centers, and the U.S. Electric Grid: A Watershed Moment.
- Pittsburgh Technology Council. TEQ Hub Report – March 2026.
- Ibid.
- Pew Research Center. What We Know About Energy Use at U.S. Data Centers Amid the AI Boom.
- Environmental and Energy Study Institute. Data Center Energy Needs Are Upending Power Grids and Threatening the Climate.
