The week kicks off with investors looking to understand the details of the agreement to end the Iran war. While oil prices slid in immediate response to the news, most analysts do not expect a return to pre-war prices any time soon.
But there is a lot more going on in the markets beyond the war and last week’s SpaceX IPO. In this week’s Markets Brief, we look at the AI buildout’s growing impact on the bond market, including a warning that harkens back to the years leading up to the Great Financial Crisis. Plus a preview of what could be store as Kevin Warsh presides over his first meeting as Federal Reserve chair.
How AI-Related Issuance Is Changing the Corporate Bond Market
SpaceX SPCX wasn’t the only notable new offering last week. Few investors in the United States probably noticed this one; in Canada, Amazon AMZN sold CAD 14 billion (roughly $10 billion) in corporate bonds. That’s the largest new issuance in the history of the Canadian bond market. The country’s previous record deal was set just last month by Google parent company Alphabet GOOG/GOOGL.
Why should US investors care? Vanguard portfolio manager Thanh Nguyen notes that the issuance reflects the scope and scale of AI-related debt hitting the investment-grade corporate bond market in the US and other key markets. Nguyen drilled into bonds sold by hyperscalers—the companies with giant cloud computing businesses that are expanding massively to support AI. Debt sold in the US by Amazon, Alphabet, Facebook parent company Meta Platforms META, and Oracle ORCL totaled $141 billion from 2020 to 2024, for an average of $28 billion per year. (Nguyen notes that Microsoft MSFT actively raises money in the bond markets, but through different structures.)
Then, as the AI infrastructure boom took off last year, hyperscaler debt jumped to $93 billion in the US. With that came EUR 17 billion from Alphabet. The pace has only accelerated in 2026. There’s already been a 50% jump in issuance from 2025, with $107 billion in US dollar debt and another $62 billion in debt denominated in Canadian dollars, Swiss francs, euros, British pounds, and Japanese yen.
What’s more, Nguyen says Vanguard estimates that hyperscalers have this year only raised somewhere over half the funding they need. She added that AI-related investments could reach $800 billion, underscoring the scale of capital required. It is no longer just a bond story, funding is happening across the full capital structure, across public and private markets and multiple currencies.
Since all this debt has been widely expected, Nguyen says the market has absorbed it well. That said, it is changing the character of the corporate bond market. The investment-grade bond market has traditionally been more heavily concentrated in financials. But Nguyen says this debt could push the tech sector to take up a bigger slice of the corporate bond market than the so-called big six financial companies, which include JPMorgan JPM and Morgan Stanley MS.
This could impact how the overall investment-grade bond market behaves relative to the stock market, where returns are dominated by many of the same mega-cap tech stocks. “It’s definitely going to be more correlated,” Nguyen says.
Pimco chief investment officer Dan Ivascyn notes that technology comes with a riskier debt profile. “It’s going to lead to more volatility, more concentration, more correlation across names,” he says. “We at Pimco are quite cautious; we don’t want to be overweight in this risk.”
The Return of Financial Engineering
This past week, Pimco unveiled its annual secular outlook report. Along with the usual discussion of macro factors, the firm observed a return to some debt issuance practices that preceded the Great Financial Crisis in 2008.
Speaking about the report at a press conference on Thursday, Ivascyn emphasized this idea: “For the first time in about 15 years … although we don’t see systemic risks today … we’re beginning to see the pace of financial engineering accelerate to a point where it bears watching.” In particular, he pointed to a “dusting off of old playbooks” with certain areas of AI-related debt. This includes “rating agency arbitrage,” wherein issuers look for friendly rating agencies to certify risky debt as safe. Here’s more of what he had to say:
People are concerned about lower-quality lending, either mid-market direct lending or even in-bank loans. So there’s gonna be a strong desire to take inherently volatile and uncertain risk and find a way to make it investment grade.
Sometimes you can create investment-grade cash flows in a very sound manner. [But] there’s a desire again to create additional ways to structure risk, so there’s going to be the temptation to get AI-related deals into [collateralized loan obligations], to get these investment-rated deals into investment-grade specialty funds to get those yields higher to add a little bit of leverage to those particular funds. So now you have a complex structure that you got rating agencies to say is investment grade.
You can quickly find a situation where you have complex transactions sitting in funds next to other complex transactions with additional layers of leverage that then—sitting in CLO, CBO, CFO structures, whatever name you want to call them—are rated investor grade. As time passes from the prior crisis, people forget about these sort of things.
Dan Ivascyn, chief investment officer at Pimco
Warsh Steps Up to the Plate
This week’s Fed meeting will be the first with Kevin Warsh running the show, although Jerome Powell will still be in the room and voting on monetary policy. It appears to be a done deal that no change in interest rates will be forthcoming. More interesting will be any hints about what direction Warsh will take on policy and communicating with the public.
Speaking at the same event, Pimco global economic advisor Richard Clarida—who served as vice chair at the Fed from 2018 through 2022—says investors should expect the changes to come over time. “We haven’t had that many Fed chair transitions historically, but at least in my professional career, going back to the ’80s, it’s understandable that when a new chair comes in, there is a period, maybe measured in weeks or months, when you’re trying to get a sense of the regime and the communication,” he said. “We saw it in the handoff from Miller to Volcker and Volcker to Greenspan, and then Greenspan to Bernanke, and then Bernanke to Yellen. So you know there will be a handoff period.”