Update: AI, Data Centers, & The Demand For Electricity
It seems like just yesterday that we did a story about the Ohio public utilities commission considering a new rate proposal that would require data centers to pay 85% of their expected utility bill every month whether they used as much electricity as anticipated or not. The reason for doing so is to ensure that utility companies they will have enough money coming in to pay for the increases in electricity generation and transmission infrastructure needed to meet the voracious appetite for electricity those data centers will have.
Here we thought we were focusing on a narrow issue that applied only to Ohio, but according to a report by Bloomberg this morning, the topic of AI and data centers is being discussed at the Future Investment Initiative, also known as Davos in the Desert, in Riyadh, Saudi Arabia, this week. Those in attendance at the Ritz-Carlton convention center include Ruth Porat from Alphabet, former Google CEO Eric Schmidt, TikTok CEO Shou Chew, Blackstone head honcho Stephen Schwarzman, and Blackrock’s Larry Fink.
The attendees listened to Masayoshi Son of Softbank as he laid out his vision of superintelligence 10,000 times that of humans by 2035. Elon Musk, in a virtual appearance, said robots would far outnumber humans by 2040. And the Saudis touted plans to make their country an AI superpower, with smart cities and personalized health care. But some of the guests at conference were focused on the potential Achilles heel of the emerging AI sector — the energy hungry data centers that underpin AI, which are already straining electricity grids around the world.
AI Will Lead To A Big Increase In Demand For Electricity
Stephen Schwarzman told the conference the AI boom could potentially propel electricity use to increase by 40% in the next decade. He warned that could slow the development of the technologies, quite aside from its disruption to global economies. In order to cash in on the expected AI boom, Blackstone is building a $25 billion data center empire. “In four years, giant economies, at the rate that expansion is going to happen, are going to stock out of electricity unless there’s either efficiencies in semiconductors and other types of things or you’re going to have to slow down the growth or have a giant expansion,” he said. If you find that statement more than a little confusing, join the club.
Some tech giants like Microsoft, Amazon, and Google are exploring nuclear energy to power their AI projects. Musk, whose autonomous cars and AI chatbot businesses require huge amounts of energy, signaled he was looking at other sources. “We need a lot of energy,” Musk acknowledged in a livestreamed conversation. “In the long term, most of the energy that we’ll get is going to come from the sun.”
The discussion in Riyadh comes as Saudi Arabia looks to pivot away from oil and diversify its economy, with an emphasis on technology and AI. “Data centers need a lower cost energy,” Saudi Arabian Oil CEO Amin Nasser told the gathering. He then invited businesses to build data centers in his country, where they could expect to pay as little as 4.8 cents per kilowatt-hour for electricity generated by burning methane gas or 6.8 cents per kilowatt-hour for electricity from renewables. “You are going to have the lowest cost energy to build data centers anywhere in the world because this is the place for it.”
Eric Schmidt agreed. “There’s every reason to think that Saudi in particular can become one of the big winners here. The simplest thing for this country to win is to use its abundant resources and create facilities that no one else can build. There is a huge shortage of electricity in the developed world.”
The AI Conundrum
The message in Riyadh this week is simple. Flood the atmosphere with even more greenhouse gas pollution so humans can make cool memes and have computers write their term papers for them. Yeah, that makes sense, right? We hear about the renewable energy revolution and how it will help decarbonize the environment but what we didn’t anticipate is that people would go gaga over artificial intelligence, which consumes ten times more electrical power than the good old fashioned ordinary intelligence we are used to.
Our email inbox here at CleanTechnica is overflowing with messages from people who want to tell us how AI will clean up emissions from cement, steelmaking, plastic recycling, or the manufacture of batteries. That is all very exciting, but the reality is that the rush to power data centers is leading to reactivating aging nuclear facilities and diverting the vast majority of new renewable energy away from meeting the needs of ordinary people and businesses so data centers can gobble it all up themselves. The newest solar farm in Texas has a nameplate capacity of 875 MW, but only 15% of that will be available to ordinary utility customers. 85% of it is going to power Google data centers in and around Dallas.
CleanTechnica readers are all above average — it’s a scientific fact! — and some of the comments made on our story about data centers in Ohio are worth reprinting. Here’s one example of superior thinking: “AI and data centers are the near term energy ‘hogs’ of which we are currently aware, but they will not be the last ‘must have’ technologies that spur future energy demand growth. I believe that very few experts in the field appreciate how much growth in energy demand we will have, and grossly underestimate the amount of wind and solar must be added over the next 30 years.”
We couldn’t agree more. There are several well known corollaries. One is that your financial needs will always exceed your income by 10%. Another is that your stuff will always need 10% more space than you have room for. The third is that you will always need 10% more time than you have available. Is it wise to operate on the principle that the electrical supply will always expand to meet the need or would it be better to put rational limits on how much electricity there is available?
Here at CleanTechnica we shudder at the idea that Saudi methane should be gleefully embraced as the way to power data centers. Just a year ago at COP 28, the Saudi oil contingent all looked like someone had just shot their dog when the conference dared suggest the world needed to move on from fossil fuels.
The Takeaway
Here’s another pearl from the comments section: “Google should give us a choice as to whether we want AI in our search results. I for one do not need a dumb C- student to do a Cliff’s Notes summary of my search results.” Right on, brother. We can’t help but wonder what the reaction would be if EV advocates suggested we should restart old nuclear reactors and build new coal-powered generating stations to provide all electricity the electric cars coming soon to the world will need. Oh, the weeping, wailing, and gnashing of teeth that would cause! And yet hardly anyone questions doing the same so people can share more videos on TikTok or so Google can review every document ever created when we ask what the time is now in Kuala Lumpur.
Humans have a problem with limits — they don’t want any. Everything should be free to expand infinitely forever. There are organisms that do that; they are called viruses. (Fans of The Matrix already know this.) They consume everything they can find until they completely devour the host organism, then they both die. Does anyone see a parallel here?
Chip in a few dollars a month to help support independent cleantech coverage that helps to accelerate the cleantech revolution!
Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.
Sign up for our daily newsletter for 15 new cleantech stories a day. Or sign up for our weekly one if daily is too frequent.
CleanTechnica uses affiliate links. See our policy here.
CleanTechnica's Comment Policy