CoreWeave CRWV stock prediction: $192 bull, $113 base, $47…

CoreWeave trades at $88.01, down 52.1% from its all-time closing high of $183.58 — and that single number hides the fact that this stock has crashed more than 50% twice in fourteen months, with a 73% rally in between. The slide most coverage describes as a grind has been nothing of the sort. Our 12-month view: $192 bull, $113 base, $47 bear.

Here is the observation that should govern how you price the risk. On 29 July 2026, CoreWeave’s five-year credit default swaps blew out to roughly 855 basis points — a spread implying something close to a 50% probability of default over five years. The cloud ETF that holds it, SKYY, closed green that day. The selling was not sector-wide and it was not about AI demand. It was the credit market, alone, repricing one leveraged borrower. Two days later the stock rose 21.5%. That is the actual character of this equity: it trades as a credit instrument wearing a growth-stock costume.

Correcting the record on the peak

Almost every current write-up of CoreWeave anchors the drawdown to a 52-week high of $153.20, or a peak close of $143.08 in October 2025. Both figures are real, and both are rolling twelve-month numbers.

CoreWeave’s actual all-time closing high is $183.58, set on 20 June 2025, with an intraday print of $187.00 the same day. It falls outside the trailing year because the company only listed in March 2025. Measured properly, the stock is −52.1% from its high, not −42.6%, and the July trough was −66.9%, not −57.5%. If you are sizing a position off the drawdown, that nine-point difference matters.

Two crashes and a 73% rally, in order

The path is worth walking because each leg had a different cause, and only one of them was about AI demand.

11 November 2025, −16.31%. CoreWeave beat on Q3 but cut full-year 2025 revenue guidance to $5.05–5.15bn from $5.35bn. The reason was mundane and physical: a third-party data-centre developer was running late on the Denton, Texas site. Chief executive Michael Intrator said plainly that “a third-party data center developer is behind schedule.” JPMorgan downgraded from Overweight to Neutral and cut its target from $135 to $110 the same week.

27 February 2026, −18.51%. Full-year 2025 came in fine. The first-quarter guide did not: $1.9–2.0bn against roughly $2.29bn expected, a miss of about 17%. In the same breath, 2026 capex guidance more than doubled to $30–35bn. Spending up, revenue down — the worst possible combination for a business funded with debt.

Then the stock rallied +73.4% from 27 February to $137.98 on 6 May, back within 3.6% of the October peak. Whatever the market was worried about in February, it stopped worrying by May.

8 May 2026, −11.40%. Q1 revenue beat, but adjusted operating margin collapsed from 17% to 1% — $163m down to $21m — and the Q2 guide came in light. Capex went up again.

1 July 2026, −13.92%. Bloomberg reported that Meta was building a cloud business to sell surplus AI compute. CoreWeave’s largest new customer commitment had just become a competitor. Meta rose 9% the same day.

24 July 2026, −11.37%. Moody’s published a sector note warning that “unprecedented” AI spending threatened credit quality, naming CoreWeave specifically.

29 July 2026, −9.63% to the $60.82 low. The CDS blowout described above. Oracle’s five-year CDS pushed past 215bp from roughly 145bp at end-2025; even Nvidia’s hit new highs. We covered the mechanism in our report on CoreWeave’s rising cost of debt.

30 July, +21.51%. Microsoft’s fiscal Q4 landed: Azure up 43%, Microsoft’s best day since 2008, and more than $130bn of data-centre leases signed in a single quarter. The AI-capex trade was not dead after all.

3 August, +19.49%. Planned strikes on Iran were called off, and lenders had repriced the Anthropic-linked loan, taking the July credit panic off the table.

The 18 August fall was a bond-market event, not a company event

CoreWeave fell 12.10% on 18 August, and the instinct was to look for something CoreWeave had done. There is nothing to find.

