Essay by Eric Worrall
“… While American AI researchers struggle with a fragile power grid, Chinese developers now treat energy availability as a solved problem. …”
Here’s what’s happening to electricity prices
PUBLISHED FRI, NOV 14 20252:49 PM ESTUPDATED MON, NOV 17 20251:26 PM EST
Spencer Kimball @SPENCEKIMBALL
Gabriel CortésElectricity prices are surging, voters are growing angry, and the artificial intelligence industry’s data centers are increasingly a target for blame with U.S. mid-term elections on the horizon.
Residential utility bills rose 6% on average nationwide in August compared with the same period in the previous year, according to the U.S. Energy Information Administration.
…
The tech companies and AI labs are building data centers that consume a gigawatt or more of electricity in some cases, equivalent to more than 800,000 homes, the size of a cityessentially.
Virginia has the highest concentration of data centers in the world. Democrat Abigail Spanberger won the state’s recent governor’s race in a landslide by campaigning on cost of living. Spanberger put at least part of the blame for rising electricity prices on data centers, promising to make tech companies “pay their own way and their fair share” of the escalating costs.
…
Read more: https://www.cnbc.com/2025/11/14/data-centers-are-concentrated-in-these-states-heres-whats-happening-to-electricity-prices-.html
Meanwhile in China;
How China’s Energy Supremacy Threatens U.S. AI Dominance
JACK MURAWCZYK
October 17, 2025 . 10:02 AMThe race for Artificial Intelligence (AI) supremacy is, fundamentally, an ever-escalating energy war. China’s centrally planned, dual-track “AI-Energy” policy is establishing a critical, long-term infrastructure advantage that threatens the U.S.’s technological leadership, forcing the West to abandon strict free-market ideology and adopt its own form of strategic state capitalism to compete.
The current limiting factors for AI development – the scarcity of compute and of world-class talent – are increasingly being overshadowed by a far more foundational constraint: energy. It is on this crucial frontier that China holds a decisive advantage.
…
China has launched what amounts to a Manhattan Project for energy infrastructure, deploying a centrally planned, dual-track strategy that grants its tech sector a decisive edge in the global AI race. While American AI researchers struggle with a fragile power grid, Chinese developers now treat energy availability as a solved problem. China’s energy strategy is characterized by aggressively deploying clean power for future dominance while expanding coal and oil production to ensure immediate stability. This has resulted in unprecedented energy construction. China currently has 32 nuclear reactors under development. The U.S., by contrast, has built two since 2014. China’s solar manufacturing capacity exceeds 1,000 gigawatts. The U.S. capacity stands at 26 gigawatts.
Read more: https://stanfordreview.org/how-chinas-energy-supremacy-threatens-u-s-ai-dominance/
…
The development of an artificial general intelligence will be a world changing event on a par with the development of the atom bomb.
Most people have tried to play that chess app on your computer, and experienced the humiliation of a few lines of code kicking your butt. Imagine that chess app AI intellectual butt kicking capability extended into every sphere of human experience, and you see what I’m concerned about.
Of course reaching the prize is not the only benefit of an AI Manhattan project. Just as the US Manhattan Project, like the Space Race, produced spinoff benefits at every advance, every Chinese advance towards super intelligence will allow more of the economy to be automated, allowing Chinese exporters to cut prices, further undermining the economies of other nations.
And that is in addition to the possibility of AI powered military devices which can outperform any human soldier. We’ve already seen a taste of battlefront AI in the Ukraine war, where Ukraine’s AI driven targeting has defeated Russian drone jamming technology, by allowing kamikaze drones to operate without a human control signal.
This is not a race anyone should want to lose.
How close are Chinese researchers to developing artificial general intelligence? It may prove to be a problem as elusive as nuclear fusion. Or it might be solved tomorrow. Having said that, I doubt it will be as difficult as fusion, because researchers already have a working template to copy – the human brain.
One thing for sure. In terms of the current most critical limiting factor, affordable energy, China is way ahead of the USA in the push to develop the world’s first super intelligence. If the USA wants to have any chance of winning this race, US energy supply and affordability must be fixed, before the growing voter backlash against skyrocketing energy prices pulls the USA out of the race.
Developing the technology alone is not enough. The USA didn’t start the current AI push. The University of Toronto developed the critical breakthrough which made the current AI revolution possible in 2012. But aside from some Canadian islands of research excellence, Canada is not a serious player in the AI race. Canada’s obsession with green energy and moribund economy gave the USA the edge – an edge the USA is now in danger of losing to China.
Some US leaders are moving in the right direction on this issue, but there are still enormous obstacles which must be overcome. So long as AI development hurts ordinary people by driving up energy prices, AI development in the USA will be in danger of stalling.
The issue is admitting/forcing an abandonment of the Green New Deal/Energiewende/Net Zero approach as an impossible to achieve failure.
