Bloomberg ran a story on 29 October 2026 headlined “Investors May Have Bought All the AI Bonds They Can Handle”, a day after its report that the highest-rated data center bonds now yield closer to 7%, up from below 5% at the start of 2026, according to Morgan Stanley strategist Carolyn Campbell.
By our arithmetic, that move adds roughly $20 million a year in interest for every $1 billion borrowed, from under $50 million to about $70 million. At the 8% level Campbell says makes issuing these bonds “significantly more challenging”, the bill would be $80 million.
Short explainer · AI data center debt
A vertical news Short narrated from the published headlines and opening lines of this piece and two related FourWeekMBA pieces.
Business Pill · Tranches
An evergreen Business Engineer explainer on tranches: how one financing is split into layers that take losses in a set order. It explains the general idea only and does not discuss the deals or yields in this piece.
The key insight: By our arithmetic, Bloomberg’s figures put the yearly interest on $1 billion of top-rated data center bonds at about $70 million, up from under $50 million at the start of 2026, roughly $20 million more.
What Bloomberg Reported
We could read only the headline of the 29 October story. Bloomberg’s 28 October report, by Neil Callanan, Caleb Mutua and Paula Doenecke, says “Early signs suggest that the credit market is approaching the limits of how much more debt it can absorb”.
Bloomberg says it spoke with more than 30 bankers, money managers, private equity executives and others in recent and current data center financing talks, and that “all acknowledged that raising cash has become significantly more difficult than it was just months ago”.
Bloomberg also reports that tech giants arranged tens of billions of dollars of new financing in the week of its report, and that Citadel Securities “isn’t predicting an imminent credit crunch”.

The Price of the Money
“At the beginning of 2026, the highest-rated data center bonds offered yields below 5%. Today similar bonds are yielding closer to 7%, and that increase matters enormously for companies trying to finance multibillion-dollar facilities”, Campbell said in a podcast, according to Bloomberg. She noted that data center debt issuance in the asset-backed and commercial mortgage-backed markets totaled roughly $22 billion this year.
Lower down the credit scale, Bloomberg says at least 18 data center financings in the junk-bond and leveraged loan markets now carry double-digit yields, according to data it compiled. JPMorgan Chase and Goldman Sachs offered steeper discounts and higher interest payments on neocloud Volta Infrastructure Holdings’ $5 billion syndicated loan, pushing its yield above 11%.
The average yield on bonds in the Bloomberg Investment Grade Technology Index is about 6%, about 1.4 percentage points higher than a year ago, Bloomberg reports. That puts the year-ago level near 4.6%, by our arithmetic.

Who Is Pulling Back
Oaktree co-chief executive Armen Panossian said in a client webcast, as quoted by Bloomberg: “We’re seeing in the last several weeks a widening of those spreads that indicates that the credit markets are becoming either full, saturated of the loans that are out there, or concerned about some of these factors that are lurking in the background”.
Bloomberg reports that a large European bank’s head of technology, media and telecom lending said the bank is approaching the limit of what it can lend to the sector, and that selling the debt or transferring the risk can take five months. Impax Asset Management is underweight data center debt, portfolio manager Tony Trzcinka told Bloomberg.
Bloomberg also reports that the cost of insuring Oracle’s debt reached a fresh record last week, as did default protection costs for SpaceX, Nvidia and Meta.
Update, 29 Oct: a lender now wants insurance on its own data center loans. Bloomberg reported on 29 Oct that HSBC is seeking to hedge data center debt with credit risk insurance; only the headline was readable to us, so the size, insurer and terms are not known. In September, Odaily reported that HSBC was in preliminary talks with investors on a significant risk transfer covering more than $1 billion of loans used to finance AI infrastructure such as data centers, citing people familiar with the matter, and said terms could still change.
A Second View: Repricing, Not Rejecting
J.P. Morgan Asset Management took the other side of the question in an August note headed “Credit markets are repricing, not rejecting, the AI buildout”. It said hyperscaler spreads “have widened by around 30 basis points (bps) year to date, compared with 2 bps widening in the broader USD investment-grade index”.
