There’s plenty of good reasons to like certain aspects of both AI and clean energy. Clean energy promises a less polluted planet, especially from emissions, and to help mitigate the dangers of global warming. AI promises significant increases in productivity and efficiency along with potentially a wider access to important information. I am certainly in many ways a supporter of both approaches. However, I’ve warned in earlier posts that there are higher level reasons to worry about too fast a transition to clean energy leading to societal collapse and that AI is potentially a stock market bubble. A key reason for the fast rollout of AI is that some of the elite believe it will provide an incredible social control mechanism, which I have argued in other posts is debatable given human nature.
In the US and Europe, at least, any expansion of clean energy usually weakens the political right, since fossil fuels have tended to support more conservative political factions as a rule. Also, both AI and clean energy require greater centralization of social power and control in that they require greater oversight, are more intrusive, and tend to favor collectivist not individual solutions. They are thus in each case more consistent with the modern surveillance state.
There is also definitely the incentive to expand both exponentially given how much money can be made from a quick rollout of these areas. Each requires tremendous allocations of resources on huge scales where billions of dollars are nothing and this rollout includes enormous amounts of fiscal spending of taxpayer dollars. Both will, no doubt, require extensive regulation from above. In the case of AI, there is also the fact that this narrative is a key to highly escalated and in many ways unrealistic stock valuations. At least as of the last quarter’s economic results, it has not led to the expected productivity gains its advocates have predicted.
However, despite all these reasons some good and some not so good for the rapid expansion of both areas, there is in fact an overwhelming problem here that is being largely hidden from the public. AI requires huge amounts of energy, amounts that are so vast and nation-scale in size that there is no way on Earth that the clean energy global carbon emissions goals can possibly be met and the scale of the AI rollout envisioned also be carried out. The contradiction between these two goals is such a big deal, in fact, that it creates potentially a gigantic problem for the elite and society as a whole. If one goes back historically, it’s hard to find another example of when the elite has pushed so hard two fundamentally contradictory things at one and the same time. As I suggested in an earlier note, are we in some strange end of history period where the elite can say and support ideas that literally totally contradict one another and either not care or not recognize the problem with their position?
Those experts who have looked at the problem of energy usage by AI point out among other things that simply changing Google’s regular searches to generative AI searches would require as much power usage as the entire nation of Ireland each year. To these staggering amounts one has to add 50 per cent for the cooling systems. If one changed all the current data centers to AI centers there would be a 1000 percent increase in energy use for this source from current levels. Currently, data centers use 1 per cent of all global electricity—-would that then jump to 10 percent?
The International Energy Administration tells us data centers, it believes, currently use 1-1.5 percent of global electricity. If you add crypto, it’s 2 percent. By 2026, it will be 4 percent. That’s equivalent to the entire energy use of the advanced Japanese economy. One study concluded 8-21 percent of global electricity use would be from computers by the year 2030 and much of it would be from AI. One researcher noted that going from non-generative to generative AI takes 30-40x more energy. Billions of computer devices will already make up 3.5 percent of global carbon emissions by 2025. All of these problems are compounded by the fact that the huge data centers necessary for generative AI require reliable energy sources, so they tend to be placed in areas where energy is very predictable but from fossil fuels. One could go on, but hopefully just these few statistics suggest the magnitude of the problem.
Now as I suggested in an earlier note, some people believe technology will save the day and make it possible for both clean energy and AI to expand simultaneously. Those technologists that believe clean energy will somehow magically be able to more quickly replace fossil fuels believe that is going to happen. But they are making a key conceptual error. Computing power may follow Moore’s Law and double in capacity for the same cost every 2 years, but energy follows no such law. In the last 50 or so years, computing power has gone up by a factor of over a billion. Incredible, isn’t it? Guess how much over the same period a solar cell has increased its energy output? Only by a factor of two. That’s a really big problem if your goal is to use only clean energy for the AI expansion.
Another possible clean energy source for this huge increase in energy use is nuclear. But nuclear has never been particularly liked by the left and will not be easy at all to get going and expand dramatically and takes a very long period to build out. In fact, the center of the global clean energy political faction, the Green Party, got Germany to close its nuclear plants as a key part of its political agenda. It will not be an easy sell to get the publics of either the US or Europe to get beyond the historical legacy of Three Mile Island, Chernobyl, and Fukushima.
