Investors are starting to push back on the debt funding the AI buildout.

Long dated bonds tied to major AI infrastructure spenders have sold off as Big Tech’s borrowing spree tests market appetite for securities linked to years of data center expansion, chip demand and AI monetization.

Amazon, Alphabet, Meta, Microsoft and Oracle have been raising capital to fund data centers, processors and power infrastructure. Big Tech borrowing tied to the AI boom has climbed to about $159 billion this year, turning one of the market’s strongest growth trades into a larger test for credit investors.

The pressure has been clearest in bonds with maturities of 10 years or more. Shorter dated debt still offers exposure to companies with strong cash flow and high credit ratings. Longer dated bonds ask investors to underwrite the economics of AI infrastructure far into the future.

That is becoming a harder sell.

The largest hyperscalers are spending at a pace that has changed the shape of the bond market. Data center demand remains strong, but investors are asking how quickly that spending will turn into durable returns.

The concern is not immediate balance sheet stress. Big Tech still has deep cash reserves, dominant businesses and broad access to capital. The concern is duration. A 30 year bond carries more exposure to technology shifts, model efficiency, energy costs and future competition than a five year note.

New chips can reduce the value of older infrastructure. More efficient models can lower compute needs. Open source systems can pressure pricing. Power constraints can raise operating costs.

The selloff does not mean investors are rejecting AI. It means they are becoming more selective about how far out they are willing to finance the boom.

The AI trade is no longer just about who can build the biggest data centers. It is also about who can fund them, how long investors are willing to wait and whether the returns arrive before the debt market demands a better price.