Since the AI boom began in 2023, the four major hyperscalers have invested more than $1 trillion in capital expenditures, with combined spending hitting $1.1 trillion by the end of June 2026. They are planning an additional $745 billion in capex for the current year, primarily directed towards data centers, advanced chips, and power infrastructure. These massive investments have led to a significant increase in financial commitments, with Meta adding $233 billion in obligations and Microsoft signing over $130 billion in data-center leases. In a single quarter, three firms alone agreed to nearly $900 billion in AI-related commitments. This aggressive spending has caused combined free cash flow to drop to a decade low of $7 billion, with only Microsoft and Meta reporting positive cash flow.
Analysts are warning that there is "basically no end in sight" for capex growth, and the long delay between upfront data-center spending and revenue generation means investors might have to wait years for substantial returns. The gap between AI infrastructure spending and actual sales is widening, with an estimated $600 billion annual revenue gap in 2025 that is growing in 2026 as capex accelerates faster than revenue. For 2026, the five largest hyperscalers are projected to spend between $700 billion and $900 billion on capex, a 36% increase over 2025. Amazon alone has guided for $200 billion in capex this year, more than doubling its 2025 outlay, and Meta raised its full-year guidance to as much as $145 billion.
This unprecedented spending is causing concern among investors and analysts. Aggregate capex is expected to reach 25% of revenue for hyperscalers in FY26, up from 17% in FY25, and management commentary suggests further acceleration into FY27. Hyperscalers are investing ahead of demand, with assets expected to pay back within 2-3 years, and long-term customer commitments provide revenue visibility. However, the accelerated AI infrastructure spending is expected to lead to a cyclical trough in Return on Invested Capital (ROIC) between FY26 and FY28, due to the lag between capital deployment and revenue monetization. With capex at such high levels, an increasing portion of investment is being funded through debt issuance and balance sheet expansion, rather than internally generated cash flows, significantly increasing the invested capital base. Individual companies like Alphabet, Amazon, and Meta have already experienced negative or significantly reduced free cash flow.
Concerns about the sustainability of this AI capex boom have been raised by financial analysts. Evercore ISI analysts have labeled it a "red flag" for stock valuations, noting that the sheer size of AI capex risks making some companies cash-flow negative. Amazon's $200 billion capex for 2026 was higher than anticipated, likely resulting in a negative free cash flow year for the company. While free cash flow generation remains positive in aggregate, the ongoing spending on AI infrastructure is becoming a critical issue. Hyperscalers' 12-month forward free cash flow has plummeted below 2022 cycle lows, and a negative aggregate free cash flow would be a major "red flag." Despite these concerns, some analysts view the AI trade as overdone rather than a bubble, suggesting that healthy rotation in the market is occurring.
