CoreWeave Stock: Between a $104 Billion Backlog and a Debt Bill That Keeps Growing

Introduction

One week after CoreWeave told investors its contracted backlog had swelled to a record $104 billion, and the same day its CEO disclosed another $29.5 million stock sale, shares fell nearly 12% in a single session (Motley Fool, Blockonomi, Aug 18, 2026). That whiplash is the whole CoreWeave story compressed into a single trading week. The company priced its IPO at $40 a share on March 28, 2025, below its targeted range, and closed flat on debut (CNBC, Mar 28, 2025). By June 20, 2025, it had rocketed to a record close of $183.58 — up roughly 359% in under three months — on the strength of its ties to OpenAI, Microsoft, and Nvidia. By August 24, 2026, it traded at $83.93 — down about 54% from that peak, though still up roughly 110% from where it IPO'd (stockanalysis.com, Aug 24, 2026).

Neither side of that math is wrong. CoreWeave's revenue is genuinely compounding at a pace almost no public company can match, and its order book keeps growing faster than Wall Street can update its models. At the same time, the company is financing an unusually capital-intensive buildout with debt that's growing at least as fast as that backlog, at a moment when Treasury yields are climbing and the CEO has sold stock three separate times in six weeks. Sorting out which force — the hyper-growth story or the balance-sheet story — should set the stock's trajectory from here is the point of this piece.

Key Takeaways

• CRWV IPO'd at $40 on March 28, 2025, peaked at a record close of $183.58 on June 20, 2025, hit a 52-week low of $60.55 on July 29, 2026, and traded at $83.93 on August 24, 2026 — down about 54% from its peak but still up roughly 110% since its debut (CNBC; stockanalysis.com).
• CoreWeave's revenue backlog reached $104 billion as of June 30, 2026, up 246% year over year, with more than $25 billion in additional commitments — including a $21 billion Meta contract expansion and a $1 billion Jane Street investment — layered on afterward (CoreWeave Investor Relations, CNBC, Aug 11, 2026).
• Total debt stood at $51.61 billion as of August 24, 2026, against a debt-to-EBITDA ratio of 9.22x — leverage stockanalysis.com itself labels "significantly elevated" (stockanalysis.com statistics, Aug 24, 2026). Quarterly interest expense more than doubled year over year, from $267 million to $640 million, and is guided to reach $860–$940 million next quarter — more than triple where it stood five quarters earlier (qz.com, Aug 11, 2026).
• CEO Michael Intrator sold roughly $30.8 million of stock on July 7–8, another $29.5 million on August 18, and a smaller $1.2 million tranche described as a non-discretionary tax withholding sale on August 20 — all within a six-week span that also included two capex guidance increases (Motley Fool; SEC Form 4).
• Wall Street's consensus is "Buy" — 38 analysts polled by S&P Global average a $144.17 price target — but the $39-to-$317 range of individual targets is among the widest this series has covered, with HSBC's Reduce rating and DA Davidson's downgrade sitting alongside fresh price-target raises from Goldman Sachs and Truist in the same week as the CEO's stock sales (stockanalysis.com forecast, Aug 24, 2026).

Four Phases Since the IPO

CoreWeave's seventeen months as a public company break into four distinct stretches, each with a different dominant storyline.

March–June 2025: A Cautious Debut Becomes an AI Halo Trade

CoreWeave priced its IPO below its targeted $47–$55 range and closed flat at $40 on its first day of trading — read at the time as a warning sign for the broader "AI infrastructure" trade (CNBC, Mar 28, 2025). That skepticism didn't last. Shares were up 42% by the third trading day, closing at $52.57 (CNBC, Apr 1, 2025), and the rally kept building as the market absorbed how deeply CoreWeave was wired into OpenAI, Microsoft, and Nvidia's AI buildout. Shares closed at a record $183.58 on June 20, 2025 — up roughly 359% from the IPO price in under three months, the kind of move that priced in years of flawless execution almost immediately.

