Hyperscalers have borrowed at record levels to fund AI infrastructure, and Wall Street is building a derivatives market around that risk. The debt still belongs to some of the strongest companies in the world, but the scale of the buildout is changing how investors think about Big Tech.
The AI boom is no longer just an equity story. It has moved into the credit markets, where banks, hedge funds and asset managers are now trading protection on the debt of Alphabet, Amazon, Meta, Microsoft and Oracle. That shift matters because the companies building the AI future are increasingly relying on borrowed money, leased data centers and structured financing to keep the infrastructure race moving.
Bloomberg reported in March that JPMorgan had launched a credit default swap basket tied to those five hyperscalers, giving clients a cleaner way to hedge or bet against the debt behind the AI buildout. The product trades in $25 million increments, a size that makes clear this is an institutional market, not a retail narrative. For corporate treasurers and credit investors, that is the point. AI infrastructure has become large enough to need its own plumbing.
The backdrop is a record wave of borrowing. The five major hyperscalers issued about $121 billion of U.S. investment-grade bonds in 2025, compared with an average of roughly $28 billion a year from 2020 through 2024, according to figures cited by Bank of America. Quartz recently noted that Bank of America raised its 2026 hyperscaler issuance forecast to $175 billion from $140 billion after Amazon came to market with a large round of debt. That is not a small adjustment. It says the market expects the AI spending cycle to keep leaning on credit.
From Cash Machines To Credit Issuers
For years, Big Tech looked almost self-funding. Microsoft, Alphabet, Amazon and Meta generated enough cash to invest heavily, buy back stock and still leave investors with balance sheets that looked unusually clean. AI has changed that rhythm. Data centers, custom chips, power contracts and networking equipment require capital at a scale that even the richest technology companies are choosing not to absorb entirely from operating cash flow.
That does not mean these companies suddenly look fragile. Alphabet, Microsoft and Meta still generate enormous cash flow, and Amazon has a cloud business that remains central to enterprise technology spending. The issue is concentration. Equity investors are already heavily exposed to AI through the largest technology stocks. Now credit investors are becoming more exposed as the same companies issue more bonds and enter the major credit indexes in larger weightings.
Oracle is the sharper test case because its balance sheet gives investors less room for comfort. The company has leaned hard into AI cloud infrastructure, including large commitments tied to data center expansion. Its credit rating sits lower than the strongest hyperscalers, and its credit default swaps have become among the most actively watched in the investment-grade technology market. If investors want a signal for how nervous credit markets are about AI spending, Oracle is where many of them look first.
The New Market For AI Credit Risk
Credit default swaps are not new, but the names driving demand are. Meta, Alphabet and Microsoft were added to the CDX Investment Grade Index, a benchmark investors use to manage corporate credit exposure. That inclusion makes sense. If hyperscalers are becoming larger bond issuers, the main tools used to hedge investment-grade credit need to reflect that reality.
The new JPMorgan basket makes the trade easier to express. Instead of buying protection on one company at a time, investors can hedge a broader AI infrastructure theme. A bank with too much exposure to hyperscaler loans can reduce risk. A hedge fund can sell protection and collect premiums if it believes fears are overdone. A portfolio manager can offset the growing overlap between technology equities and technology credit. The market is doing what markets do when a funding wave becomes too large to ignore.
This is where the comparison with 2008 needs care. The debt here is not subprime mortgage debt, and the borrowers are not weak consumers with poor credit histories. These are profitable companies with real assets, large customer bases and deep access to capital. The more practical risk is spread widening. If AI revenue takes longer to arrive than expected, investors may demand higher compensation to hold the debt, even if default remains unlikely.
The Hidden Leverage Question
The harder issue is that not all AI financing sits neatly in headline bond totals. Data center leases, special purpose vehicles and long-term purchase commitments can make leverage look cleaner than the economic obligation really is. That is especially important in AI infrastructure, where the useful life of assets, the pace of chip upgrades and the timing of customer demand are all moving quickly.
For founders, the impact is indirect but real. If credit spreads widen or lenders become more cautious, the pace of data center construction could slow. That would feed into cloud capacity, GPU availability and eventually the price startups pay to train or run AI products. The biggest companies may still get the power and chips they need, but smaller companies usually feel the squeeze later through pricing and access.
For investors, the lesson is that AI exposure is no longer confined to share prices. It now runs through corporate bonds, derivatives, leases and bank balance sheets. That creates new ways to hedge, but it also creates new ways for stress to travel through portfolios that once looked diversified.
The next milestone is monetization. If enterprise AI adoption produces enough durable revenue, the debt will look like the cost of building a dominant infrastructure layer. If revenue disappoints, the market will not need defaults to reprice the story. Wider spreads would be enough. Wall Street has built the credit machinery for the AI boom. Now investors have to watch whether the cash flows arrive on time.
Also read: Putting The Senses In AI • Samsung’s AI memory bonus fight is testing its chip supply chain • Rising U.S. debt is starting to test the AI funding boom
