Bond markets are sending clear warning signals to investors regarding technology giants. The cost of insurance against default (CDS) for Oracle, Amazon, Google, and Microsoft has soared to multi-year highs, while credit spreads are increasingly diverging from stock prices. The artificial intelligence race appears to be triggering not euphoria, but well-founded anxiety.
The situation is unfolding on multiple fronts simultaneously, each deserving close attention. Let's start with the main fear indicator — credit default swaps.
Record cost of protection against Big Tech risks
Five-year CDS spreads for Oracle, Amazon, Google, and Microsoft have risen to approximately 75 basis points. This is not just an increase — it is nearly a seven-year high. Excluding Oracle from the sample paints a slightly milder, but still significant, picture: spreads for the same group of companies have risen to 49 basis points, the highest level since at least 2018.
Key point: both indicators have more than doubled since the start of 2025. They already significantly exceed the peak values of the 2022 bear market. The reason for this dynamic is clear — record borrowing. In 2026, Amazon, Google, Nvidia, Meta, Oracle, and SpaceX placed an astronomical $182 billion in investment-grade bonds. This represents a 1300% increase year-over-year. Tech giants are actively borrowing to finance their AI ambitions, and the market is beginning to price in the risks that these investments may not pay off.
Divergence between credit and stocks: a classic "red flag"
The second alarming signal is the growing divergence between the credit and stock markets. Credit spreads for the largest data center operators have widened to 153.5 basis points over government bonds, compared to approximately 118 basis points in February. This is the highest level since the launch of the corresponding Goldman Sachs basket.
Meanwhile, the stocks of these same companies have gained only about 3% over the same period. This is a classic sign that debt investors (more conservative and "smart" money) are already pricing in problems, while the stock market is still holding on to enthusiasm. Demand for the bonds themselves is also weakening: the coverage ratio for issuances has fallen from approximately 5 in February to less than 2 in July.
The reason is the scale of future spending. Estimates suggest that data center operators could spend around $5.5 trillion on AI by 2030, with roughly half of that amount needing to be financed through the bond market. Any misstep by the Federal Reserve would leave the market without a safety margin.
Speculative AI stocks are collapsing
The third signal is the collapse of the most volatile securities. The Goldman Sachs basket of high-risk stocks with strong momentum risks falling 23% in a month. This is the worst result in 17 years. The basket includes Nvidia, Super Micro Computer, Palantir, D-Wave Quantum, and Navitas Semiconductor — the very "pumped-up" assets particularly favored by retail investors.
Their collapse hits private traders hardest. The reason for the sell-off is growing doubt: are data center operators building excessive AI capacity that may turn out to be unneeded?
My expert conclusion: All three signals form a single picture. Credit markets, which historically sense problems first, are already pricing in the risks of the AI race. Stocks are holding up for now, but this tension cannot last forever. If it continues to build, echoes will sooner or later be felt by those betting on AI through the stock market. Investors should prepare for increased volatility.