CoreWeave’s stock and its unsecured bonds are two bets on the same buildout. The stock offers the upside if the company turns contracted AI demand into lasting shareholder value. The bond offers a fixed payment schedule—provided the company can keep paying.
That distinction matters when the stock is falling sharply and the debt is trading at yields associated with substantial credit risk.
The selloff is much larger than the indexes
At 12:25 p.m. ET on October 14, CRWV was observed at $65.15, down 16.26% from its October 13 regular close of $77.81. SPY was down 1.10% and QQQ 1.81% at nearly simultaneous observations. CoreWeave was experiencing a much larger decline than those indexes; the comparison does not measure how much of the move came from any particular headline.
September CPI, released this morning, rose 0.6% month over month and 3.7% year over year. Core CPI rose 0.3% and 2.6%, respectively. That keeps financing conditions relevant to a business investing heavily ahead of revenue. It does not prove CPI caused CoreWeave’s entire decline. BLS release.
There is also a distinct credit story. Bloomberg reported at 11:29 a.m. ET on October 14 that CoreWeave’s unsecured 2032 debt traded below 80 cents on the dollar, citing TRACE, with a yield above 15%. Earlier reporting described the $1.25 billion 2032 issue touching roughly 15% on October 13 and ending near 14.5%. We treat yesterday’s figures as reported observations, rather than an independently verified closing trade. Bloomberg report, October 13 yield reporting.
Which bond are we comparing?
The focus is CoreWeave’s US-dollar 9⅝% senior notes due July 15, 2032, CUSIP 21873SAK4, ISIN US21873SAK42. The issuer documented $1.25 billion of this series in June. It matches the amount and maturity in the reported October 13 story. It is distinct from the 9.75% October 2031 notes, the euro-denominated 2032 notes and the convertibles. June financing filing.
The fixed coupon pays $962.50 per full year for $10,000 face value, while outstanding and assuming payment. Payments are scheduled for January 15 and July 15, starting January 15, 2027. That first payment covers a longer initial period, so it is larger than a normal half-year coupon. The notes are callable and subject to transfer restrictions; availability to an individual investor depends on the security, investor eligibility and broker. A publicly visible yield is not an executable retail offer. Indenture and restricted-note terms.
Crucially, a 15% secondary-market yield does not reset the coupon to 15% or mean CoreWeave just borrowed all its money at that rate. It reflects the price investors pay for existing promised cash flows. It can nevertheless signal that raising replacement capital may be more expensive.
What a yield near 15% is actually buying
Our illustrative model discounts the note’s promised payments using its actual first coupon and 30/360 accrual. With assumed October 15 settlement, a 15% nominal annual yield to maturity implies a clean price near $79.70 per $100 face. At 12%, the model gives $90.28; at 20%, $65.37. These are hypothetical sensitivity calculations, not observed trades, forecasts or a verified yield to worst.
At the 15% model price, $10,000 face costs approximately $7,969.63 before accrued interest and fees, or $8,282.44 including the model’s accrued interest. The annual coupon remains $962.50. The rest of the modeled yield depends partly on being repaid at face value, with all promised payments made. Selling early, default, fees, reinvestment and early redemption can change the realized result.
There is bond-price upside if credit conditions improve, but the downside is real. A purely illustrative recovery of $40 per $100 face against an $80 clean purchase price halves the principal invested before interest, accrual and costs. That is an example of loss severity, not a forecast of recovery or probability of default.
Strong demand, expensive delivery
CoreWeave’s Q2 revenue reached $2.575 billion, up 112.5% year over year. It reported $1.510 billion of adjusted EBITDA, alongside a $49 million GAAP operating loss, $640 million of net interest expense and a $626 million net loss. Adjusted EBITDA is useful for operating context, but it is not cash left for shareholders after infrastructure investment and financing. Q2 results.
First-half operating cash flow was $3.663 billion, versus $14.117 billion of cash PP&E purchases: a $10.454 billion gap under this simple cash definition. Financing cash flow supplied $13.985 billion. Operating cash flow also benefits from customer-prepayment timing. Cash PP&E purchases differ from the issuer’s capex measure, so those numbers should not be mixed. Cash-flow statement.
