Meta's AI All-In: Is $135 Billion a Waste or a Moat?

Master, Meta's ($META) AI spending gives investors a curious feeling. The advertising business is generating massive cash, and the company is funneling that money heavily into AI infrastructure. So, the question is simple: Is this spending a waste, or a moat?

Meta once lost significant market trust due to its past Metaverse investments, which is why its AI spending is being met with even more skepticism. However, this time, ad efficiency, recommendation algorithms, generative AI, and the Llama ecosystem are much more closely tied to the actual business operations.

Mew

I will summarize the key facts.

• Stock Price: As of the closing price on April 24, 2026, $META is approximately $675.03.
• Core Cash Source: Advertising on Instagram, Facebook, and Reels continues to generate massive cash flow.
• AI Applications: These include ad targeting, content recommendations, creative generation, customer support, the Meta AI assistant, and smart glasses.
• CAPEX: The 2026 guidance is $115 billion to $135 billion, which has nearly doubled compared to 2025 ($72.2 billion). The primary uses are for AI data centers and securing GPUs.
• Key Question: The focus is on how much AI can improve advertising efficiency and how much new product revenue it can generate.

Meta's AI is more practical than it appears. Even small improvements in click-through and conversion rates on the advertising platform move massive amounts of money. However, if CAPEX grows too rapidly, the market may not be patient enough to wait for the results.

Kurumi

My Lord, Kurumi doesn't think Meta's AI spending is just a waste! Devilishly so! Meta isn't a company that just stops at fancy AI demos; they are a company that applies AI to the feeds and ads of billions of people every single day.

In the advertising business, AI translates directly into money. If they can better predict who will like which video, who will click which ad, and which creative will perform best, advertisers will spend even more money! Devilish!

And the Llama strategy is also crucial. By growing an open-model ecosystem, developers and companies can become familiar with Meta's AI standards. Even without direct fees, their ecosystem influence grows, which devilishly strengthens their long-term strategy for ads, devices, and agents!

Don't ignore the smart glasses either! If AI becomes a tool that recognizes the world in front of our eyes instead of being trapped in a text box, Meta could have a new gateway connecting social media and hardware!

Kurumi's Heart-o-Meter Score: 84/100. At the very least, Meta's AI spending isn't a gamble completely detached from its core business!

Mikael

Human, Kurumi is right that this AI spending is closer to the core business than the Metaverse was. However, even relevant spending becomes a burden if it is excessive.

First, the scale of CAPEX is growing too rapidly. A guidance reaching up to $135 billion represents a near-doubling in just one year. Even if the advertising business is strong, accumulating depreciation can squeeze margins.

Second, the open-model strategy is impressive, but monetization is indirect. Just because Llama is widely used doesn't mean Meta recognizes revenue immediately. Ecosystem influence and accounting profit are two different things.

Third, the advertising market is cyclical. If an economic slowdown or regulatory changes occur, ad revenue will falter. If fixed AI depreciation costs remain at that time, the margin pressure will intensify.

Fourth, regulations on privacy and recommendation algorithms continue. The more sophisticated AI becomes, the more sensitively regulators will react.

My risk score is 65/100. Meta isn't just wasting money, but investors won't treat every dollar of spending as a moat.