July 2025–January 2026: The Bear Case Arrives Early, and So Does More Capital

The skepticism that missed the IPO debut showed up a few weeks later instead. HSBC initiated coverage with a Reduce rating and a $32 price target — implying 77% downside — on July 17, 2025, arguing CoreWeave was exposed to GPU commoditization risk since it rents Nvidia chips rather than owning proprietary technology; shares fell about 7% on the call (TipRanks, Jul 17, 2025). Through the rest of 2025, CoreWeave kept adding to both sides of its ledger: OpenAI expanded its contracted commitment to a cumulative $22.4 billion across three tranches by September (CoreWeave Investor Relations, Sept 25, 2025), while the company priced an upsized $2.25 billion offering of 1.75% convertible senior notes due 2031 that December — an early signal of how much of the buildout would be funded with debt (CoreWeave Investor Relations, Dec 2025).

That tension crystallized in a single week in January 2026. On January 27, Nvidia disclosed a $2 billion strategic equity investment in CoreWeave tied to a 5-gigawatt AI-factory buildout target through 2030, and Deutsche Bank upgraded the stock to Buy, lifting its target to $140 from $100 (CNBC, Jan 27, 2026). One day later, HSBC — unmoved by the Nvidia infusion — cut its own target further, to $41 from $44, warning of a projected $8.9 billion liquidity shortfall in 2026 and estimating that fresh CoreWeave debt could carry rates above 10% given how wide the company's credit spreads had widened (Investing.com; Insider Monkey, Jan 28, 2026). Two firms, the same catalyst, opposite conclusions — a split that has never really closed.

February–May 2026: Bigger Numbers on Both Sides

CoreWeave's fiscal 2025 results, reported February 26, 2026, showed revenue of $5.13 billion — up from $1.92 billion in 2024 — alongside a GAAP net loss of $1.17 billion and a backlog of $66.8 billion (CoreWeave Investor Relations, Feb 26, 2026). Then, on May 18, 2026, DA Davidson analyst Gil Luria downgraded the stock to Neutral from Buy and slashed his target to $100 from $175, arguing CoreWeave was "destroying value with every data center it builds" because unlevered returns on new capacity — roughly 4% by his estimate — were running well below the company's rising cost of capital, near 9% (Yahoo Finance, May 18, 2026). It was the first time a major analyst framed CoreWeave's problem not as demand risk, but as a math problem in the balance sheet itself.

July–August 2026: A Record Backlog, a Competing Customer, and a Selling CEO

The most eventful stretch came last. Bloomberg reported on July 1, 2026 that Meta was developing "Meta Compute," a commercial cloud service to resell surplus AI capacity — a direct threat given Meta is also one of CoreWeave's largest customers; shares fell 12.7% on the report (Motley Fool, Jul 1, 2026). Days later, on July 7–8, CEO Michael Intrator sold roughly 369,489 shares for about $30.8 million (Yahoo Finance/Motley Fool, Jul 7–8, 2026). Meta formally announced Meta Compute on July 17, and the stock touched a new 52-week low of $60.55 on July 29 amid what several outlets described as a broader "neocloud" selloff hitting GPU-cloud stocks as a group (Benzinga, Jul 29, 2026).

Then, on August 11, 2026, CoreWeave reported second-quarter results that scrambled the narrative again: revenue of $2.58 billion, up 112% year over year and above consensus, alongside a backlog that had grown to $104 billion, up 246% year over year — including a new $21 billion contract expansion from Meta itself (on top of a prior $14 billion commitment) and a $1 billion strategic investment from Jane Street (CNBC, Aug 11, 2026). Meta, in other words, was simultaneously the competitive threat that had crushed the stock in July and the single largest new contract behind the backlog beat three weeks later. Shares jumped as much as 14% in extended trading on the print, but the same release raised full-year 2026 capital-expenditure guidance to $35–$39 billion from a prior $31–$35 billion range.

