According to a report by UBP, major technology companies, including Meta Platforms (NASDAQ: META), are increasingly relying on debt markets to fund artificial intelligence infrastructure. The analysis suggests that the five largest hyperscalers could spend between $1 trillion and $1.3 trillion on capital expenditures in 2027, a figure that exceeds the cash flow generated by these companies.
UBP estimates that in 2026, the five hyperscalers will spend approximately $820 billion on capital expenditures, compared to about $750 billion in operating cash flow. This gap indicates that earnings alone are insufficient to cover the cost of the AI buildout. As a result, companies are turning to global credit markets for financing.
The report highlights that bonds are a primary source of this capital, but other channels are also being utilized. These include project finance, securitization, leveraged loans, and chip-backed financing. The Financial Times adds that investors have provided roughly $500 billion of financing to AI-linked groups this year, with the five hyperscalers accounting for about $200 billion of that figure.
The European Central Bank (ECB) notes that hyperscalers have issued more than $100 billion of bonds in the previous year, accounting for nearly a tenth of new non-financial corporate bond issuance in the euro area. This surge in borrowing is creating a feedback loop where AI infrastructure requires more capital, which in turn increases the supply of bonds investors must absorb.
The cumulative investment by these five hyperscalers could reach $5.6 trillion through 2030. The report warns that if spending continues to accelerate while returns lag, the pressure could manifest in bond spreads, borrowing costs, and future capital expenditure decisions rather than just stock valuations.
