AI’s central storyline today is the same one investors keep testing in the numbers: hyperscalers are spending to expand capacity faster than demand for AI services can fill it, and the bills are showing up before the margins do. Microsoft delivered strong Azure growth while pushing capex expectations to about $170 billion for calendar-year 2026, including more than $50 billion in the current quarter. The company also posted quarterly profit of over $36.6 billion and sweetened the deal for cost discipline, offering voluntary buyouts to around 60,000 workers after earlier layoffs of more than a quarter of its workforce. The message is clear: scale the AI backend now, worry about profitability later.
Alphabet’s Google Cloud echoed the strain with less comforting optics. Cloud revenue doubled to $30.7 billion in the third quarter, but Alphabet shares fell more than 11% after-hours as free cash flow stayed deeply negative at -$9.1 billion. Management is also steering toward roughly $200 billion in capex in 2026, and the market’s focus shifted from “cloud demand” to whether that spending can be funded without bleeding.
Even on the software side, the scrutiny tightened. Amazon warned engineers against “tokenmaxxing,” calling token and lines-of-code counting “deeply flawed” and cautioning that more AI usage can mean worse work. Instead, teams are nudged to document when tooling doesn’t help and to compare 10th- and 90th-percentile AI users—an oddly practical corrective to all this hardware optimism.
