Goldman Sachs Launches AI Junk Bond Basket Trading Product
Claire Weston
Goldman Sachs this week rolled out a basket product covering 18 US high-yield bonds tied to AI and data-center issuers, with an average yield of 7.45% and a spread of 319 basis points — both above the broader high-yield market — signalling that Wall Street is now pricing AI-debt risk and building dedicated hedging tools around it.
What is inside Goldman's basket?
The basket holds 18 US high-yield bonds from AI and data-center issuers including CoreWeave, Applied Digital, and Cipher Digital.
Inquiry sizes range from $50 million to $250 million. Investors can buy the physical bonds outright or trade via a total return swap (TRS) — a derivative that transfers a bond's gains and losses without actual ownership.
This means → one trade lets an investor build or hedge exposure to the entire AI credit sector, instead of negotiating bond by bond.
How much extra yield does the market demand — and why?
The basket's average yield is 7.45% with a 319-basis-point spread. The broader high-yield market averages 7.3% and 267 basis points.
In plain terms = AI-linked bonds must pay roughly 52 extra basis points of interest to attract buyers — the market sees these issuers' repayment risk as higher than typical junk debt.
This reflects investors pricing in the "burn fast, earn later" profile of AI infrastructure projects.
Is JPMorgan doing the same thing?
JPMorgan launched three basket products the same week, each targeting a different risk tier:
Investment-grade, long-duration basket — covering 11 hyperscale cloud and project-finance issuers including Microsoft, Meta, Amazon, Alphabet, and Oracle.
High-yield AI basket — covering 15 issuers such as CoreWeave and Applied Digital; a third basket targets 16 semiconductor and hardware names including Nvidia.
This means → two of the largest investment banks entered the market almost simultaneously, a clear sign that hedging demand around AI debt exposure has reached a scale worth packaging into standardised products.
Why are these tools appearing now?
AI and data-center debt is expanding at an unprecedented pace, and bondholders need instruments to manage that exposure in bulk.
Shorecliff Asset Management founder Grant Nachman noted that basket products offer scale, breadth, and liquidity — long-only managers can adjust overall risk in a single trade, while hedge funds can toggle directional bets quickly.
JPMorgan had already launched a credit-default swap (CDS) basket tied to hyperscalers in February; Goldman told clients in March that a product for shorting the leveraged-loan market was not yet ready — that gap is now being closed.
What do the latest market signals tell us?
Alphabet's bonds came under pressure after this week's earnings — the company raised its full-year capex guidance and posted its first negative quarterly free cash flow in over twenty years as a public company.
Oracle's CDS protection cost rose to an all-time high, reflecting an extreme market repricing of its AI-spending burden.
In plain terms = even companies on the scale of Microsoft and Google are making bond investors nervous with their AI outlays — and that anxiety is precisely where demand for these hedging tools comes from. Whether the tools ultimately spread risk effectively will depend on the market's final verdict on when the AI capex cycle peaks.
Content is for reference only, not financial advice.