No 8-K was filed between 11 and 26 August. No agency took action in the window. The 10-Q states that “the Company was in compliance with all covenants under its delayed draw term loan facility agreements.” And the stock gapped down at the open — $101.40 against a $106.00 prior close — which means the driver was overnight and global, not a disclosure.

What actually happened is that the 30-year Treasury yield reached its highest level in 19 years. The Treasury’s own daily constant-maturity series shows 5.31% on 17 August and 5.28% on 18 August; scanning every year back through 2007, the last higher print was 5.35% on 12 June 2007. For comparison, the 2023 maximum was 5.11% and 2024’s was 4.82%.

Gil Luria of D.A. Davidson put the transmission mechanism to CNBC that day:

“For them this is more existential, because they’re not getting very good returns and their cost of debt is very high… Therefore, every little change in interest rate could impact their ability and plans to build more data centers.”

— Gil Luria, D.A. Davidson, 18 August 2026

That is the whole CoreWeave thesis in two sentences. This is a company whose equity value is a residual claim on a spread — the gap between what GPUs earn and what money costs. When the long end moves, the residual moves violently.

Index inclusion made it worse, not better

CoreWeave joined the Nasdaq-100 before the open on 22 June 2026, one of five additions alongside Astera Labs, Nebius, Rocket Lab and Teradyne. It is worth noting for the record that this was the quarterly reconstitution, not the annual one — the annual reshuffle happens in December — because several write-ups have the label wrong.

Index inclusion is conventionally treated as a positive catalyst. It was not. CoreWeave fell 5.65% on the effective date and then closed lower for five consecutive sessions. The mechanism is well understood by anyone who has traded a reconstitution: index funds must buy before the effective date, which front-loads the demand into the announcement window, and everyone who bought the announcement sells into the passive bid on the day. What looks like a structural buyer is in practice a one-off liquidity event that discretionary holders exit into.

The same pattern hit Rocket Lab, added on the same day, which we examine in our Rocket Lab analysis. For CoreWeave the timing was particularly unhelpful: the sell-the-news came nine days before the Meta Compute report and five weeks before the credit-spread blowout, so the index bid was fully absorbed before the stock needed it.

The leverage, stated accurately

CoreWeave carries $35.1bn of total debt, $27.56bn of it non-current, against $5.02bn of equity. That is a debt-to-equity ratio of 6.98x.

A figure of “over 14x” has circulated, including in our own earlier coverage, and we should correct it here: 14.3x is total liabilities divided by equity ($72.05bn / $5.02bn), which is a different and less meaningful measure for a company of this shape. The debt ratio is 6.98x. That is high enough without overstating it.

The more revealing disclosure is the covenant relief. On 31 December 2025 CoreWeave amended its DDTL 3.0 facility: minimum liquidity was cut to $100m for March and April 2026, the first test of the debt-service coverage covenant was pushed out to 31 October 2027, and unlimited equity cures were permitted before 28 October 2026 — explicitly, in the filing’s framing, to align the facility with the timing of deliveries.

Read that plainly: the lenders moved the goalposts and allowed the company to plug covenant breaches with fresh equity, indefinitely, for ten months. That is not a company in distress, but it is not a company with covenant headroom either. It is a company whose lenders have chosen to be flexible while the build-out completes.

What the credit agencies actually did

Given how much of this story is credit, the ratings record is thinner than the drama implies. The only agency action inside the window was S&P on 9 April 2026, revising the outlook to positive and affirming the B+ issuer credit rating, with an upgrade to BB− contingent on remediating internal-control material weaknesses by the end of 2026. The prior baseline, from May 2025, was Moody’s Ba3 corporate family rating and B1 unsecured.

So: an outlook improvement in April, a critical sector note in July, and a CDS market that briefly priced a coin-flip default. Those three things cannot all be right, and the equity has been trading the most alarming of them.

The numbers: $192 bull, $113 base, $47 bear

Bull — $192 (+118%). Long rates fall, the credit spread normalises, and the $104.2bn backlog converts on schedule while operating margin recovers from the 1% trough. This level is slightly above the all-time high, which sounds aggressive until you note the stock covered a 73% rally in ten weeks earlier this year. In a business this levered, multiple expansion is violent in both directions.