We are talking about three different but related topics here. One energy, two CAGW and three AI. Like it or not clean, safe and affordable energy is number one no question. We have wasted trillions of dollars because of the false belief in number two CAGW. Stop doing that and number one is pretty easy, we know how to produce clean, safe and affordable energy. That leaves number three, AI. AI is inevitable it will be a good thing and it will be a bad thing. Our job is to recognize the bad and do our best to prevent it. That leaves the good. The good presents two problems. Number one they use so much energy that they could raise the cost of energy for the rest of us. Like it or not we live in a regulated society, light regulation is a good thing. We control who has access to the grid. I want AI to have access to all the energy they need. For that privilege they can help with the generation and transmission needed. They have mighty deep pockets. Number two is the race with other nations and the perceived threat from other nations. If we build generators that work and stop building stuff that doesn’t with help from AI we should be able to achieve all our goals. Remove our road blocks and we are capable of out competing anyone.
At the moment, AI seems to do nothing more than collate stuff produced by others, and throw in the odd hallucination or fantasy to keep us on our toes.
The problem with the majority view it that it is not always correct. AI programs now tend to carry a warning that the information provided might contain mistakes – which rather reduces the benefit of AI, if you need to use natural intelligence to try to figure out how much of the AI output is factual, and how much is not.
Another AI problem is that AI hallucinations are collated uncritically by AI, with the result that the problem will become worse, with AI “slop” gradually becoming the majority data.
Not to worry, hopefully the “experts” guiding the expenditure of hundreds of billions of dollars into AI rather than other things, have thought this through. On the other hand, Richard Feynman said “Science is belief in the ignorance of experts”.
I’ll wait for a bit. Time will tell.
“Time will tell who has fell and who’s been left behind when you go your way and I go mine”
Bob Dylan
AI is much more than pattern matching, though that is part of it.
Yes they produce new sentences, sometimes amazingly.
For example https://www.cfact.org/2025/03/04/ai-emulates-abstract-thinking-about-kipling-lady-gaga-and-the-rolling-stones/
Truly amazing.
China went large on the power sector while Americans and American media slept, as usual. They did the same in many industries, plus dozens of international airport complexes and high-speed rail lines. We built a comparable industrial complex of lawyers, judges, and district courts. And when we attempt to do something large outside the services sector it looks more like the Obama Biden fiascos of uneconomic projects and insider deals. There you have it, shiny versus old and rusty.
And still, the Chinese squat to crap in holes in the floor, and often don’t have toilet paper or they ration it.
I’m skeptical that AI has driven the price of electricity up nationally.
Maybe in a few localized areas, but not everywhere because AI is just getting established.
Trump says AI companies should supply their own electricity so as not to raise electricity prices for the Public.
The Public better keep an eye on how these companies do things in their area. If they just hook up to the grid, it’s going to cost the Public a lot of money.
It looks like people in my area are going to have to start paying attention. Google just announced they were going to build two new data centers in my area. We already have one that located here last year. I suppose the fact that there is a coal-fired power plant in the area might have something to do with their locating here.
I don’t want to subsidize Google. I don’t want my electric rates to go higher because of Google.
I’m curious why the higher price signals aren’t stimulating enough new plant construction.
But the PR problems occur even when the data centers build their own plants, then the complaints are the data centers use too much water or clog up traffic.
I suspect the holdup is public suspicion of AI rather than inability to meet everyone’s power requirements.
A huge amount of new gas generation has been announced. Back around 2000 we brought 200,000 MW of new gas online. Looks like we will do that again.
New plants aren’t being constructed because we have to save the world.
Don’t worry. All will be fine – unless Google find they can’t actually make a profit, and shut the data center down.
Ah well, the surplus electricity can be used to generate heat, which climate scientists assure us can be stored and accumulated. The stored heat can no doubt be sold to people in cold places, like Siberia or Alaska – at a profit!
Problem solved.
Good luck with copying a human brain – nobody has the faintest idea how it does what it does. For example –
I remain optimistic that Nature will continue to thwart our efforts to understand her. We can’t even figure out why the “double slit experiment” does what it does, or why nobody has found an example that disproves the uncertainty principle.
All good fun.
AI will never be able to value or discern truth; it is fully incapable now, and in the future of ever possessing wisdom. It can never appreciate things like beauty, order or grace thus it can’t ever make decisions or be able to filter out evil.
AI will never naturally defend humanity it can only evaluate humanity in terms of materialism. If anything at all, humanity is a competition. Humanity is a consuming danger. Humanity is worth nothing more than the minerals or energy extracted from it.
Religion says that Creation is to glorify God. Any other view devolves into nihilism and the philosophy of Malthusians. AI will never glorify God and can’t be programmed to do so, therefore at best, we get SkyNet and a world looking for John Connor.
Researchers do not need a lot of juice. That is for mass users, an entirely different issue.
For some research, sure. But some research efforts require brute force to try to beat known limits of current approaches, though I guess University of Toronto can rent time from US data centers when they need a boost.
Fusion is the most expensive, and therefore the stupidest, way to generate electricity ever proposed.
No kidding. Everyone knows the future is Matter-Antimatter Reaction moderated by dillithium crystals. Go long on deuterium and antihydrogen.
Helicopters are the most expensive, and therefore the stupidest way, to fly passengers, ever proposed.
As to grid electricity, I don’t believe that any has been produced by fusion, so you may just be confused, rather than mindlessly hammering on your keyboard.