The six largest technology companies are now borrowing at such a scale that they are effectively competing with the US government for investors' capital.

JPMorgan strategist Michael Cembalest noted that hyperscalers — the group comprising Amazon, Alphabet, Meta, Microsoft, Oracle, and Nvidia — have collectively issued roughly $320 billion in new long-term investment-grade debt and related financing through late August 2026. That figure represents approximately 68-70% of total new long-term US Treasury bond issuance over the same period. The borrowing is funding a historic buildout of AI data centers and the power infrastructure needed to run them, a capital-expenditure cycle whose scale has few precedents in corporate history.

The AI Spending Binge Behind the Borrowing

Before 2025, these same companies issued under $50 billion in debt annually, relying instead on the enormous cash piles generated by their core businesses. In 2025, hyperscaler debt issuance climbed to roughly $180-205 billion. By 2026, it had nearly doubled again. The shift marks a departure from the sector's longstanding practice of self-funding expansion from operating cash flow, and signals that planned AI infrastructure spending now exceeds what internal cash generation alone can cover.

The six largest hyperscalers now account for approximately 5% of the US investment-grade bond index, double their share from just two years ago. That growing weight matters for bond investors broadly, since index-tracking funds and passive strategies are mechanically increasing their exposure to these issuers as their share rises.

The Treasury Yield Pressure

When high-quality corporate borrowers flood the market with new bonds, they compete directly with Treasuries for the same pool of fixed-income capital. That dynamic can push Treasury yields higher on the margin, as the government must offer more attractive rates to remain competitive. Treasury yields serve as the benchmark for borrowing costs across the economy, including mortgages and corporate financing, so any upward pressure at the margin carries implications well beyond the bond market itself.

Credit spreads on hyperscaler bonds have widened modestly, with median increases of 10-22 basis points across some tenors. The broader AI-linked bond issuance picture is even larger: related sales across currencies have exceeded $220 billion in 2026.

Can the Market Handle Another $1.5 Trillion?

JPMorgan global market strategist Stephanie Aliaga estimates that the hyperscaler group could comfortably absorb an additional $1.5 trillion in debt without pushing their leverage ratios above the broader market average. How this plays out will be visible in several places: the pace of upcoming hyperscaler bond sales, the trajectory of credit spreads as supply continues, and whether Treasury issuance calendars adjust to the added competition for fixed-income capital.

For fixed-income investors, hyperscaler bonds offer exposure to some of the most creditworthy companies on earth, backed by recurring revenue and strong cash generation. However, the sheer volume of issuance means these bonds are no longer scarce, and the crowding-out effect on Treasuries introduces an additional variable into the rates equation.