The firm estimated that hyperscalers “could issue USD 300 billion in investment-grade bonds over the next year alone” and said: “We do not expect this new wave of supply to derail the market, but it is reshaping it”.
Connecting the Coverage
The first part of Bloomberg’s 28 October report, and our count of October’s AI compute debt, are in our piece on the $103 billion of leases. The record borrowing behind it is in our piece on $6.6 trillion of bond sales. A loan that banks are holding longer is in our report on the Texas loan, and the neocloud balance sheets most exposed to the cost of money are in our CoreWeave debt piece and our Nscale IPO piece.
The Structural Read
As we read Bloomberg’s report, the pressure runs across the ratings scale: double-digit yields on at least 18 junk and leveraged-loan financings, an investment-grade technology index yield that has risen about 1.4 points to about 6%, and record insurance costs on several of the biggest names.
As we read Bloomberg’s report, the limit described is capacity as well as price: a European bank says it is approaching its lending limit to the sector and that selling the debt or transferring the risk can take five months, while insurers, according to a Willis Towers Watson survey Bloomberg cites, were indicating months ago that they were nearing or had reached their data center exposure limits.
As we read J.P. Morgan’s August note, it called the widening a repricing of heavy supply rather than a rejection of the buildout, and its 30 basis points was a year-to-date figure for hyperscaler spreads.
Carolyn Campbell, Morgan Stanley, as quoted by Bloomberg
“When borrowing costs approach 8%, issuing these bonds becomes significantly more challenging.”
Three Implications
FOR DATA CENTER DEVELOPERS By our arithmetic, at 8% the yearly interest on $1 billion would be $80 million, the level Campbell says makes issuing these bonds significantly more challenging.
FOR NEOCLOUDS Bloomberg reports Volta Infrastructure’s $5 billion loan priced above 11%, and that offerings from Elice, Firmus and Nscale stumbled.
FOR HYPERSCALERS Bloomberg reports record costs to insure the debt of Oracle, SpaceX, Nvidia and Meta, and Citadel’s Tucker Roberts says chipmakers and hyperscalers could shoulder more of the risk.
The Business Engineer Lens
This story maps onto the Business Engineer framework The AI Capex Map.
The framework sets out to map the financing of AI: “How the money flows, where it cracks, what it buys, who can absorb it, and what happens when revenue can’t catch up.”
As we read it, Bloomberg’s report is about the “who can absorb it” part: the yield buyers now ask for data center debt. The framework was published in May 2026 and does not discuss these figures.
What Is Not Established
We did not read the 29 October Bloomberg story beyond its headline, so its figures and sources are not reflected here. The yields are as Bloomberg reports them, rounded; real coupons, fees and discounts differ by deal.
The $20 million, $80 million and 4.6% figures are our arithmetic on those rounded yields, for one year, and do not reflect any specific company’s bonds. J.P. Morgan’s view dates from 20 August 2026, before the moves Bloomberg describes in October.
The Bottom Line
Bloomberg reports that the highest-rated data center bonds yield closer to 7%, up from below 5% at the start of 2026, and headlined on 29 October that investors may have bought all the AI bonds they can handle. By our arithmetic, that is roughly $20 million more a year in interest for every $1 billion borrowed.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Only the headline of Bloomberg’s 29 October story was read. Yields, deals and quotations come from Bloomberg’s 28 October report as republished in full by Yahoo Finance; the second view comes from J.P. Morgan Asset Management’s note of 20 August 2026. Interest per $1 billion and the index’s year-ago yield are our arithmetic; linked earlier pieces carry their own sources.
Sources: Bloomberg (via Google News): Investors May Have Bought All the AI Bonds They Can Handle (29 Oct 2026, headline) · Bloomberg via Yahoo Finance: Wall Street’s AI Debt Machine Is Starting to Choke on Supply (28 Oct 2026) · J.P. Morgan Asset Management: Hyperscalers, now also a credit story (20 Aug 2026)









