Generative AI is stupid. Is Yann LeCun’s opinionone of the godfathers of the artificial intelligencewhich has grown tired of how the AI majors seek AGI and it seems that he is going to set up his startup to achieve it. To make AI more “smart” you have to train it, and for that you have to build data centers. And boy is it being done.
To the point that there are already those who calculate that the rise of AI threatens the plan of reindustrialization of the United States.
AI walks or doesn’t walk. The United States has a plan: invest whatever it takes to achieve superintelligence before China. China is also investing, but while what it seeks is a cheap and functional AI to monetize nowwhat the US wants is artificial general intelligence, or AGI. That costs money and, above all, investment in huge data centers.
One of the Donald Trump’s election promises During his two campaigns he orbited around the commitment to return millions of jobs to Americans. To achieve this, the opening of new factories on national soil through tax incentives and an “America First” policy that we have seen echo in the rest of the world in the form of tariffs.
Capital redistribution. ANDfactories are opening and reopeningbut perhaps not as many expected. In Bloomberg They point to a devastating fact: spending on new data centers has increased by 18% in the last seven months. This is a colossal increase, but it goes hand in hand with another fact: spending on new factories has fallen 2.5% this year.
While large technology companies are committed to building data centers, the policies of recent months, immigration restrictions, withdrawal of support for electric vehicles and tariffs are generating uncertainty in the market that slows down investment aimed at opening other types of factories.
Not only are factories not opened, but they are laid off. American manufacturing heavyweights are not only facing the biggest corporate tax hike since the 1990s, but are estimated to have lost 38,000 manufacturing jobs this year. Mostly in sectors such as electronics, automobiles or household appliances. In August alone, 12,000 people lost their jobs (and why don’t we include those from the video game industry here…).
brutal difference. Estimates suggest that the monthly spending of manufacturing plants will situates at $18.8 billion, but while the trend is downward, if we look at spending on AI, we see a radically different scenario. Among the big four technology companies (Amazon, Microsoft, Meta and Alphabet) $400 billion will go to AI infrastructure in 2025 alone.
This is an increase of 60% compared to last year and it is not a peak: it is something sustained. In fact, the investment in 2026 is expected to be higher. There are other companies with their own plans, such as OpenAI what is the most valuable private company and can afford lose 11.5 billion in just the last 90 days which is making an investment of between 400,000 and 500,000 million dollars between 2025 and 2027.
{“videoId”:”x9sjece”,”autoplay”:true,”title”:”CHINA is WINNING the TECH WAR because they planned it that way 10 YEARS AGO”, “tag”:”china”, “duration”:”721″}
Help Uncle Sam. This AI boom is driving other directly linked sectors, such as the construction of the data centers themselves (someone has to build them as long as they do not use already manufactured facilities) and that of energy. Because these facilities need ridiculous amounts of energy to runso much so that Google wants to take them to space and China is submerging them in the sea to spend less on dissipation.
Thus, reopening nuclear power plants or investing in modernizing gas turbines to supply data centers is on the horizon, but it is still something that does not impact the American worker, they are not new factories that need personnel. And part of the money needed is coming from the state itself.
Recently, AMD announced that the United States Department of Energy had allocated 1,000 million public to power the infrastructure. And both OpenAI and NVIDIA have dropped the need for the United States to get involved to sustain this new industry, which is already awakening bubble feelings.
Echoes of the 2008 blow. When we talk about such astronomical figures, it is very difficult to get an idea. It was already happening with the 70,000 million dollars that Microsoft paid to take over Activisionand if we now go to amounts of 400,000 or 500,000 million, things are going to get worse. What is evident is that, as we say, these investments fly over the fear of the bubble bursting.
If in July of this year 37% of fund managers believed that we were facing a bubble, in October the figure increase up to 54%, although from the technology industry itself It seems that there is no one who brings sanity. Because it is spending a lot, a lot, more than during the dotcom era which did not end too well for many, and even figures as interested as Mark Zuckerberg, CEO of Meta, have commented that, while it is true that many are oversizing their investments, it is better than being left behind.
Only time will tell how everything turns out, obviously, but the article Bloomberg It closes in a quite interesting way. Arno Hill, former mayor of Lordstown, a municipality where there was a large GM plant already closed and which is now part of SoftBank and Foxconn’s plans to create a data center, says that he does not know what will happen with AI, but that people will always need cars.
Image | Google Data Centers
In Xataka | The world of AI has a problem: there is no energy for so many chips
(function() { window._JS_MODULES = window._JS_MODULES || {}; var headElement = document.getElementsByTagName(‘head’)(0); if (_JS_MODULES.instagram) { var instagramScript = document.createElement(‘script’); instagramScript.src=”https://platform.instagram.com/en_US/embeds.js”; instagramScript.async = true; instagramScript.defer = true; headElement.appendChild(instagramScript);
was originally published in
Xataka
by
Alejandro Alcolea
.
GIPHY App Key not set. Please check settings