Wisconsin leaders oblivious to issues driving AI, energy and climate change | Opinion

Recent articles in the Milwaukee Journal Sentinel tout the return of coal and the expanding use of natural gas as energy sources for the future. Only in America and Russia are there still leaders who do not understand that burning these fossil fuels has brought the Earth to the brink of catastrophic climate change.

China recognizes that rapidly and completely shifting to renewable energy production. Short and long-term energy storage provides the only path to a livable planetary future. Yearly, China installs about 60% of the world’s new wind and solar energy-based electrical power and is positioned to dominate the global market for these new technologies.

In contrast to old fossil fuel technology, renewables use virtually no water, produce no heat, toxic air pollutants and reconnect rich and poor societies, alike, with the unlimited inexpensive source of energy for our planet, the sun.   How to make the transition to renewable solar energy quickly is a prime example of a multi-faceted, real world, wicked problem. Consider We Energies’ plan to produce electricity solely with net-zero carbon (greenhouse gas) emissions by 2050. Between 2005 and 2023, its emissions declined 54% as coal was replaced by natural gas. Renewable energy infrastructure would be built between 2023 and 2030 that reduced total carbon emissions another 26%.

Data centers, like Port Washington, present a ‘wicked’ problem

Something unexpected, something wicked, happened between 2023 and 2025. New energy demand nationwide shot up with the appearance of artificial intelligence applications or “AI” that use energy devouring, large language model computer programs like ChatGPT. In the next 5 years, data centers that house the hardware for such programs are projected to require 20-50% of all the new energy output of the U.S.

Letters: The future I see with AI is more folks unemployed and homeless. Leaders not looking ahead.

Locally, We Energies’ plan must now accommodate the Microsoft data centers under construction or planned in Mount Pleasant and Kenosha WI, which will utilize more energy than all the homes in the city of Madison. A much larger data center envisioned for 1900 acres of farmland west of Port Washington will eventually require as much energy as all the homes in the state of Wisconsin! Together, these data centers will impose at least a huge 40% increase in the production of base-load power during the next 5 years.

How would such a precipitous increase in energy generation impact the fundamental objective to make energy production carbon-free by 2050? There is little clarity about modes of energy generation for these data centers. Cloverleaf in Port Washington will only commit to employing 30-40% renewable energy for the first part of its gigantic facility.

Opinion: We Energies natural gas plants will help power data centers driving job growth

With two other data centers in Beaver Dam and Wisconsin Rapids also being planned, Wisconsinites are facing the likelihood that carbon dioxide emissions will increase substantially not decrease in the future.  In turn, efforts to get to net-zero carbon emissions, will be significantly slowed, if not thwarted.

Wisconsin leaders seem oblivious to impacts of data centers

The state legislature and executive branch seem oblivious to the outsized impacts that huge data centers will have on state-wide energy demand, availability, and cost, let alone their deleterious impact on bottom-line climate goals. It appears that legislators have not considered the environmental consequences of the enormous need for water to cool data centers.

Moreover, after construction, new sprawling data centers will provide few on-site jobs; most profits from their operations will be made almost entirely outside Wisconsin. Exactly, how will Wisconsin benefit from a plethora of data centers?

Finally, no one is talking about the gorilla in the room: A Chinese company, DeepSeek, has developed a generative AI program of comparable sophistication to American versions that operates with about 10% of the energy of its US counterparts. Its computer code is open-sourced so that anyone can use it and compete with big tech. Why do we need all of these voracious data centers when 10% of them will do?

We are faced with a wicked knot of interconnected problems and issues. Currently, requests for new energy infrastructure are considered independently by the Public Service Commission without evident recognition of the intertwined concerns of environment, human health, and the common good.

Collectively, we got ourselves into the existential climate change dilemma by not understanding and establishing policy based broadly on planetary limits not simply economic considerations. We can amend our ways by adopting “Integrated Resource Planning” to coordinate and integrate education, evaluation and decision-making about all facets of energy production and use.  The “wicked knot” of climate change and data centers demands it.

David H. Petering is a distinguished professor emeritus of chemistry and biochemistry at the University of Wisconsin-Milwaukee.

This article originally appeared on Milwaukee Journal Sentinel: Energy demand for AI data centers only worsens climate | Opinion