Big Tech’s AI capex surge pressures cloud providers; Microsoft faces Copilot, cash‑flow scrutiny

Big Tech’s AI spending spree forces cloud providers, Microsoft faces scrutiny

Big technology firms are ramping up capital spending to build the data‑center and semiconductor capacity needed for generative AI, creating immediate operational and strategic pressure on cloud providers such as Microsoft. Companies including Amazon, Alphabet, Microsoft and Meta disclose roughly $120 billion of combined capex in the fourth quarter alone, and industry estimates point to aggregate capex approaching $660 billion in 2026. Executives including Nvidia’s Jensen Huang frame AI as “fundamental infrastructure,” pushing hyperscalers to expand facilities, chips and networking at a pace that tests supply chains and project timelines.

The shift is sharpening debate over near‑term cash flow and the timing of returns, with analysts warning free cash flow could dip as firms front‑load spending to seize AI market share. Melius Research lowers its view on Microsoft in part because it foresees a near‑term free‑cash‑flow shortfall if subscription growth slows and Copilot uptake remains modest. The firm highlights that Copilot reports about 15 million paid users after three years, and suggests Microsoft risks margin pressure if it discounts Copilot to sustain Office 365 relevance — a move that could consume Azure capacity and complicate resource planning.

Operationally, the surge in AI demand is changing enterprise purchasing patterns and platform engineering. Databricks reports agent‑built databases now account for about 80% of workloads on its platform, signalling that customers are automating more of their software development and increasing dependency on model hosting, data pipelines and orchestration — services that run on Azure and rival clouds. Analysts say investors and corporate customers will closely watch quarterly reports, detailed guidance on capex timing and utilization rates for data centre capacity and semiconductor supply trends to judge whether the heavy spending converts into durable enterprise adoption or leaves providers with excess infrastructure.

Anthropic takes AI to the Super Bowl, escalating consumer‑facing competition

AI startups are moving from technical messaging to mass marketing: Anthropic runs Super Bowl ads that directly challenge rival monetisation models, framing its Claude assistant as ad‑free and pitching consumer trust as a competitive edge. The campaign signals deep‑tech firms are prioritising brand differentiation as they compete for both enterprise and consumer mindshare.

Microsoft name used in phishing scams, watchdogs warn

Security commentators flag a surge in convincing scam emails posing as urgent Microsoft messages about Copilot and account issues. Cybersecurity advisories urge users to verify sender domains and treat unexpected Copilot links as potential phishing attempts, underscoring how rising AI product visibility invites new fraud risks.