The AI data center arms race is still escalating! Can Meta's additional $40 billion bet extinguish the argument of excess computing power?

According to Zhitong Finance APP, Meta Platforms (META.US) has committed to an additional investment of $40 billion in its massive data center campus in Louisiana, bringing the total expected investment at the site to over $250 billion as the company continues to expand its artificial intelligence computing landscape.

On Monday, Meta announced that it would expand the computing capacity of this project located in rural Louisiana to at least 5 gigawatts, at a cost of $50 billion. Previously, the company had announced a $10 billion investment in the data center and surrounding communities.

In May, media reports indicated that Meta planned to add $200 billion in spending on the project, primarily for the purchase of expensive computing chips to be housed in the nearly 4,000-acre campus. According to an anonymous insider, this brings the expected total for the site to at least $250 billion.

In addition to the $50 billion, Meta has not publicly disclosed any other costs associated with the project.

CEO Mark Zuckerberg has made aggressive investments over the past two years to build infrastructure such as data centers, which he believes are crucial for achieving superintelligence in artificial intelligence. The company currently has 33 data centers that are either completed or actively under development, and Zuckerberg has committed to investing at least $600 billion in infrastructure projects in the U.S. over the next few years. Last week, Meta pledged to spend $10 billion to build its first data center in Canada.

The data center in Richland Parish, Louisiana, is Meta's largest and most ambitious project to date, and the high costs have prompted external investors to seek additional financing. Blue Owl Capital Inc., which owns 80% of the project, has sought billions of dollars from Wall Street to help support the construction. Meanwhile, Entergy Louisiana is spending billions to build 10 new gas power plants to supply electricity to the campus. While the data center will utilize 5 gigawatts of computing capacity, over 2 gigawatts will be used to meet the broader power needs of the campus.

In an interview last week, Zuckerberg stated that Meta is still seeking all the computing capacity it can obtain. However, the founder is also considering whether to lease some of the capacity to external parties, as media reports earlier this month indicated that Meta is developing a plan for a cloud infrastructure business. Discussions include plans to sell access to "native" computing capacity, similar to other so-called "new cloud" businesses (such as CoreWeave Inc.).

On Monday, Meta stated in a press release that the Richland Parish data center, once operational, will provide 1,000 jobs, doubling its previous employment commitment. The company also noted that since breaking ground at the site in December 2024, it has awarded over $1.6 billion in contracts to businesses in Louisiana

Can a $250 billion investment alleviate the "overcapacity" theory?

From the current market dynamics, Meta's $250 billion "astronomical" investment serves as both a strong boost for the bulls and a wake-up call that deepens the anxiety of the bears. It cannot completely "eliminate" concerns, but it successfully shifts the focus of market debates from "is there an overcapacity in computing power" to deeper propositions.

Previously, the market was most afraid of "tech giants being unable to move" or "the AI boom being just a temporary impulse." Meta's additional $40 billion, pushing the total to $250 billion, counters this with the most hardcore financial commitment:

This proves that tech giants' demand for top-tier computing chips, high-speed storage (such as SanDisk and Micron's enterprise SSDs), and advanced packaging (such as TSMC) is not a temporary peak but a long-term infrastructure necessity.

Zuckerberg's revealed "computing power leasing" plan dispels market doubts about Meta's ROI from "buying so many chips just to sell ads." If computing power can be leased to society like oil, then as long as the spark of global AI startups does not extinguish, Meta's inventory will always have buyers. From this perspective, it greatly alleviates the valuation crisis in the hardware supply chain.

On one side is high demand, while on the other is terrifying costs. Although Meta's investment proves that there is no overcapacity in computing power, it opens up a new battlefield of concerns:

Wall Street is beginning to doubt whether, even if there is no overcapacity, this "frenzied arms race" among giants will drag tech stocks down to become high-debt "traditional utility companies" or "real estate developers."

Even a strong player like Meta, facing such a massive funding gap, has to bring in Blue Owl to Wall Street for structured financing worth billions. This reliance on external leverage has made risk-averse analysts wary of potential financial structural pressures.

This investment successfully alleviates the physical anxiety of a short-term slowdown in the hardware supply chain (foundry, chips, high-end storage), proving to the market that super computing power remains the industrial blood of the digital economy.

However, it has not cured Wall Street's spiritual exhaustion regarding AI monetization (ROI). It has transformed the question from "why buy so many chips" to "after leasing 10 power plants to burn cash, when can we recoup the first $250 billion?" This bet between analysts and tech giants is likely to reach a watershed moment in the upcoming August earnings season