Meta Accumulates $825 Billion in AI Commitments by 2036

When Mark Zuckerberg went all-in on the metaverse in 2021, the market punished Meta with a drop of over 60% in its stock. Now, the bet is on artificial intelligence, and the check is considerably larger. The company revealed in a regulatory filing that it has accumulated approximately $825 billion in future contractual commitments, most of which are related to the computational infrastructure needed to compete in the AI race.

The figure is impressive not only for its absolute size but also for the speed of escalation. In July of this year, the reported amount was close to $700 billion. In just three months, Meta added over $125 billion in future obligations. For comparison, the annual GDP of countries like Portugal or Chile hovers around this quarterly increase.

How Meta's $825 Billion is Divided

The commitments are divided into two major categories. The first involves $355.7 billion in operational and financial lease contracts that have not yet been recorded on the company's balance sheet. These agreements cover leasing of data centers and energy infrastructure expected to begin between the end of 2026 and 2036.

The second slice, amounting to $454.5 billion, corresponds to short- and long-term contracts that cannot be canceled. This amount includes the purchase of cloud computing capacity from third parties, power supply agreements for data centers, and hardware aimed at the Reality Labs division, which handles the company's virtual and augmented reality products.

There is also an additional commitment of up to $14.7 billion in cloud computing capacity over five years, a value that can be reduced if the supplier can reallocate part of this capacity to other clients. The strategy reflects a company that does not want to rely solely on its own infrastructure, diversifying its sources of computational power to train and run AI models at scale.

Strong Revenue, But Market Looks at Future Costs

Meta's immediate financial context does not warrant alarm. Quarterly revenue reached $64.6 billion, surpassing analysts' expectations, which projected $63.34 billion. The digital advertising business continues to generate robust cash flow, supported by a base of billions of users across Facebook, Instagram, and WhatsApp.

Still, the market's reaction to the results was negative. The explanation lies precisely in the escalation of capex. The company raised its investment projection for the year once again and announced a new class of non-voting shares. This structure could be used in the future to finance acquisitions and cover expenses without diluting Zuckerberg's voting power, who maintains almost absolute control over the company's strategic decisions.

The dilemma for investors is clear: Meta generates enough cash to support the bet today, but the commitments extend over a decade. If generative AI does not translate into proportional revenue, the impact on financial results could be severe in the coming years.

The Billion-Dollar Race of Big Techs for AI

Meta is not alone in this escalation. Alphabet, Microsoft, Amazon, and Apple have also drastically increased their investments in AI infrastructure over the past two years. What differentiates Meta is the concentration of revenue in digital advertising, while competitors like Microsoft and Amazon have cloud divisions that already directly monetize the demand for AI computing from corporate clients.

The central question for Meta is to demonstrate how models like Llama, its family of open-source AI, convert into revenue. So far, the main visible application is in optimizing content recommendation algorithms and personalizing ads. Generative AI tools for content creators and advertising companies are beginning to appear, but they still do not represent a separate and material revenue line on the balance sheets.

Energy supply is another relevant bottleneck. Large data centers consume electricity equivalent to entire cities, and Meta has signed long-term contracts to ensure stable energy supply. This is one of the largest costs embedded in the $825 billion and one of the most difficult to renegotiate, given the global energy transition scenario and the growing demand for renewable sources.

What This Means for Industry Observers

The volume of commitments from Meta illustrates a structural transformation in the technology industry. Big tech companies are essentially mortgaging a decade of future cash flow on the premise that artificial intelligence will be as transformative as the internet was in the 2000s.

For investors, the point of concern is not whether AI will have an impact, but whether the return on these investments will be proportional to the capital allocated. Meta generates approximately $260 billion in annualized revenue. Committing more than three times that amount in future obligations creates an asymmetry that can only be justified by significant revenue acceleration in the coming years.

The new class of non-voting shares adds another layer of complexity. If the company needs to raise additional capital in the market without diluting Zuckerberg's control, minority shareholders may see their economic stake shrink without any power to contest it. This is the type of structure that works well when the bets pay off and becomes a source of friction when results take time to materialize.

The jump from $700 billion to $825 billion in three months suggests that the investment acceleration phase has not yet peaked. For the technology ecosystem as a whole, including chip suppliers, data center builders, and energy companies, this demand represents a concrete opportunity. For Meta, it is the largest bet in the company's history, even surpassing the controversial rebranding to the metaverse, which ironically now shares space with AI within the same Reality Labs division.

This content is informational and educational and does not constitute investment advice. Past performance is not indicative of future results.