Google and Microsoft slow AI build-out as costs bite

Google and Microsoft both reported strong demand for their AI services this week, but neither company raised their capital expenditure forecasts. According to CNBC and Fox Business, both firms' latest earnings show AI is driving growth, but the pace of new spending on infrastructure is slowing.

Why it matters

For the past year, the biggest cloud companies have been pouring money into data centers to train and run AI models. Wall Street and enterprise buyers have watched these investments closely, since they set the pace for the entire sector. Now, with Alphabet and Microsoft both reporting that their capital expenditures fell short of estimates, some are asking whether the breakneck race to build AI infrastructure is starting to cool.

Strong demand for AI-powered products remains. Fox Business says both companies' earnings were lifted by customers adopting new AI tools in cloud and productivity products. But the decision to hold off on raising capex forecasts, as reported by CNBC, suggests management is watching costs more carefully. This shift could affect how quickly new AI features land in mainstream business software and how much capacity is available for new projects.

Data center build-outs slowed last quarter, raising questions about future AI capacity.

AI demand is up, but spending is under review

Alphabet and Microsoft both highlighted surging demand for AI services in their earnings calls. This matches what customers have seen: new models, more AI features in cloud products, and rapid expansion in tools for developers and businesses. But CNBC reports that, despite this demand, capital expenditures did not meet analysts' expectations. Both companies are investing heavily in data centers, but the pace is no longer accelerating as quickly as it was earlier this year.

Some industry analysts say this pause is a response to higher costs for AI hardware and energy. Others suggest the companies are waiting to see how much further demand will rise before making new commitments. The move may also reflect a focus on efficiency, with firms like Google DeepMind working to cut AI energy use, as reported in our recent coverage of Gemini 4 Carbon.

Fiber optic networks are a key bottleneck as cloud giants weigh new infrastructure plans.

What this means if you run a business

If your company relies on cloud AI services from Microsoft or Google, the good news is that demand remains strong and the core offerings are not being cut. However, the slower growth in infrastructure spending could mean longer lead times for new features or capacity increases. For businesses building on top of these platforms, it's a signal to keep an eye on service availability and consider contingency plans for high-demand workloads.

This shift also hints at a more cautious approach to AI rollouts. Vendors may prioritize enterprise accounts or high-margin products when allocating new AI resources. If your business is experimenting with custom AI models or planning to scale usage, it may be worth reaching out to your cloud provider early to discuss capacity and support. As the largest players balance growth with cost, smaller firms and startups should watch for potential impacts on pricing or access.

For more detail on how tech companies are optimizing AI infrastructure, see our report on Google's Gemini 4 Carbon model and its energy focus.