Microsoft shares rose 15.5% on July 30, whereas Meta fell 8%, giving traders two sharply alternative ways to evaluate Huge Tech’s AI infrastructure increase. Microsoft paired heavy spending with sooner Azure development, rising Copilot adoption and a bigger business backlog. Meta’s promoting enterprise grew strongly, however bills climbed and free money movement dropped.
Microsoft shouldn’t be spending cautiously. It reported $41 billion in quarterly capital expenditures, whereas Meta expects complete 2026 capital expenditures of $130 billion to $145 billion. The intervals and accounting particulars differ, so the figures aren’t immediately comparable. Buyers responded extra favorably to Microsoft as a result of its outcomes confirmed clearer hyperlinks between infrastructure funding, cloud consumption, software program subscriptions and contracted demand.
Microsoft hyperlinks spending to present demand
The July 30 market response adopted Microsoft’s fiscal fourth-quarter outcomes. Quarterly income and working revenue each rose 18%. Azure income elevated 43%, whereas full-year Azure income exceeded $100 billion for the primary time.
Microsoft additionally reported greater than 30 million paid Microsoft 365 Copilot seats, with internet additions greater than doubling from the earlier quarter. Its business remaining efficiency obligation — contracted income not but acknowledged — rose 84% to $678 billion. About 30% is predicted to be acknowledged inside 12 months.
These disclosures give traders a number of measurable demand indicators. Clients are shopping for Copilot licenses, consuming Azure companies and making multiyear commitments. Microsoft additionally operates a number of business platforms that may use the identical infrastructure, together with Azure, Microsoft 365 and GitHub.
The buildout nonetheless carries monetary stress. Microsoft reported $19.6 billion in quarterly free money movement as capital expenditures elevated. Its firm gross margin share fell yr over yr, pushed partly by the gross sales combine shifting towards Azure and continued AI infrastructure funding. Microsoft forecasts Azure development of roughly 45% in fixed foreign money for its subsequent quarter.
Vendor development doesn’t assure that every deployment will produce an appropriate return. As a result of AI agent cloud prices can differ sharply between comparable workflows, IT groups want workload-level controls, adoption information and measurable operational outcomes earlier than increasing Azure AI or Copilot contracts.
Meta’s AI returns stay more durable to separate
Meta raised the decrease finish of its 2026 capital-expenditure forecast from $125 billion to $130 billion, leaving the higher finish at $145 billion. The forecast contains principal funds on finance leases. Quarterly free money movement fell to $784 million from $8.55 billion a yr earlier.
Meta’s core enterprise remained sturdy. Based on Meta’s second-quarter outcomes, income elevated 28% to $60.8 billion. Advert impressions rose 14%, whereas common advert costs elevated 12%.
Meta says AI is enhancing its core enterprise and supporting new merchandise, nevertheless it doesn’t report AI-attributable income individually. Complete prices and bills rose 55%, whereas working revenue declined 8%. The quarter included $2.4 billion in authorized costs and $1.18 billion in severance bills, so AI funding alone didn’t trigger the decline.
Meta expects its infrastructure to help bigger fashions, private brokers and future enterprise merchandise. It is usually reportedly contemplating whether or not to promote entry to extra AI compute, nevertheless it has not introduced a business service, pricing, availability or monetary targets.
Buyers are making use of an identical check elsewhere. Google just lately raised its 2026 capital-spending forecast to as a lot as $205 billion as cloud demand accelerated. Microsoft at the moment gives stronger proof of income development and contracted demand, whereas Meta’s newer AI companies stay financially much less seen.
Learn extra: Bigger capital budgets don’t remove each infrastructure constraint; AI information heart networking is changing into a separate bottleneck as GPU clusters place extra stress on switches, interconnects and congestion controls.