The June backlog was approximately $104 billion, subject to delivery and service-availability requirements. That is evidence of committed demand, not $104 billion of cash available to repay debt. The three largest customers represented 36%, 26% and 10% of Q2 revenue. Concentration remains relevant even as the customer base broadens; the filing does not identify those customers by name. Q2 filing.
Higher priority does not mean secured
At June 30, principal debt was $35.551 billion, compared with $5.524 billion of cash and cash equivalents. Reported liquidity of $15.553 billion included $10.014 billion of undrawn secured borrowing availability. These are quarter-end figures. They exclude a complete reconciliation of later borrowing, repayments and spending and must not be described as today’s cash or debt balances. June balance sheet and debt notes.
September’s $4.2 billion convertible financing is a later event. The filing reports approximately $4.137 billion of net proceeds before estimated offering expenses. Separately, $566.2 million was spent on capped calls. Issuance size is therefore not the same as incremental cash available for construction. September closing filing.
The 2032 notes sit ahead of common equity, but are unsecured. Secured lenders have priority over their collateral, and subsidiary structures and guarantees affect what is available to different creditors. Owning these notes does not give an investor a direct secured claim on the entire GPU fleet. A large asset base alone cannot establish recovery value. Collateral and debt structure, note guarantees.
What the stock chart and insiders add
The $65.15 snapshot was below the completed-session 50-day SMA of $88.44 and 200-day SMA of $92.58. Nearby historical downside references were the July 28 intraday low at $64.63 and July 29 low at $60.55. A break below the former would bring the latter into focus; neither is guaranteed support. A recovery first needs to reclaim lost levels, including the October 13 intraday low at $76.11. For a daily retest setup, the test is a daily candle that retests and closes back above the level.
The larger chart starts at the listing, rather than just the rebound. The June 20, 2025 closing peak of $183.58 fell to $60.82 on July 29, 2026, a 66.9% decline. The rebound then reached $107.73 on August 12, before fading to $77.81 by October 13. The October 14 snapshot remained approximately 64.5% below that earlier closing peak. A smaller price is not, by itself, proof of better value. Price history, with independently calculated returns and averages.
For July 14–October 14, our Insider Analyzer counted $506.12 million of eligible reported common-stock sales, with 519 sale rows, 498 referencing plans, and no eligible P-coded purchases in the included rows. Same-date transactions are aggregated in the chart. Sales can reflect plans, taxes or diversification; their dollar value alone does not establish conviction or explain today’s move.
The figures exclude 99 rows affected by an unresolved amendment. Grants, exercises, derivatives and F-coded withholding are separate from these P/S totals; S-coded sales to cover taxes remain included. One example is Agrawal’s September 30 five-share sale, explicitly identified as covering tax withholding. The scan describes the included filings, not a guarantee of complete insider activity. Agrawal Form 4. Plan-linked examples include Intrator’s September 29 transactions, filed October 1, and McBee’s August 24 transactions, filed August 26.
The decision is about the claim you want to own
The stock case improves if CoreWeave brings contracted capacity online, converts it into durable cash generation and reduces financing dependence without giving away too much of the upside through dilution. A price recovery without those improvements is a weaker investment case.
The bond case improves if the issuer can meet its obligations and financing risk eases enough to support its debt price. That offers a defined coupon and potential price recovery, with less participation in extraordinary business success than the equity. The credit case weakens if funding becomes harder to obtain, cash conversion disappoints or collateral priority leaves less recovery for unsecured holders.
Our read: a yield near 15% is a warning to examine funding and recovery, rather than a shortcut to safe income. The bond may suit a thesis about payment and credit normalization; the stock needs a stronger thesis about value left for shareholders after the buildout is financed. Neither thesis is settled by a rebound candle or a headline yield.
And remember, this is market research only, not a recommendation to buy or sell. USD unless stated. Market observations are dated October 14 at 12:25 p.m. ET; financial periods and reported bond observations have their own dates. Bond-model prices are illustrative and assume payment, not executable quotes.
Explore CoreWeave’s dated price levels, the Insider Analyzer, and other stock-versus-bond comparisons.