BlackRock’s Rosenberg says AI fears broke the oil–bond yield link in October

Weakness in AI stocks is becoming a warning about the wider U.S. economy, not just technology portfolios, BlackRock portfolio manager Jeffrey Rosenberg argues. In a CNBC discussion published November 1, he pointed to October’s split between rising oil prices and falling Treasury yields as evidence that growth fears had overtaken another powerful market concern.

October’s break with the oil trade

Rosenberg, a senior portfolio manager for BlackRock’s Systematic Alternatives ETF, told CNBC’s ETF Edge that oil and Treasury yields had an approximately 80% correlation during 2026. They had largely moved together. In October, that relationship stopped holding as leading AI stocks retreated and yields moved lower despite rising oil futures.

His explanation centers on two channels: spending on AI infrastructure and consumption supported by gains in AI-related wealth. If confidence in that trade falters, he argues, investors see a threat to both business investment and household spending. That makes an AI selloff a potential signal of weaker growth and easing inflation.

The investment numbers help explain his concern. Third-quarter U.S. business investment rose 12.7%, its largest increase outside the pandemic since 2011. Investment in information-processing equipment increased 37%. Industrial equipment investment rose nearly 25%, a gain CNBC described as partly reflecting data-center spending rather than being a pure AI measure.

That interpretation is contested. At the October 28 Federal Reserve press conference, Chair Kevin Warsh argued that capital spending was broadening, making it harder to describe economic strength as confined to a few technology-driven pockets. Rosenberg’s assessment concerns what October’s markets appeared to believe, not a settled finding that AI alone powers growth.

A hedge with inflation still attached

For Rosenberg, October showed why longer-term Treasuries can offer protection when AI weakness threatens the economy and reduces inflation pressure. But he explicitly rejected treating the month as a return to a dependable pattern in which bonds offset stock losses. The forces working against bonds have not disappeared.

AI borrowing brings a more visible test

The same discussion examined how AI’s financing needs are changing credit markets. Rosenberg said major cloud companies’ capital spending is increasingly overwhelming their free cash flow. As those companies approach the financing limits of bank loans and private credit, he expects more borrowing to move into public bond markets.

He said a global bond market of $40 trillion to $60 trillion could absorb an expected $4 trillion in AI-related debt over four years. That is a forecast, not completed borrowing. He nevertheless warned that capacity is finite and that repeated large offerings would pressure bond prices and increase the extra yield investors demand.

Microsoft’s return to public bond markets after nearly a decade illustrated the shift, although its offering attracted more demand than available bonds. Rosenberg welcomed the greater transparency of public borrowing: bond prices and credit-default swaps, which provide protection against default, give investors visible ways to price and trade concerns about the buildout.

He cautioned against reading those credit-market signals simply as predictions of default. In his view, they also reflect financial intermediaries’ growing difficulty finding counterparties willing to take the risk as more companies seek debt financing. Public markets make that constraint more visible, even while the underlying AI investment story remains positive in his assessment.