Press the heart button on this article, yes you! I would greatly appreciate it :)
Nvidia has been the darling stock as of late, at one point contributing to nearly 4% of the S&P 500 price appreciation. The success of Nvidia is well earned as the stock has been growing revenues and earnings in great excess of estimations from a few years ago. While the stock may be overvalued on a historical basis, it is poised to take advantage of the latest trends and on the forefront of computational technology development.
Artificial intelligence (AI) like the internet, is poised to change the way we interface with the world and provide productivity boosts to certain sectors of the economy. Like the lightbulb, computers, cars, and others, AI has been met with criticism and worry. Some are worried about AI overtaking human intelligence and revolting, some are worried about rapid job loss from human replacement, while others are skeptical of its energy consumption.
In one of my more popular articles, Keep AI On Energy, I wrote about the concerns of AI energy consumption. Some estimates put it near Bitcoin in terms of energy consumption within a few years. In the grand scheme neither are all that much when compared to some other sectors. As with Bitcoin, the argument stems from a deeper moral conundrum of whether we should be using energy for these new things that require lots of energy. I say that the human spirit of innovation should not be stunted arbitrarily, but that doesn’t stop others from putting the energy use in a bad light, even in large publications like the Wall Street Journal.
The revolution in artificial intelligence may soon require more electricity than all electric vehicles combined. - WSJ
The conundrum is how much carbon emissions are justified given the complicated analysis of how beneficial AI will be. If AI is amazing, then perhaps people will be happy burning coal to get it. If it is just not as amazing as it is hyped up to be, then people may demand that its energy use be constrained to only essential items or it has to be powered by cleaner sources. Many view burning fossil fuels as immoral and as an extension, even if AI is a miracle it is not worth the emissions to the planet. Luckily there are solutions to both these problems available to us today and some companies are already looking at ways to solve them.
Nuclear energy is clean and provides technological benefits over renewable energy sources, chiefly location and energy density. With the need to provide consistent energy to servers to keep them online and provide a solid product to consumers, companies are beginning to look to nuclear energy as a solution. With most top companies having some sort of environmental target these days, options are limited for reliable electricity. Wind, solar, and batteries may be too expensive or not provide the power needed for large-scale data centers for artificial intelligence while real estate and infrastructure limit availability near large hydroelectric facilities. Nuclear energy has the best capacity factors lending itself to the reliability companies are searching for with little concern about the amount of power required.
Google, Microsoft, Apple, and Amazon are among the companies looking to leverage AI to their advantage through their products. CEOs from these companies and other companies are moving to nuclear over time. Microsoft founder Bill Gates has been supportive of advanced nuclear energy, as evidenced by his share of TerraPower. Now, Amazon Web Services (AWS) has acquired a large data center in Pennsylvania which sits directly next to a nuclear power plant. Talen, the owner of the nuclear plant can provide long-term contracts to provide power to AWS and others. A neighboring site will also support a crypto-mining data center powered by the same nuclear plant.
This is not a one-off thing, OpenAI founder Sam Altman told Bloomberg we need a power breakthrough and that he wishes the world was more pro-nuclear. He even has a stake in Oklo, a nuclear power company. Microsoft has invited billions in OpenAI and has a nuclear vision of its own. The company recently hired a director of nuclear energy to work alongside the company’s head of nuclear development acceleration. This comes after the company made small module reactors a part of its company strategy last year. Microsoft has also procured credits from a Canadian energy firm and another in Virginia to power its data centers with nuclear energy.
Energy companies are eager to get in on the trend as well. Constellation Energy's CEO is saying that the AI boom needs nuclear for reliable electricity. With nuclear assets in the US, he is certainly talking his book, but even the US government is thinking about urging companies to look into advanced nuclear reactors for their data centers.
It is no surprise AI companies are attracted to nuclear power. With a capacity factor in the 90s which is far superior to any other source, reliability is without a doubt. This is not the same for wind or solar which are the other options for these companies who seek to reduce their emissions. Battery storage hasn’t been the savior for renewables yet, and large operations would be required for the energy-intensive AI data centers. Whether small module reactors or other advanced nuclear reactors can compete on the commercial scale has yet to be seen. In the meantime, major companies are already leaning on operating traditional reactors for their new power-hungry data centers. While the popularity of nuclear energy has been on a resurgence around the world, AI could be a spark that ignites the trend here in the US. Until next week,
-Grayson
Leave a like and let me know what you think!
If you haven’t already, follow me at TwitterX @graysonhoteling and check out my latest post on notes.
Socials
Twitter/X - @graysonhoteling
LinkedIn - Grayson Hoteling
Archive - The Gray Area
Let someone know about The Gray Area and spread the word!