Scot might not be able to help, but A.I. might

ChatGPT, We Need More Power!

Kim Bellard
5 min readApr 29, 2024

I know: you want to live in the promised world where we get all our electric power from renewables, such as solar or wind. We’ll finally stop burning all those fossil fuels. And you also want to live in the promised world where artificial intelligence — A.I. — makes our lives easier, more prosperous, and safer.

The problem is, right now those are two different worlds.

We may think of A.I. as just being there, in the cloud, but we sometimes forget that cloud computing requires vast data centers, each of which consume massive amounts of electricity. Companies interested in cloud computing and/or A.I. are investing in data centers like crazy. Reuters reports: “Nine of the top 10 U.S. electric utilities said data centers were a main source of customer growth, leading many to revise up capital expenditure plans and demand forecasts.”

The article goes on to note:

Overall, power use from the thousands of giant computing warehouses that comprise data centers is expected to triple globally from less than 15 terawatt-hours (TWh) in 2023 to 46 TWh this year, according to Morgan Stanley research.

“The truth of the matter is these things (data centers) are pigs when it comes to energy use, and now they’re the size of an elephant,” said Eric Woodell, an expert who specializes in data center operations.

Credit: Washington Post

As a result, The Washington Post warns: “Vast swaths of the United States are at risk of running short of power as electricity-hungry data centers and clean-technology factories proliferate around the country, leaving utilities and regulators grasping for credible plans to expand the nation’s creaking power grid.”

“When you look at the numbers, it is staggering,” Jason Shaw, chairman of the Georgia Public Service Commission, told The Post. “It makes you scratch your head and wonder how we ended up in this situation. How were the projections that far off? This has created a challenge like we have never seen before.”

The key culprit? A.I. The Post says: “A major factor behind the skyrocketing demand is the rapid innovation in artificial intelligence, which is driving the construction of large warehouses of computing infrastructure that require exponentially more power than traditional data centers.”

The problem isn’t only from A.I. — there’s also the impact of crypto-mining, remote work, and EV charging — but put A.I. at the top of the list.

Well, then, we’ll just build more electric plants, right? Not so fast. In the first place, that requires lots of capital investment, something power companies are stingy about. In the second place, nothing about building power plants is fast, from the permitting process to the construction to the connecting to the power grid. In the third place, generating power using what?

“The big utilities are typically most comfortable with one way of doing things: building those big, conventional power plants,” Heather O’Neill, president of Advanced Energy United, a trade group representing low-carbon technology companies, observed in NYT.

President Biden promised to shift electric power generation to all renewables by 2035, but that promise was made before the impact of A.I. on power use was fully evident. “The growth we’re seeing is historic in scale and speed,” Kendal Bowman, president of Duke Energy’s operations in North Carolina, told NYT. “But it’s also going to be a challenge, particularly in the near term, to see carbon reduction at the same time we’ve got this unprecedented growth.”

And it’s not that the power grid is quite as robust or technologically advanced as we’d like. “One of the biggest hurdles is the outdated power system infrastructure,” three Rand researchers wrote in The National Interest. “Many of these systems were built decades ago and are not equipped to handle the demands of rapidly emerging technologies and changing consumer needs. Therefore, significant investments will be required to update the grid and realize the benefits of emerging technology.”

As a result: “What we’re seeing in the market is that these projects are not coming online fast enough to meet the local demand for the for the data centers,” Rystad Energy analyst Geoff Hebertson told Reuters.

Credit: Washington Post

This is impacting those climate goals. The Wall Street Journal reports:

For many utilities, the solution to rising demand is to keep coal-fired power plants burning for longer and add natural-gas power plants to balance big expansions of renewables…Utilities in Georgia and North Carolina are adding fossil-fuel power or considering delaying the shutdown of coal-fired plants to meet the demands of data centers and other industries.

That doesn’t make many consumers, or even big Tech companies, happy. “No data center wants to be tied to the need for new fossil resources, that’s the problem,” Brian Janous, former vice president of energy at Microsoft, told WSJ. “You can’t throw this much [data-center] capacity at the system and not have some degree of fossil resources to support it.”

The result, WSJ says, is “a four-way battle among electric utilities trying to keep the lights on, tech companies that like to tout their climate credentials, consumers angry at rising electricity prices and regulators overseeing investments in the grid and trying to turn it green.”

Try to pick the winner of that.

Many are hoping that A.I. can help solve some of the power problems it is causing. Rand reported:

Overall, it could optimize energy consumption to reduce waste while improving efficiency and comfort levels. AI could also better forecast energy demand and supply, allowing energy providers to adjust their production and distribution to increase flexibility and reduce the risk of blackouts. AI tools could open new ways of interacting within the electricity grid, such as the dynamic charging and discharging of electric vehicle batteries.

Furthermore, AI could help integrate various renewable energy sources into the grid.

Similarly, Javad Mohammadi, a University of Texas assistant professor, said: “If you were to coordinate all these devices, that would give you an army of intelligent devices that could talk to each other and respond to emergencies. If the grid is in a bad situation, and there is a need to reduce energy use, all the different devices could coordinate among themselves to reduce their energy usage.”

Geri Richmond, undersecretary for science and innovation at the Department of Energy, believes: “We have the opportunity to build a power grid that’s better and cleaner with the help of AI.”

We have an early 20th century energy grid infrastructure that was already struggling to integrate early 21st century renewables technologies now being faced with the mid-21st century technology of A.I. Something has to give.

Perhaps some of those engineers working on A.I. should focus instead on reinventing the power grid.



Kim Bellard

Curious about many things, some of which I write about — usually health care, innovation, technology, or public policy. Never stop asking “why” or “why not”!