Microsoft, Amazon, Alphabet and Meta collectively spent $165 billion on capital expenditures in the second quarter while generating just $7 billion in free cash flow, a gap that is putting significant pressure on Big Tech's cash generation as the four hyperscalers ramp up investment in AI infrastructure.

Jefferies' Chris Wood points to a hidden risk inside this spending surge: the industry's growing reliance on the cash-burning AI firms OpenAI and Anthropic.

Capex pressure on free cash flow

Free cash flow — the cash a company generates from operations after capital expenditures — is being compressed as the hyperscalers accelerate the buildout of data centers, chips and networking capacity needed to run AI workloads. The contrast between $165 billion of quarterly capital spending and $7 billion of free cash flow illustrates the scale of the current investment cycle and the extent to which it is reshaping Big Tech's cash flows.

The compression matters beyond the AI story itself: the four companies have for years ranked among the world's largest buyers of their own stock, and free cash flow is the pool from which buybacks, dividends and further investment are funded. The strain is already visible in financing behaviour — Meta raised $30 billion in the US corporate bond market in October 2025, one of the largest corporate debt offerings on record, and Alphabet also tapped debt markets during the year, a shift for companies long accustomed to funding expansion from operating cash. Several of the four lifted their full-year capex guidance as 2025 progressed, signalling that the buildout is set to continue.

The OpenAI and Anthropic connection

OpenAI, the developer of ChatGPT, and Anthropic, the maker of the Claude AI models, are among the most prominent names in artificial intelligence, and both continue to burn substantial cash as they scale up computing capacity. Their operations are closely tied to the hyperscalers now spending so heavily on AI: Microsoft is OpenAI's principal investor, while Amazon and Alphabet have both made multibillion-dollar investments in Anthropic. Such arrangements have included commitments by the AI firms to run their workloads on their backers' cloud platforms, meaning a significant share of hyperscaler AI capex ultimately supports the compute needs of cash-burning AI developers — the interdependence Wood identifies as a hidden risk. The structure is also circular: cash invested in the AI labs flows back to their backers as cloud revenue, and OpenAI's compute commitments extend beyond its direct investors to other infrastructure partners, widening the web of obligations behind the buildout.

Background

Chris Wood is Global Head of Equity Strategy at Jefferies and a veteran market strategist known for his weekly "Greed & Fear" note, in which he sets out his view on global markets. His latest comments draw attention to how the AI infrastructure boom is being financed, and to the degree of interdependence between Big Tech's balance sheets and the leading AI startups. The open question for the sector is whether cloud and AI-related revenue grows quickly enough to cover the cost of the capacity being installed — the arithmetic on which the sustainability of the current pace of spending depends.

Reported by Economic Times Markets (original article).