Meta Gears Up For Super Artificial Intelligence

CEO of Meta (META) Mark Zuckerberg announced that we are getting closer to Super AI.

It means his team is making progress.

Meta is infamously playing catch-up after a botched Llama 4 model didn’t make the cut.

Now, Zuck is spending big bucks and plans to spend $72 billion in infrastructure this year related to AI.

In mid-2025, Mark Zuckerberg announced Meta’s ambitious push toward achieving super artificial intelligence (SAI) through the creation of Meta Superintelligence Labs (MSL), a new division consolidating the company’s AI efforts.

Zuckerberg has committed “hundreds of billions of dollars” to build massive AI infrastructure, including multi-gigawatt data centers like Prometheus, set to come online in 2026, and Hyperion, scalable to 5 gigawatts.

These facilities are designed to provide industry-leading compute power, critical for training advanced AI models. Additionally, Meta has aggressively recruited top talent, offering multimillion-dollar compensation packages—some as high as $100 million—to poach experts like Alexandr Wang, former Scale AI CEO, and Nat Friedman, ex-GitHub chief.

Zuckerberg’s vision, articulated in a June 2025 memo, emphasizes “personal superintelligence for everyone,” suggesting applications in Meta’s core products like Facebook, Instagram, WhatsApp, and emerging areas like the Meta AI app and smart glasses.

Meta’s advertising business, which generated $165 billion in 2024, could leverage SAI to enhance ad targeting, content personalization, and user engagement, driving revenue growth.

The massive capital expenditure—$64 billion to $72 billion projected for 2025 alone—raises concerns about profitability, especially if SAI development timelines slip or fail to deliver commercial results.

For instance, AI capable of self-improvement could analyze user behavior across Meta’s platforms (Facebook, Instagram, WhatsApp) with unprecedented precision, optimizing ad delivery to maximize engagement and conversion rates. This could increase ad revenue per user, particularly in markets where growth has plateaued. Additionally, SAI-powered content creation tools could enable advertisers to generate hyper-personalized campaigns at scale, potentially introducing premium ad services as a new revenue tier.

For example, a premium Meta AI subscription might offer users SAI-powered productivity tools, personalized education, or creative content generation.

I don’t believe that AI stocks will go up in a straight line. Just take for instance, if there is a massive recession, then that will throttle the momentum by a few years if not more.

We are on the brink of a recession, and we have been for some time.

The $72 billion in infrastructure spend for this year appears reckless, almost as if the deep-pocket tech companies are throwing everything they can at the AI business and see what sticks.

It doesn’t seem like this is too strategic, along with poaching top engineers for $100 million per year pay packages.

A lot of this smacks of FOMO (fear of missing out), and instead of building AI in an orderly fashion, Meta is going for the winner-takes-all as soon as possible strategy.

Then there is the case that AI integrated into its marketing is a total bust, and people can’t stand it.

There are a lot of risks, and the amount spent is crazy.

Any type of measured approach has been thrown out of the window.

However, the stock keeps cheering AI like it’s a guaranteed revenue generator when it is anything but.

There will be an event that causes a huge pullback, and readers should be aware of it, but until then, the pain trade is higher in all premium AI stocks.