Mark Zuckerberg is doubling down on AI, and Wall Street is starting to sweat.
Meta announced Wednesday it's committing an additional $21 billion to CoreWeave, a cloud infrastructure provider that powers AI workloads. This isn't a one-off splurge, it's part of Meta's massive bet that AI will justify the astronomical capital expenditures the company has been racking up.
Let's put that number in context. Meta's total AI spending for 2026 is expected to exceed $60 billion, according to analyst estimates. That's not pocket change, even for a company that prints money from ads. For comparison, Meta's entire 2023 capital expenditure budget was around $30 billion. They've doubled their spending in three years, all in the name of AI infrastructure.
The problem? Investors are still waiting to see meaningful revenue from all this spending.
Sure, Meta's AI-powered ad targeting has improved, and their Llama large language models have gotten rave reviews from developers. But when you're dropping $60 billion a year, you need more than "improved ad targeting" to justify the expense. Shareholders want to know when AI stops being a cost center and starts being a profit engine.
The CoreWeave deal is particularly interesting because it signals Meta is locking in long-term compute capacity. CoreWeave specializes in GPU infrastructure, the kind of hardware you need to train and run massive AI models. By committing $21 billion upfront, Meta is essentially pre-paying for years of computing power, which protects them from price spikes and capacity shortages.
Smart? Absolutely. But it also means Meta is all in on AI, with no easy way to pull back if the returns don't materialize.
Wall Street's concern isn't that Meta is investing in AI, it's that they're investing this much without a clear path to profitability. Every tech giant is dumping money into AI right now, but Meta's spending-to-revenue ratio is starting to look uncomfortable.
The bull case: Meta is positioning itself as the infrastructure leader in AI, similar to how Amazon dominated cloud computing with AWS. If Meta's AI investments pay off, this $60 billion could look like a bargain in five years.
The bear case: Meta is burning cash on technology that may not generate corresponding revenue growth, and if AI doesn't deliver, shareholders are going to demand answers.
For now, Zuckerberg is betting that AI justifies the expense. Investors are holding their breath, hoping he's right.
