Over the past month, oil prices have risen, yet in a departure from the 80% correlation that had persisted throughout this year, US Treasury yields moved lower over the same period. What changed? According to Jeffrey Rosenberg, Senior Portfolio Manager at the world's largest asset manager BlackRock's Systematic Alternatives ETF, it was the decline of leading AI stocks—the Magnificent Seven.
"Oil and Treasury yields had this amazing correlation, and then in October it just stopped," he said. "That tells us something important. In the market's mind, geopolitical risk is secondary; AI risk is primary."
A divergence after an FOMC meeting
If there was one thing that held steady in 2026, it was that US Treasury yields would climb alongside rising crude oil futures contracts. Over the past month, that pattern came to an end.
In the view of Jeffrey Rosenberg, who works at the world's largest asset manager BlackRock's Systematic Alternatives ETF (IALT), the breakdown of this trade carries major implications for the market. "Oil and Treasury yields had this amazing correlation, and then in October it just stopped," Rosenberg said.
The relationship between oil prices and bond yields also surfaced at Fed Chair Kevin Warsh's FOMC press conference on Wednesday—Warsh noted that persistently rising crude oil futures were a factor the central bank must consider in its fight against inflation. But Rosenberg said what Warsh did not address was "the most important conclusion."
"US Treasuries and oil had an 80% correlation this year," Rosenberg said, which is "absurdly high" relative to normal levels. "We've basically been trading off oil, and for good reason."
What happened?
Rosenberg said looking at the deflated AI trade sums it up. As the market broke away from this oil-bond trading correlation in October, "it tells us something important," he said. "In the market's mind, geopolitical risk is secondary; AI risk is primary."
Over the past month—most painfully illustrated by Friday's Nasdaq selloff and Alphabet's largest single-day decline in its history earlier this week, which wiped out roughly $500 billion in market value—the AI trade has undergone a pullback. Rosenberg said bond yields moved lower because the market equated weakness in the AI trade with broader weakness in the US economy. He also said last week's GDP report supported that view.
"AI is everything, everywhere, all at once," he said. Growth on the investment side of the economy, as well as the wealth effect on the consumption side, is all AI, he said. "Third-quarter GDP was a blowout number, and it was all AI-related," Rosenberg said. He pointed to increased equipment investment and far-better-than-expected consumption, both driven by the AI wealth effect—particularly notable given that US real incomes have been declining. Third-quarter business investment grew 12.7%, the largest increase since 2011 excluding the pandemic period in 2026, led by a 37% jump in information processing equipment investment. Industrial equipment investment rose nearly 25%, which, while not directly an AI trade, partly reflects data center spending.
Warsh disagreed with Rosenberg's assessment of the economy. At the FOMC press conference, he said: "It's easy to attribute all of this economic strength to this new wave of technology. In fact, if you look at the data—really the data from the last meeting to this one—the capex story isn't just strong, it isn't just powerful, it's broadening. So if you want to tell a story that the strength we're seeing is only in a few narrow areas, that story is much harder to tell now than it was a few quarters ago."
Indeed, market breadth—a measure of how many S&P 500 stocks are showing positive trading momentum—improved significantly in October, after touching a 2026 low at the end of September, alongside strength in defensive sectors such as healthcare and consumer staples.
But Rosenberg said the overall conclusion the market drew in October was that any "wobble" in the AI trade would be seen as a wobble in the US growth outlook—and one that is disinflationary. And in a disinflationary environment, "only one thing works as a hedge," he said, and that is "the thing everyone loves to hate…the long end of the Treasury curve." "That's the truly important lesson from last month's market performance," he said.
Despite holding this view, Rosenberg remains cautious about the claim that this dynamic makes bonds a safe choice. "The simplest answer is that this isn't a return to normal—normal used to be bonds hedging stocks," he said. The reason is that the forces facing bonds are conflicting. Rising oil prices, and inflation that remains above target, will not go away as challenges to fixed income assets and to the "effectiveness" of fixed income as a hedge. "Inflation is too high, which limits the Fed's ability to provide easing, and we haven't escaped that yet," he said. "But what we learned last month is that if it's big enough—if the AI trend shifts in a direction that works against the macroeconomic and inflation trends, it can bring about a complete reversal in that direction, creating disinflation, and bonds could react as they did in October."
Still, he said investors should not bet that this trade will be one-way or without potential costs. Investors need to weigh the diversification benefits bonds may offer based on October's moves against the scenario in which AI does not fracture the economy. And the longer the AI boom persists, combined with broader US fiscal challenges, "the more pressure bonds will face," he said.
How to trade the AI hyperscalers' debt boom
The AI trade itself is a very large part of the bond market, something made vividly clear over the past week—with Alphabet's post-earnings decline on one side and Microsoft's return to the public bond market on the other. "This was a surprise to the market. It hadn't come back for nearly a decade," Rosenberg said.
As tech giants' capex needs begin to exceed their free cash flow more frequently, the public bond market is absorbing an expected flood of issuance. "When the bond market, the public bond market, looks at where all this supply is coming from…this is a massive marginal issuer, and it keeps coming, again and again," he said.
Hyperscalers are approaching the financing limits that bank loans and private credit markets can provide; Rosenberg emphasized that the $40 trillion to $60 trillion global bond market can absorb the estimated $4 trillion in AI debt expected over the next four years, but "the market's capacity is not unlimited."
That demand will lead to lower prices and wider spreads, even though Microsoft's deal attracted substantial investor interest and was oversubscribed. Rosenberg said it is a good thing that more hyperscaler debt is being pushed into the public bond market, because it means more transparency in the form of bond prices and CDS spreads.
Microsoft's (MSFT.US) credit default swap (CDS) spread is currently trading near a five-year high of about 55; by contrast, its CDS spread is worlds apart from Oracle's (ORCL.US), which is near 300. This offers investors another type of trade: those who, given the US inflation environment and fiscal challenges, may not want to buy bonds directly but still want a hedge to bet on the AI boom. "It's that old tool from the great financial crisis," he said. "Short it. In the public credit market you can trade both ways, and in bank loans and private credit you can't. But once they start entering the public market, you can."
Rosenberg said Alphabet's (GOOGL.US) single-day 12% decline "isn't even on the chart historically." "So it's entering a different 'zip code'—if you will—a different zone of market pricing, market volatility and market indigestion. It's not the end of the world, just a different world during the AI buildout," he said, adding that greater transparency in the public bond market gives investors more ability to trade "the emotional swing between fear and greed."
Rosenberg believes what the CDS market is telling investors about large hyperscalers is not that default probability is very high, but rather that as all these companies try to obtain more debt financing, banks' and financial intermediaries' ability to find counterparties for these risks is becoming increasingly constrained. "Think about financial markets and these massive investment booms. They don't move in one direction. They go up in steps, then reset." But as far as the AI boom is concerned, so far there has only been "the positive part," Rosenberg said. "It's still an extremely positive story, but you'll see some balance against AI pessimism."
Comments