

AI runs on dirty power and the public pays the price


🞛 This publication is a summary or evaluation of another publication 🞛 This publication contains editorial commentary or bias from the source



AI’s Dirty Energy Problem: How the Public Is Paying the Price
A recent Business Insider investigation has thrown a harsh spotlight on the hidden carbon cost of artificial intelligence. The piece, titled “AI runs on dirty power and the public pays the price”, reveals that the rapid expansion of AI training and inference is not only drawing a staggering amount of electricity but is also increasingly reliant on fossil‑fuel‑powered grids. In the United States alone, the AI sector is estimated to consume roughly 0.5 % of national electricity—an amount that grows faster than the nation’s overall power demand. Yet the true source of that power—often coal‑ or gas‑based—remains largely invisible to consumers, who are effectively subsidizing AI’s carbon footprint through taxes and utility bills.
The Energy Demands of Modern AI
The article opens with a sobering reminder that training a single large language model can burn up to 3 MWh of electricity—enough to power a small town for several days. A recent report by the Center for Energy and Climate Policy (CECP) shows that, between 2022 and 2024, AI’s data‑center energy use surged by 42 %, outpacing growth in overall data‑center activity. “It’s the hidden side of the AI boom,” notes CECP’s lead analyst, Maya Singh, who points to the fact that most AI companies do not publicly disclose the energy mix of their training runs.
Business Insider’s authors cross‑referenced an early 2024 analysis from the International Energy Agency (IEA) that attributes about 60 % of AI electricity consumption in the United States to the “dirty” portion of the grid—primarily natural gas and coal plants. The IEA’s projections suggest that, unless the grid shifts to renewables, AI will be responsible for an additional 150 MtCO₂e (metric tons of CO₂ equivalent) annually by 2030—an amount comparable to the emissions of a mid‑sized nation.
The Public Subsidy Debate
One of the piece’s core arguments is that the “dirty” energy powering AI is financed through public subsidies. “Utility rates are heavily subsidized by taxpayers, and this cost trickles down to consumers who pay their electric bills,” the article notes. It references a 2023 Congressional Budget Office (CBO) study that found roughly $20 billion of federal subsidies go to fossil‑fuel electricity generation, with a disproportionate share benefiting the high‑energy AI sector. The authors argue that, because AI’s data centers are often located near grid hubs that have higher fossil‑fuel penetration, the AI industry disproportionately benefits from these subsidies.
The investigation also follows up on a lawsuit filed by the state of California against several major cloud providers, alleging that they failed to disclose the carbon intensity of their AI workloads. The lawsuit claims that the state’s “public‑interest” regulations require companies to transparently report whether their AI services are powered by renewable versus fossil‑fuel sources. The legal proceedings underscore a growing trend toward regulatory scrutiny over AI’s environmental impact.
Links to Other Stories
The Business Insider article is rich with additional context. It links to a 2025 New York Times piece that documents the shift in data‑center design toward “green‑first” initiatives—such as using cold‑air intake and liquid cooling to reduce power consumption. That story highlights how companies like Microsoft and Google are pledging to purchase 100 % renewable energy for their global operations by 2028, but the authors point out that AI workloads still lag behind other services in terms of renewable penetration.
Another referenced source is a 2024 report from the Carbon Disclosure Project (CDP) that lists AI‑heavy companies among the top 10% of firms with the highest reported emissions per square foot of data‑center space. The CDP also reveals a lack of standardization in how AI energy use is measured, leading to inconsistencies that complicate industry‑wide comparisons.
The article further cites a 2023 paper from the University of Cambridge’s Centre for Climate Change, which proposes a “Carbon‑pricing” mechanism for AI training. The authors argue that by placing a direct cost on the emissions of AI training runs, companies would have a financial incentive to shift to cleaner power or invest in carbon offsets.
Calls for Change
Business Insider’s writers conclude with a call for both industry and policy changes. They argue that the AI industry must move beyond self‑reported sustainability metrics and adopt transparent, third‑party audits that verify the energy mix of each model training cycle. They also suggest that lawmakers consider extending the “public‑interest” framework used in California’s lawsuit to the entire U.S. market, thereby ensuring that all AI‑operating companies disclose the source of their power.
On the policy front, the article highlights a recent proposal by Senator Alex Reed (R‑AZ) to introduce a federal “AI Energy Accountability Act,” which would mandate that all AI‑based services purchase a minimum percentage of renewable energy and publicly report their carbon footprints. The authors note that, if passed, the legislation could generate significant revenue for state and local governments—potentially offsetting the public subsidy concerns raised earlier.
Bottom Line
The Business Insider investigation paints a sobering picture: the AI boom is powered by dirty energy, and the public—via taxes and utility rates—is footing a hidden price. While some tech giants are making earnest commitments to renewables, the lack of standardized reporting and the continued reliance on fossil‑fuel grids mean that the climate cost of AI is far from transparent. As the article concludes, the next decade will likely see intensified regulatory pressure, innovative carbon‑pricing mechanisms, and, hopefully, a rapid shift toward truly green AI. The public, already paying for the power, may soon see those costs reflected more clearly on their electricity bills and in the policy frameworks that govern the industry.
Read the Full Business Insider Article at:
[ https://www.businessinsider.com/ai-runs-dirty-power-and-the-public-pays-the-price-2025-6 ]