CoreWeave’s debt other than its convertible notes carried an average effective interest rate of about 9.9% at 30 June, by our arithmetic from the line-by-line table in its second-quarter 10-Q. Its finance chief’s guide for the next investment-grade loan, a 7% to 7.5% yield, is 2.4 to 2.9 percentage points below that.
With the two low-coupon convertibles included, the average was about 8.4%, by our arithmetic, on $35,551 million of principal. Of the 16 lines in the table, only the convertibles, at 2%, sit below the guide, and only the investment-grade DDTL 4.0, at 7%, sits at its low end.
Business Pill · Neoclouds Explained
An evergreen Business Engineer explainer on neoclouds, which rent out AI chips by the hour. It teaches the general idea only and does not discuss CoreWeave’s debt.
The key insight: Set the convertibles aside and CoreWeave’s debt cost about 9.9% at 30 June, by our arithmetic from its 10-Q; with them, 8.4%. The CFO’s 7% to 7.5% guide for the next investment-grade loan is 2.4 to 2.9 points under the non-convertible average and below every non-convertible line except DDTL 4.0.
The 10-Q lists each facility with its maturity and effective interest rate. By our tally of that table, the parent-guaranteed contract loans, DDTL 1.0 to 5.0, were the largest group at $10,806 million, 30.4% of principal, at a weighted 10.3%. Five unsecured senior notes came next at $10,029 million, 28.2%, at a weighted 9.8%.
The two convertible notes due 2031 and 2032 were $6,588 million, 18.5%, at 2%. OEM and software license financing was $5,102 million, 14.4%, at a weighted 10.7%. The non-recourse, investment-grade DDTL 4.0 was $2,837 million, 8.0%, at 7%, and a loan from Magnetar was $189 million at 12%.
The principal adds up to $35,551 million, the same total as the 10-Q’s schedule of future principal payments, by our sum. The rates are the 10-Q’s effective interest rates, printed to the whole percent, so the averages are approximate.
CoreWeave’s debt principal by instrument group at 30 June 2026, with each group’s principal-weighted effective interest rate, from Note 10 of its second-quarter 10-Q; the 30 September total is from its Q3 release and its split awaits the Q3 10-Q. Group rates and the 8.4% and 9.9% averages are our arithmetic.Effective interest rate on each CoreWeave debt line at 30 June 2026, from Note 10 of its second-quarter 10-Q, against the CFO’s 7%-7.5% yield guide for the next investment-grade DDTL (10 November call). The 9.9% line is our arithmetic.
Where the Cost Sits
The most expensive line was DDTL 1.0 at 15%, followed by the Magnetar loan at 12% and DDTL 2.0 and recourse OEM financing at 11%. The early loans are shrinking: DDTL 1.0 fell to $1,300 million at 30 June from $1,553 million at the end of 2025, and DDTL 2.0 to $3,190 million from $5,037 million, the 10-Q shows.
Chief executive Michael Intrator said on the 10 November call that this includes “paying down nearly half of our first two delayed draw term loans, which are our most expensive facilities”. Every unsecured note carried 9% or 10%.
Against that table, the CFO’s guide is cheaper than every non-convertible line except DDTL 4.0. Nitin Agrawal said the next DDTL should “price at a spread that implies a 7% to 7.5% yield based on current SOFR while financing more than 90% of the associated CapEx”. Our explainer on how CoreWeave borrows covers why contract loans and parent notes are priced so differently.
What Changed by 30 September
The facility split for 30 September is not yet public: the third-quarter 10-Q was not on EDGAR when we checked at 00:05 UTC on 12 November. The results release gives the total, which our loan piece summed from its balance sheet; it is $11,805 million higher than the $35,068 million carrying value at 30 June, by our arithmetic.
Two Q3 deals are named in the filings. CoreWeave “Raised approximately $4.2 billion through an upsized offering of convertible senior notes”, the release says, priced at 2.875% and due 2033, per its 18 September release. It also closed the $2.6 billion DDTL 5.5 at SOFR + 5.50%, of which it had drawn $1.2 billion by the 10-Q’s filing in August.
Intrator said that at the end of Q3 the “weighted average cost of debt across our entire business was 7.9%, down 250 basis points year-over-year”. That is the company’s own measure; the call did not give its method, so it is not directly comparable with our 8.4% from the 30 June table.
The schedule also shows how soon the stack turns over. Of the $35,551 million of principal at 30 June, $10,597 million, or 29.8%, was due by the end of 2027, by our arithmetic from the 10-Q. Our call piece covers the company’s plan to fund that period without new parent-level unsecured debt at current yields.
The Structural Read
As we read the 10-Q table, the low headline cost leans on the convertibles: $6,588 million at 2% pulls a 9.9% non-convertible average down to 8.4%, by our arithmetic, and convertibles carry a share-conversion feature that the other lines do not.
As we read the guide against the table, an investment-grade contract loan at 7% to 7.5% would be cheaper than every DDTL CoreWeave had before DDTL 4.0, and than every senior note, which is the economics behind the pledge to avoid new parent-level unsecured debt at current yields.
Michael Intrator, CoreWeave CEO, Q3 2026 earnings call, 10 Nov 2026
“For the facilities funding long-term commitments, by the time the customer contract ends, the debt is repaid, and the infrastructure is owned free and clear by CoreWeave”
Three Implications
FOR THE COST OF DEBT Each new loan at 7% to 7.5% would sit below the 9.9% non-convertible average at 30 June, by our arithmetic; the 10-Q does not show where Q3 left it.
FOR REFINANCING $10,597 million of principal, 29.8%, was due by the end of 2027 at 30 June, by our arithmetic from the 10-Q schedule.
FOR THE NEXT FILING The Q3 10-Q, which the release says is filed or to be filed with the SEC, is where the 30 September table, including the 2033 convertibles and DDTL 5.5, would appear.
The framework reads: “The bubble can pop on the short cycle while the supercycle continues uninterrupted on the long cycle, exactly the way the dot-com bust did not stop the internet from becoming the substrate of the next thirty years.”
As we read it, CoreWeave’s table holds both clocks: contract loans that amortise over customer terms of a few years, and parent notes and convertibles due 2030 to 2032 that are priced on the company’s own credit, per its 10-Q. The framework predates the filing and does not discuss it.
What Is Not Established
The 30 September split by instrument is not established until the Q3 10-Q is filed. The weighted averages use whole-percent effective rates and principal, not carrying values, and are our arithmetic. The 7% to 7.5% figure is a guide for a loan not yet signed; whether it is the $4 billion loan Bloomberg reported is not established. This is not investment advice.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
CoreWeave’s non-convertible debt carried about 9.9% at 30 June, and 8.4% with the convertibles, by our arithmetic from its 10-Q. The CFO’s 7% to 7.5% guide for the next investment-grade loan is below every non-convertible line in that table except DDTL 4.0. How the $4.2 billion of 2033 convertibles and DDTL 5.5 changed that mix is in the Q3 10-Q, not yet on EDGAR.
A note on sourcing. Principal, maturities and effective interest rates by facility are from Note 10 of CoreWeave’s Form 10-Q for the quarter ended 30 June 2026, and the DDTL 5.5 draw from its Note 16. The 30 September balance sheet and the Q3 financing items are from CoreWeave’s third-quarter results release of 10 November 2026; convertible and DDTL 5.5 terms are from its 18 September and 10 August releases; quotes are from its 10 November earnings call (proofread transcript).
Group totals, weighted averages, the gap to the guide and the share due by 2027 are our arithmetic. We did not contact CoreWeave for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
