Microsoft shares surged 9 percent in premarket trading while Meta slid 9 percent after the two companies delivered quarterly reports that investors read as a referendum on whether massive AI spending is starting to pay off. The split verdict underscores a market that is no longer rewarding ambition alone, it wants evidence that the billions flowing into data centers are converting into recurring revenue.
The numbers tell the story
Microsoft’s fiscal fourth quarter beat consensus on revenue and showed Azure growing 43 percent, ahead of expectations. The company reiterated its 2026 capital expenditure forecast and signaled spending could expand further in fiscal 2027, a stance that would normally spook a market jittery over AI costs. Instead the stock popped, erasing a portion of the roughly 24 percent decline it had accumulated this year. Meta, by contrast, missed earnings estimates and guided current-quarter revenue between $61 billion and $64 billion, $62.5 billion at the midpoint, against a $63.15 billion consensus. Free cash flow collapsed 91 percent year over year to $784 million. The shares, already down about 16 percent in 2026, extended losses.
Microsoft's Copilot traction
The differentiator appears to be Microsoft 365 Copilot. Paid seats surpassed 30 million, up from more than 20 million in April, suggesting the AI assistant is moving from pilot projects into broad enterprise adoption. Forrester’s Tracy Woo argued the acceleration signals that Microsoft’s $190 billion data-center buildout is beginning to deliver returns. The market accepted that narrative even as the company kept the capex spigot open for next year.
Meta's cash burn and compute pivot
Meta’s story is messier. Mark Zuckerberg said the company is fielding offers for compute capacity at a “significant premium” to its own cost, hinting at a potential business leasing excess infrastructure to third parties. He also acknowledged Meta must retain enough compute for its own product roadmap. Quilter Cheviot’s Ben Barringer called the narrative light on detail and reliant on future possibilities, leaving both costs and revenues looking volatile. The 91 percent free-cash-flow drop is the receipt for that uncertainty.
What to watch next
The divergence sets up a clear test for the rest of earnings season: can other hyperscalers show the same conversion from capex to recurring AI revenue that Microsoft just demonstrated? Meta’s compute-leasing experiment bears monitoring, but until it materializes with pricing and volume disclosure, the market is pricing it as optionality rather than income. For now, the AI trade has split into two buckets, companies with visible seat counts and companies with promises.