Base — $113 (+28%). Rates stay roughly where they are, the build-out completes broadly on time, and CoreWeave grows into a still-elevated debt load without another guidance cut. Consensus sits around $142; we are deliberately below it, because the sell side raised price targets through both a CDS blowout and a 19-year high in long rates without cutting a single rating, and we think that is complacent.

Bear — $47 (−46%). Long rates grind higher, a customer commitment slips or a competitor like Meta Compute takes share, and the equity-cure window closes in October 2026 with covenants tested against a weaker book. A 46% fall sounds extreme; this stock has fallen more than 50% twice in the last fourteen months. The bear case here is not a tail scenario, it is realised behaviour.

What would change our mind

Three observable triggers, in order of information value. First, the CDS spread — it moved before the equity in July and it is the single best real-time read on this name; a sustained return toward 400bp would support the bull case more than any earnings beat. Second, the October 2026 expiry of the unlimited-equity-cure provision: how CoreWeave enters that date, and whether the facility is amended again, tells you what the lenders actually think. Third, customer concentration disclosures — Meta building competing capacity is the structural risk, and the next 10-Q’s customer breakdown is where it will show up first.

For the wider AI-infrastructure complex, see our current views on Nebius (NBIS), Nvidia’s Q2 margins, Nvidia’s AI server price increases and Anthropic’s reported $2 trillion IPO ambitions.

FAQ

Why did CoreWeave stock crash?

It crashed twice. From June to December 2025 it fell 64.8%, driven by a guidance cut caused by a late third-party data-centre developer and a JPMorgan downgrade. After rallying 73.4% between February and May 2026, it fell another 55.9% to a 29 July low on an operating-margin collapse, news that Meta was building competing cloud capacity, a Moody’s sector warning, and a credit-default-swap blowout to roughly 855 basis points.

Is CoreWeave in financial trouble?

Its credit is stressed but not distressed on the disclosed facts. Debt-to-equity is 6.98x on $35.1bn of total debt, and lenders amended the DDTL 3.0 facility in December 2025 to delay the first debt-service coverage test to October 2027 and permit unlimited equity cures. Against that, S&P revised the outlook to positive in April 2026 and the company reported compliance with all covenants in its most recent 10-Q.

What is CoreWeave’s price target?

Our 12-month view is $192 bull, $113 base and $47 bear against a spot of $88.01. Sell-side consensus is around $142. We sit deliberately below consensus in the base case because analysts raised targets through both a credit-spread blowout and a 19-year high in long-term rates without cutting ratings.

Why does CoreWeave stock fall when interest rates rise?

Because its equity is a residual claim on the spread between what GPU capacity earns and what debt costs. With $35.1bn of debt and a capex programme guided at $30–35bn, a move in long-term yields changes both the cost of the existing book and the viability of the next data centre. On 18 August 2026 the 30-year Treasury hit its highest level in 19 years and CoreWeave fell 12.1% having filed no news at all.

Who are CoreWeave’s biggest customers?

Customer concentration is the structural risk in this business, and the position shifted materially in 2026. The most consequential development was Bloomberg’s 1 July report that Meta — the source of CoreWeave’s largest new commitment — was building its own cloud business to sell surplus AI compute, turning a customer into a competitor. CoreWeave shares fell 13.9% that day. Verify the current breakdown in the latest 10-Q before relying on any figure.

This article is informational analysis only and is not investment advice, a recommendation, or a solicitation. Price targets are FinanceFeeds estimates and represent scenario analysis, not forecasts. Spot of $88.01 is the close of 26 August 2026 per stockanalysis.com; the stock traded at $92.73 pre-market on 27 August. CoreWeave is a highly leveraged, capital-intensive business whose equity has twice fallen more than 50% within the past fourteen months; you can lose money. Always do your own research.