There are multiple reasons for the dramatic increases in our electricity rates; the demands of Artificial Intelligence (AI) are potentially the most egregious consumer of electricity. AI electricity consumption may be exponentially higher than residential customers’ needs.
A recent article in the Economic Times, attributed to a forecast by Morgan Stanley, provided the following AI summary:
“AI power shortage warning: America may soon run out of electricity as AI data centers drain the grid, Morgan Stanley warns. The bank projects a massive 20% U.S. power shortfall through 2028 — about 44 gigawatts, enough to power 33 million homes. Surging AI demand from Microsoft, Google, and Amazon is pushing the grid to its limits. New data centers, nuclear plans, and gas turbines are racing against time. The AI boom is real, but without power, “
The solution to the problem lies with the electricity generators, grid operators, and the various utility commissions. Whether or not they are positioned, motivated, and conscientiously inclined to manage growth is the real issue. The Electricity Industry is not well prepared to address the service demands of AI. Closely related problems are aging generation facilities and the inability to rapidly adapt and integrate newer technologies. Couple all of these issues and hurdles with a political environment that favors carbon-based fuel sources and one that is inhibiting the growth of wind, solar, geothermal, and nuclear solutions.
Increasing capacity cannot be accomplished overnight, and it should not, or need not, be diverted to AI use at the detriment of traditional customers, especially residential users. One needs to measure the value and benefits of this new technology as capacity is increased. Increasing capacity should recognize the alternatives to carbon-based fuels and the environmental impacts and benefits that can be derived from nuclear, solar, wind, wave, thermal, etc. With the advent of AI, the tech-industry giants seemingly expect everyone to jump on the bandwagon and start shoveling more coal or pumping more gas and oil! Currently, alternative fuel sources provide miniscule amounts of added capacity relative to carbon-based fuels.
The operation of new data-storage facilities, which house the volumes of data essential for AI applications to provide “meaningful” information responses, are significant and possibly disproportionate to the demands for residential customers. In many instances the size of these “data-centers” is being expressed in Acres vs. Square-Feet! The jargon de-jour is “AI Factories”. Often these factories are encroaching upon suburban, residential communities and having a degrading effect on the quality of life, in terms of aesthetics and noise levels.
Countering the increasing demands, being attributable to data-centers, are recently completed studies that identify the requirements to upgrade and/or replace ageing grid components such as wiring and poles. (Reference: Washington Post, 10/26/25, page A13).
In addition to the resource issues, there are other important factors that should not be ignored. Some of these factors are: quality and depth of the AI response, cognitive decline (aka dumbing-down), and unemployment. There is a growing body of literature regarding the societal impacts of AI, and the great majority of it is negative. These elements should not be ignored or minimized.
Electricity Utilities (generations, grids, and commissions) hold the keys to the kingdom! They need to aggressively manage the growth of the AI providers in terms of overall capacity and they need to establish methods of service allocations and pricing for residential, industrial, commercial, manufacturing, AI processes, and AI queries.
The capabilities of the electricity providers are inextricably linked with politics, government, public service, self-interest, profitability, and stockholders. The challenge is to find the “middle ground” where providers can manage the industry’s growth with confidence and sound judgment.
At this juncture, it seems that the majority of companies are implementing AI interfaces with their application systems. One might question the need to do so and companies should assess the benefits to their company and to those using the system. For example: Often seen are opportunities to summarize an article or document: to what avail and/or purpose? What was left out, overlooked, minimized, etc.? The summarization process may inadvertently add employee time and electricity usage to ensure important details were not omitted. Would you sign a contract after having read a summary rather than the fine-print? Companies need to properly asses the value of applied AI.
The big-4 AI Factory developers are Amazon, Microsoft, Google, and Meta. Each of these privately managed corporations is developing its own chip designs and is partnering with AI software developers with tightly coupled relationships and funding. Notable AI language models are Anthropic, ChatGPT, and Gemini.
These tools can be used for real-time queries, work processes, and for fun. There is a high probability that a significant AI load will not be business-related but a form of entertainment, often in poor or bad taste and offensive. There are a number of such examples: many directed at actors, pop musicians, celebrities, and the like. The most offensive one I have seen was a Trump flyover at a No Kings demonstration on Saturday, 10/19 (origin unknown, but attributed to White House staff).
Users of AI should pay for the services they receive. A comprehensive system for managing AI usage and billing is an essential requirement and should be based on the various types of usage. Traditional electricity users (e.g., residential customers) should not have to incur the expenses of AI operations.
The employment reductions envisioned and the impacts that will inevitably occur need to be addressed concurrently. Who will be prepared to address the problems AI may exacerbate on the working population of our country? The societal, social, and economic impacts may be well beyond our comprehension. There are many far-reaching concerns that need to be addressed now!
Calvin Yowell
Talbot County
Mary Smith says
Thank you for this thoughtful piece. It’s especially relevant because data centers are already headed for Talbot County. Maryland has shown, as it did with solar and affordable housing, that it will override county zoning to quickly meet state-level objectives. Decades of imposing strict conservation zoning on land has left many Talbot landowners with few options beyond farming, which is why solar and data centers have become the default economic lifeline. If counties want better outcomes, they need to open the door to more commercial and light industrial uses around villages and key intersections, instead of freezing everything out.
Kent Robertson says
Excellent article, Mr Yowell. If we are to halt the I exirable march of power-hungry AI, there will have to be a massive grass roots campaign across the state (and country). Is anyone leading such an effort? Local Eastern Shore efforts to stop large solar farm construction have fallen on deaf ears in Annapolis. I suspect it will take blackouts in Baltimore and Annapolis to get people to protest loudly enough to be heard.