Meta is betting that artificial intelligence will let it do something far more valuable than showing ads: it will predict what you want to buy before you know it yourself. The company is doubling its annual infrastructure spending to roughly $125 billion to $145 billion in 2026, up from $72.2 billion in 2025, to build the computing power needed to make this shift. If the gamble works, Meta stops looking like a media company and starts looking like a toll on commerce itself.

What Is Meta's AI Commerce Strategy?

Meta's recommendation engine has evolved from a simple system that shows you ads based on basic information into an "agentic" system powered by artificial intelligence. This means the system doesn't just guess what you might like; it learns from your history, interests, relationships, and behavior to predict what you'll buy next, sometimes before you realize you want it. On Meta's Q4 2025 earnings call, Mark Zuckerberg stated plainly: "We're starting to see agents really work," describing AI that understands context drawn from a person's history, interests, content and relationships, and pointing directly at agentic shopping tools that help people find the right product from a business inside Meta's catalogue.

Mark Zuckerberg

The numbers already show the pull. In Q1 2026, ad impressions across Meta's family of apps rose 19 percent and the average price per ad rose 12 percent, both year on year. More inventory and higher prices at the same time signals demand outrunning supply. Improvements to Meta's AI recommendation systems lifted ad conversions on Instagram by roughly 5 percent and on Facebook by around 3 percent, with online commerce the single largest contributor to ad revenue growth.

Think of it this way: instead of a toilet paper brand spending $100,000 on a TV ad that reaches parents, teenagers, and babies indiscriminately, that same budget now goes to Meta, where targeting is built around an actual profile of intent. The advertiser sees the return on that investment, keeps the full $100,000 in the channel, and books double the sales. This is Jevons paradox at work. When something gets cheaper and more efficient, we don't use less of it. We use more.

How Is Meta Building This Infrastructure?

The spending is not subtle. Meta's capital expenditure ran to $72.2 billion in 2025, and guidance for 2026 has been lifted to a range of $125 billion to $145 billion. The company is roughly doubling its infrastructure bill in a single year. For context, Meta did $200.97 billion of revenue in 2025, up 22 percent, with advertising still around 97 percent of the total. The ad machine is funding the compute machine, and the bet is that the compute machine turns the ad machine into something bigger.

Meta is building what it calls a "supercluster" of computing power. Bigger context, sharper inference, a system that doesn't just know roughly who you are but can call what you want next, sometimes before you can. This is not speculation about the far future. The company has already shown the conversion math works at smaller scales.

What Are the Key Risks and Challenges?

The payback on $125 billion to $145 billion of annual capex is unproven, and several headwinds could derail the strategy:

• Margin Pressure: Depreciation from the infrastructure build flows straight through the income statement for years, and Meta's operating margin has already slipped from a 2024 peak near 48 percent toward 41 percent in Q1 2026.
• Reality Labs Losses: Meta's metaverse division, Reality Labs, is still losing roughly $4 billion a quarter, adding to the company's overall cash burn.
• Regulatory Risk: Regulation in the European Union and United States remains live, and the entire targeting advantage rests on a data and privacy regime that politicians can change.
• Unproven Revenue Model: Agentic commerce is a roadmap item, not a revenue line yet, meaning the company is spending billions on a strategy that has not yet generated meaningful income.

A large, far-off opportunity is exactly the kind of catalyst that is easy to narrate and hard to underwrite. When Meta raised its capex guidance in April, the stock fell on the news, suggesting investors are skeptical about whether the company can deliver returns on such massive spending.

How Does This Compare to Payment Networks?

If Meta's strategy succeeds, the better comparison isn't a media company at all. It's a payments network, something closer to Mastercard or Visa: a toll on activity that happens whether or not you think about the toll. The difference is that Meta would also be the thing that created the intent in the first place. Picture a cookie shop owner who is told, "I can send you 100 people a day who I know, with near certainty, will buy from you, and in return I want 30 percent of the sale." That owner takes that deal every time. At 50 percent, they're probably still in, depending on how long the arrangement can last.

Meta is building toward that position. Not a billboard. Not a media channel. An operator that sits between intent and purchase and clips a ticket on the transaction. The company has a captured base of over 3 billion daily active people with strong network effects to protect it, making it an appealing proposition at a time when other pockets of the AI trade are fetching multiples of 60, 90, or even 400 times earnings.

The bet is audacious, the spending is real, and the outcome remains uncertain. But if Meta can prove that AI-powered agentic commerce works at scale, it will have transformed itself from an advertising company into something far more powerful: a commerce infrastructure layer that extracts value from every transaction that happens on its platform.