On July 15, leaders from the energy and technology sectors joined US government officials at Carnegie Mellon University (CMU) for the first-ever Energy and Innovation Summit. The gathering put a spotlight on Pennsylvania’s potential to power the AI revolution, but it also exposed a troubling reality. Meeting the massive energy needs of AI data centers could mean a big win for fossil fuels.

The numbers are staggering. The International Energy Agency (IEA) projects that by 2030, global data centers will consume more electricity than Japan does today. In the US, AI-related data processing could require more power than manufacturing aluminum, steel, cement, and chemicals combined. Without sustainable solutions, decades of hard-won climate progress could be reversed.

Despite this looming risk, sustainable energy was far from the main focus at the CMU summit. “You need the natural gas or coal infrastructure in order to provide these giant AI data centers the power that they need,” said Commerce Secretary Howard Lutnick, according to Axios. Of the $90 billion in Pennsylvania investments announced, at least 25% will go toward fossil fuel energy production. Wind and solar, cheaper and faster-growing sources of electricity, were barely mentioned and even dismissed by some attendees.

Read More: Green Chemistry Gains Ground: Entrepreneurs Unite for a Sustainable Future

The federal AI energy plan under the Donald J. Trump administration prioritizes geothermal and nuclear power, with no mention of solar or wind. While these sources can be part of the solution, their limited existing infrastructure means they can’t immediately meet AI’s skyrocketing energy demand.

Cooling technology presents another challenge. Around 40% of a data center’s electricity use goes toward cooling high-performance processors, often through chilled-water systems that consume thousands of gallons a day. This strains local water supplies while driving up energy use.

Innovative solutions are emerging. In China’s Hainan Province, researchers are experimenting with underwater data centers located near offshore wind farms. By submerging servers in the sea, cooling energy requirements drop by 30%, and proximity to renewable energy sources reduces dependence on fossil fuels. While still experimental, such ideas hint at how creative engineering, particularly from chemical engineers, could make AI more sustainable.

China’s energy mix still relies heavily on coal, which supplies 60% of its power. However, the nation has added more solar and wind capacity in the past five years than the rest of the world combined. By 2030, renewable output is expected to surpass coal, and by 2035, the IEA estimates renewables and nuclear together will supply nearly 60% of the electricity for Chinese data centers.

Also Read: Climate Lawsuits Rising in Nearly 60 Countries, Says Report

For AI, the stakes are high. Its applications, from healthcare breakthroughs to climate modeling, can be transformative, but the nearly 24/7 energy demand of AI data centers risks undermining efforts to decarbonize electricity grids. Chemical engineers could help tip the balance through innovations in liquid cooling, waste heat recovery, and new clean energy technologies.

If US policymakers and tech leaders truly want to lead in AI, they must prioritize powering it sustainably. Winning the AI race should not mean losing the planet.

Follow more news and views via our Environment and Featured Articles sections, and stay updated on top ESG events to attend in 2025 for industry insights and networking.

Source: c&en