If there is one relatively safe bet for 2026, it is that the yield on U.S. Treasury bonds will rise as crude oil futures increase. However, over the past month, this relationship has come to an end. BlackRock's Systematic Alternatives ETF ( IALT ) Senior Portfolio Manager Jeffrey Rosenberg stated that the failure of this trading logic has had a significant impact on the market.

In the most recent issue of CNBC's ETF Edge program, Rosenberg said: 'There is a surprising correlation between oil prices and Treasury yields, but it suddenly stopped in October.'

Federal Reserve Chairman Kevin Warsh also mentioned the relationship between oil prices and bond yields at a press conference on FOMC Wednesday. He pointed out that the continuous rise in crude oil futures is a factor that central banks must take into consideration when fighting inflation. However, Rosenberg stated that Warsh did not discuss the “most important conclusion.”

"The correlation between this year's national debt and oil prices has reached 80%," said Rosenberg, which is considered "abnormally high" compared to the normal correlation. "We have basically been trading around oil prices, and there are sufficient reasons for doing so."

So what happened? Rosenberg believes that the answer can be found in the cooling down of AI trading. As the market gradually disengaged from the correlation between oil prices and treasury bond yields in October, "it tells us something important," he said. "In the eyes of the market, geopolitical risks are secondary; AI risks are the main concern."

Over the past month, this was most evident in the heavy selling on the NASDAQ market on Fridays, as well as earlier this week when Google recorded its largest single-day decline in history, losing a market value of $500 billion. Rosenberg stated that AI trades saw a pullback, and bond yields followed suit as the market equated the weakening of AI trades with a weakening of the overall U.S. economy. He also mentioned that last week's GDP report confirmed this view.

“AI is everything; it is omnipresent and affects the market at all times,” he said. Whether it’s the investment side of economic growth or the wealth effect on the consumer side, both are related to AI.

In the extended conversation of the podcast “ETF Edge”, Rosenberg said: “The third quarter’s GDP is an astonishing figure, and it’s all related to AI.” He pointed out that increased equipment investment and consumption exceeding expectations were both driven by the AI wealth effect; the latter is particularly noteworthy since actual income in the United States has been on the decline.

In the third quarter, corporate investment grew by 12.7%, which is the largest increase since 2011, excluding the pandemic period. Among this, investment in information processing equipment increased by 37%. Investment in industrial equipment grew by nearly 25%, and although this is not a direct AI transaction, it does reflect some of the expenditures on data centers.

At the press conference held on FOMC, Warsh did not agree with this view regarding the economy. He said, "It is convenient to attribute strong economic performance to this wave of new technologies. In fact, if we look at the data, including that from the last meeting to this one, the story of capital expenditure is not only robust but also expanding. Therefore, it is much more difficult now to claim that this strong performance is limited to a few narrow sectors, as it was just a few quarters ago."

However, Rosenberg indicates that the market in October did not come to such a conclusion. Overall, he believes that any “volatility” in AI transactions will be seen as volatility in the prospects for U.S. growth and is deflationary. In a deflationary environment, “there is only one thing that can hedge against it,” he said, and that is “the long end of the Treasury yield curve, which is something that no one really wants…”

"This is the truly important lesson brought about by the market performance over the past month," he said.

Nevertheless, Rosenberg remains cautious about interpreting this development as a safe bet on bonds. "In simple terms, this is not a return to normality – by normality, I mean bonds hedging stocks," he said.

The reason lies in the fact that bonds are subject to conflicting forces. Rising oil prices and inflation remaining above target levels will not disappear, and these challenges to fixed-income assets will also undermine the 'effectiveness' of fixed income as a hedging tool, says Rosenberg.

"Inflation is too high, which limits the Federal Reserve's ability to provide easing measures, and we haven't yet overcome that," he said. "But what we've learned over the past month is that if the scale is large enough, and if the AI trend begins to be unfavorable for the macroeconomy and inflation trends, it can trigger a comprehensive change in direction, leading to deflation, and bonds may react in the same way as they did in October."

However, he stated that investors should not bet on such transactions developing in a one-way direction, nor should they assume that there are no costs involved. Investors need to weigh the following: on one hand, the diversified returns of bonds as shown by the market in October; on the other hand, the scenario where AI does not drag down the economy. Moreover, the longer AI remains prosperous, and coupled with the broader fiscal challenges in the United States, "bonds will be put to an even greater test," he said.

How to trade in the debt boom of AI – the super-large cloud service providers

AI Trading itself is an important component of the bond market, and over the past week, Google's financial report followed by a decline in its stock price, as well as Microsoft's return to the public bond market, have made this point particularly evident. "This came as a surprise to the market. Microsoft hasn't been in this market for nearly a decade," said Rosenberg.

As the capital expenditure needs of tech giants increasingly exceed their free cash flows, the public bond market is gradually absorbing the expected flood of new issuances. "When the bond market, specifically the public bond market, sees where all this money is coming from... that represents a huge number of new issuers, and more and more funds are flowing here time and again," he said.

Super-large cloud providers are approaching the upper limits of financing that bank loans and private credit markets can provide. Rosenberg emphasizes that the global bond market, worth $40 trillion to $60 trillion, can absorb the expected $4 trillion in AI debt over the next four years, but "the market's capacity is not unlimited."

This demand will lead to a decline in prices and an widening of interest rate spreads. However, Microsoft's offering has attracted a large amount of investor interest and resulted in over-subscription.

Rosenberg indicates that it is a good thing that more large-scale cloud vendors are pushing their debts into the public bond market, as this means that information such as bond prices and credit default swaps (CDS) spreads will become more transparent. Currently, the spread trading for Microsoft's credit default swaps is near a five-year high, at around 55; in contrast, Oracle's (CDS) spread is close to 300. For those investors who do not want to directly buy bonds in an inflationary environment and amidst the fiscal challenges of the United States but still wish to hedge their bets on (AI) prosperity, this provides another trading opportunity.

"That's an old trick from the time of the great financial crisis," he said, "short it. You can trade in both directions on the [public] credit market, which is not possible with bank loans and private credit. But once they enter the public market, you can do that."

Rosenberg said that Google's 12% decline in a single day "is not even recorded on the charts historically." "So it is entering a different 'zip code area,' if you will, which means different market pricing, market volatility, and market digestion ranges. This is not the end of the world; it's just that AI has entered a different realm," he said, noting that the higher transparency of the public bond market gives investors the ability to better trade amidst the emotional fluctuations between fear and greed.

In the view of Rosenberg, the message conveyed to investors by the CDS market corresponding to the names of large, ultra-large-scale cloud vendors is not that the default risk is very high, but rather that as all companies attempt to obtain more debt financing, the ability of banks and financial intermediaries to find counterparties for these risks is becoming more limited.

"Think about financial markets and these cycles of large-scale investment prosperity. They don't move in a straight line in one direction; instead, they rise step by step and then reset." However, during this current cycle of prosperity, only the "positive aspects" have been observed so far, says Rosenberg. "This is still an extremely positive story, but you will begin to see a bit of the balance brought about by some AI pessimistic sentiments."

Subscribe to our weekly newsletter, and in addition to live content, you will also gain a deeper understanding of the trends and data that shape the ETF market.