Much more likely is that more fossil fuels use will be necessary to accomplish this goal of the AI rollout. For over 150 years, fossil fuels have powered our civilization and even now despite all the efforts at clean energy development in recent years the total percentage of energy from fossil fuels is still over 80 percent. If you go to the US Energy Information Administration, the part of our government that concerns itself with these kinds of things, fossil fuels they conclude will still in 2050 make up a huge amount of overall energy needs, more than two-thirds in both OECD and non-OECD countries.
Since the best analysis by the US government itself says that more than two-thirds of energy use will still be fossil fuels by 2050 that huge an increase in energy demand required for AI more or less guarantees no real reduction and probably even a quite significant increase in fossil fuel use and resulting carbon emissions. If no such increase is to occur, then what other energy need of society will have to be cut to fuel the AI expansion? I wrote in an important earlier post about Tainter’s advanced research on societal collapse. In his work, he shows that societies tend to collapse if the increase in complexity outstrips the energy needed to support that increase. How is such a vast increase in complexity as involved in AI possibly going to succeed without significant increases in fossil fuel use?
Now, politicians can say anything they want on paper but there’s still reality. The G7, consisting of the US, UK, Germany, Japan, France, Italy, and Canada recently made a historic agreement to eliminate all coal usage for the sake of lower carbon emissions by 2035. Great idea, who wants more emissions. The problem though is that coal currently makes up 15 percent of electricity production in those countries. In Germany it makes up 25 percent, and in Japan no less than 29 percent. How on Earth can you possibly meet those goals, if you’re expanding AI to the extent the very same elite is talking about?
As one commentator noted, if you eliminate coal you will have to rely on ever more gas use since you can’t fill the gap with clean energy. The result will be higher gas prices, which will in turn force China, India, and other less wealthy countries to increase not decrease their coal use defeating the entire point of the exercise.
In any case, in both politics and markets narrative is a crucial organizing principle and source of social order. Right now, the AI narrative is the primary support of the stock market gains and clean energy has long been the unifying narrative for global elites with lots of fiscal spending programs. While perhaps you can have both narratives coexisting for a short while, the contradictions between these two narratives will become more and more apparent over time. Keeping both narratives running at full speed is not maintainable given reality—it’s just too big a contradiction. My guess would be that the global elite will ultimately jettison or vastly redefine the clean energy narrative and is already preparing to do so.
In terms of investing, it also suggests potentially a key market contradiction. It is doubtful that the clean energy, and AI emphasis can continue unabated. One of those two sectors is likely to continue to be a source of very high levels of capital investing while the other will have to be reduced in emphasis or both will become less leading sectors.
If I was advising some of these leaders, I would be telling them they need to figure this one out, and it’s going to be very difficult indeed. One of the problems is that both ideas as they are being defined are very millennial and utopian, we are going to make a perfect world like. When ideas that are so utopian come to the forefront politically, often society can continue for quite some time before finally confronting reality. It is hard for politicians who often use these ideas to get elected or anyone else to stand up against things which at least on paper sound great. Again, I am not opposed to either area, but I fear that if we do not take reality seriously enough, solutions to this huge contradiction may simply not be found and an unrealistic extreme rollout risks failing where a more thoughtful implementation might have succeeded. Failure if done badly could even be very extreme.
Sir Thomas More once wrote a work he called Utopia about a society where everything was “perfect.” Most serious scholars believe the book was a work of satire since the word utopia itself actually means “nowhere.” It is upsetting though to have, at least on the surface, such scientific ideas as energy and AI become part of what seems to be a millennial dream world that either the people responsible know is a dream world or in which they themselves are caught in the dream.
Disclaimer– the information discussed is simply one person’s opinion nothing more or less. It is only for entertainment purposes. By using this blog, you assume all risks associated with using this advice, suggestions, information, conclusions and everything else contained here-in and that you completely and fully understand that you and you alone are 100 per cent responsible for anything that occurs from using this information and material in anyway whatsoever–regardless of how you interpret any discussion, conclusions or advice contained here-in. Any discussion of actual stocks or investments is in no way a recommendation and is only for educational purposes. You should listen to many competing opinions, consider all the counterfactuals to what is argued, seek out always if necessary professional advice, and of course ultimately make your own decisions about the markets.