America's seven biggest technology companies have piled up $578 billion in combined debt as they race to build AI infrastructure, according to a new SBI Research report warning the borrowing binge now carries real financial risk.

The Magnificent Seven — Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla — saw their combined debt rise nearly 45% from around $399 billion in 2025 to $578 billion by mid-2026, the report said. That is roughly 1.7 times the level the group carried in 2023, a jump the researchers tie directly to spending on data centres, chips and power needed to run AI systems at scale.

The findings landed on Monday, November 2, as markets were already on edge. Asian shares slid at the week's start amid worries over the AI investment boom, with Apple, AMD and Palantir due to report earnings soon and US midterm elections on Tuesday.

Debt is rising faster than the revenues meant to pay for it

SBI Research flagged more than just the headline debt figure. The report also pointed to substantial off-balance-sheet commitments — obligations tied to purchases and investments that do not show up in the main debt total — as companies keep expanding their AI infrastructure. Goldman Sachs analysts expect capital spending by the big hyperscalers to almost double to $800 billion this year, and to rise further to $1.2 trillion in 2027.

The concern, the researchers said, is what happens if borrowing costs keep rising. Higher borrowing costs and rising Treasury yields could put pressure on the companies' valuations by increasing the discount rate applied to future earnings, the report cautioned. When investors demand higher returns for holding risk, the present value of profits expected years from now falls — and for companies priced on decades of future growth, that arithmetic bites hard.

Using what it calls an equity duration approach, SBI Research estimated that a rise in yields could result in a potential market capitalisation decline of around $1.67 trillion to $2 trillion for the seven companies, whose combined market value was estimated at approximately $23 trillion.

October's rout showed how quickly sentiment can turn

The warning comes days after a brutal October for technology stocks. The Nasdaq Composite fell 2.9% on Friday, its largest single-day drop since June, closing out a month in which an estimated $2 trillion was erased from the Nasdaq-100 as investors questioned whether AI spending will ever pay off. Alphabet shares plunged 12.6% on Thursday, its worst day on record, while Intel fell 28.6% after forecasting lower profits.

The report also highlighted a quieter knock-on effect: the Federal Reserve's unrealised losses on Treasury holdings rose $26.3 billion, from $535.3 billion in December 2025 to $561.5 billion in June 2026, as rising yields cut the value of the bonds it holds.

Anthropic's listing is the next test

The report also pointed to the expected entry of Anthropic into public equity markets, likely around November 12, with a target valuation of $1.5 trillion to $2 trillion. It described the proposed listing as a "high-stakes litmus test" for the financial viability of the AI boom, giving public investors a chance to judge whether rising enterprise revenues can justify the substantial infrastructure and computing costs involved.

That listing, alongside Apple earnings on Monday and AMD on Tuesday, will give the market its next hard evidence on the AI trade. If the numbers hold up, the debt pile looks like a calculated bet. If they disappoint, the warning about $23 trillion of market value will feel a lot less theoretical.

By Hamza Faisal