Hyperscaler earnings: AI is booming, and so is the bill

This week's hyperscaler earnings showed the same core signal across big tech: AI demand is real. Microsoft, Meta, Amazon, and Alphabet all reported strong growth in cloud, ads, or AI-related workloads. The split in share-price reaction came down less to demand and more to whether investors could see a credible path from spending to returns.

What the quarter actually showed

Microsoft

$MSFT reported that Azure and other cloud services revenue grew 40% year over year, while its AI business surpassed a $37 billion annual revenue run rate. That is a clear demand signal. But capital expenditure rose to $31.9 billion in the quarter, and the stock still traded lower after the release as investors weighed how long margin pressure may last.

Meta

$META reported first-quarter revenue of $56.3 billion, up 33% year over year. The cleaner takeaway was top-line growth and stable 41% operating margin, rather than the headline EPS figure alone, which benefited from an $8.03 billion tax benefit. The market focused on capex guidance rising to $125 billion to $145 billion, and shares fell in extended trading.

Amazon

$AMZN reported that AWS revenue grew 28% to $37.6 billion, its fastest growth in 15 quarters. That was paired with a sharp drop in trailing 12-month free cash flow to $1.2 billion from $25.9 billion, driven primarily by higher property and equipment spending tied to AI infrastructure. Unlike Microsoft and Meta, Amazon's initial reaction was more constructive, but the underlying trade-off was still the same: strong demand, heavy upfront spend.

Alphabet

$GOOGL reported first-quarter revenue of $109.9 billion, up 22% year over year, and Google Cloud revenue of $20.0 billion, up 63%. Backlog expanded to more than $460 billion, and Gemini models were processing more than 16 billion tokens per minute via direct customer API use. The headline EPS figure of $5.11 was boosted by large unrealized gains on non-marketable equity securities, but the operating numbers were still strong enough that investors rewarded the release.

The market's real concern: timing, not direction

What markets seem to be testing is not whether AI demand exists. By April 29 and April 30, 2026, that question looked much less controversial. The harder question is whether hyperscalers can convert today's data center, chip, and energy spend into durable revenue and margin expansion quickly enough to justify the cash burn happening now.

That helps explain why similar quarters got different reactions. When growth is visible, monetization looks broad-based, and management frames the path to returns clearly, the market is willing to look through capex. When spending grows faster than the evidence of near-term monetization, even strong numbers can feel incomplete.

What comes next

For the next few quarters, investors will likely focus on three things:

• whether AI revenue growth keeps outpacing infrastructure spend
• whether margins stabilize as utilization improves
• whether management teams can show that current capacity build-outs are backed by durable customer demand rather than short-term experimentation

The broad takeaway still looks constructive. These results suggest AI infrastructure demand is real and scaling. The unresolved question is how quickly that demand turns into cleaner cash flow and operating leverage.