Microsoft's stock jumped 15% Thursday as investors rewarded the company for demonstrating that its massive artificial intelligence infrastructure investments are finally generating returns. Meta's shares plunged 9% the same day, caught in a brutal market reckoning over bloated spending and murky future revenue prospects.
The divergence reveals a hard truth about the AI boom: execution matters, and so does fiscal discipline.
Microsoft posted fiscal fourth-quarter revenue that beat analyst estimates, with its Azure cloud business growing 43%—also ahead of market expectations. The company now has over 30 million paid seats for Microsoft 365 Copilot, its AI work assistant, up from more than 20 million as of April. That's the kind of concrete user adoption that investors want to see when a company commits hundreds of billions to infrastructure.
"Microsoft's strong revenue performance, combined with accelerating Copilot adoption, signals that its $190 billion data-center buildout is beginning to deliver returns," Tracy Woo, principal analyst at Forrester, said. The stock gain marks Microsoft's best day since March 13, 2020—six years ago. Remarkably, the company achieved this surge even as it reiterated its 2026 capital expenditure forecast and signaled a potential spending expansion in its 2027 fiscal year, at a time when the market is jittery over the cost of AI.
The Meta Problem: Spending Without Clarity
Meta's situation tells a different story. The company missed investor expectations on earnings and its revenue guidance for the current quarter. Meta said it expects revenue this quarter between $61 billion and $64 billion, or $62.5 billion at the middle of the range. Analysts were expecting $63.15 billion, according to LSEG. That miss matters.
Worse, Meta's free cash flow plunged 91% year-on-year to $784 million as the company continues to spend heavily on AI investments. The stock is on a record losing streak, on pace to hit 11 days, and is down over 20% in that time. Investors are asking a simple question: where's the payoff?
Meta CEO Mark Zuckerberg acknowledged the company is "getting a lot of offers for compute at a significant premium" over what Meta paid for it. He suggested the company might lease excess computing capacity to third parties, a potential revenue stream that could offset some of its massive infrastructure costs. But he offered few details on what that business could look like or when it might materialize.
"Right now, the narrative from Mark Zuckerberg is a little light on detail and relying on what could be done in the future," Ben Barringer, head of technology research at Quilter Cheviot, said. "Meta still has a crucial role to play in the AI world, but it is still finding its way somewhat and that is why we see both costs and revenues looking a little volatile."
Zuckerberg also acknowledged that Meta will need to keep compute resources for itself to develop new products—a reminder that the company can't simply monetize its way out of its spending problem.
Market Discipline at Work
Thursday's market reaction demonstrates something important about capital allocation. Microsoft made big bets on AI infrastructure and is now showing customers will pay for the tools built on top of it. The company can justify continued spending because it's generating revenue growth that exceeds expectations.
Meta made similar infrastructure bets but hasn't yet proven the business model works. Zuckerberg's vague talk about future possibilities and potential compute leasing arrangements isn't enough to convince investors that the company's spending trajectory is sustainable. The market is demanding accountability.
This isn't about whether AI is valuable. It's about whether companies can actually convert infrastructure investments into profitable operations. Microsoft has shown it can. Meta is still trying to figure that out.
Why This Matters:
These earnings results expose a fundamental tension in the AI economy: massive capital expenditures only make sense if they generate returns. Microsoft's ability to show 43% growth in its cloud business and rapidly expanding Copilot adoption justifies its $190 billion infrastructure commitment to investors. Meta's inability to demonstrate comparable returns—combined with a 91% collapse in free cash flow—raises serious questions about whether its spending is sustainable. For taxpayers and policymakers watching corporate capital allocation, the lesson is clear: government should be skeptical of subsidizing or directing private AI investment. Markets are already enforcing discipline. Companies that can't show returns will face consequences. That's how it should work.