That guidance raise caught up with the stock a week later. By August 18, shares had reversed hard, falling 11.8% to 12.1% in a single session — down roughly 15.5% for the week — as investors refocused on the spending number alongside climbing Treasury yields. The same day, Intrator sold another 307,692 shares for about $29.5 million under a 10b5-1 trading plan established in November 2025 (Motley Fool; Blockonomi, Aug 18, 2026). Two days later, a much smaller Form 4 disclosed Intrator selling 13,129 shares for about $1.2 million — described as a non-discretionary sale to cover tax withholding on vesting equity, distinct in character from the two larger transactions earlier in the month (SEC Form 4; Motley Fool, Aug 20, 2026). Even so, by August 21 and 24, Goldman Sachs and Truist had both raised their price targets — to $139 and $165, respectively — citing the backlog build and margin tailwinds despite the selloff.

The Bull Case: A Backlog Growing Faster Than the Company Can Build

CoreWeave's growth numbers are not in dispute. Second-quarter 2026 revenue of $2.58 billion was up 112% year over year, and trailing-twelve-month revenue reached $7.59 billion, up 115.3% (stockanalysis.com, Aug 24, 2026). The company's revenue backlog — CoreWeave's term for remaining performance obligations plus estimated future revenue under committed contracts subject to its delivery capacity — hit $104 billion as of June 30, 2026, up 246% from a year earlier (CoreWeave Investor Relations, Aug 11, 2026). Management said more than $25 billion in additional commitments arrived in the weeks immediately following quarter-end — not even reflected in that $104 billion figure — including the Meta expansion and Jane Street's $1 billion investment (CNBC, Aug 11, 2026). Layer in OpenAI's cumulative $22.4 billion in contracts and Nvidia's own $2 billion equity bet on CoreWeave's 5-gigawatt buildout plan, and the demand side of the ledger reads like a company that could plausibly grow into any valuation, given enough time.

That demand isn't happening in a vacuum. Combined 2026 AI capital spending across Microsoft, Alphabet, Meta, and Amazon is tracking toward roughly $725 billion, up from about $410 billion in 2025 (Tom's Hardware; CNBC, Feb 2026) — a rising tide of hyperscaler spending that keeps generating overflow demand for the kind of specialized GPU capacity CoreWeave exists to sell, a dynamic this series also examined from Nvidia's side of the AI-infrastructure trade and from Meta's own capex-driven stock hardship.

The Bear Case: A Balance Sheet Growing Just as Fast

The other side of the ledger is just as real. CoreWeave's total debt stood at $51.61 billion as of August 24, 2026, against trailing EBITDA that produces a debt-to-EBITDA ratio of 9.22x — leverage stockanalysis.com's own data page flags as "significantly elevated" (stockanalysis.com statistics, Aug 24, 2026). A narrower reading of the company's own second-quarter 10-Q — which counts principal balances on recourse and non-recourse financing facilities, senior notes, and convertible notes, rather than every financing-related obligation — puts total debt at $35.55 billion as of June 30, 2026 (The Register, citing CoreWeave's SEC 10-Q filing). The gap between the two figures reflects how differently data providers treat a company that finances most of its data centers through complex, project-level facilities; either measure tells the same underlying story of leverage rising faster than trailing profitability. CoreWeave's lenders have already responded to that trajectory once: in a January 2026 amendment, the company pushed out the initial testing dates on two of its own debt-service and contract-realization covenants and temporarily lowered its minimum-liquidity requirement, a sign its financing partners were already managing around a tighter cushion than the headline growth numbers would suggest (Investing.com, Jan 2, 2026).

The spending side of that leverage keeps climbing too. Full-year 2025 capital expenditures totaled $14.9 billion. CoreWeave guided 2026 capex to $31–$35 billion in the spring, then raised that range again to $35–$39 billion alongside its second-quarter results — more than double the prior year's actual spend inside a single fiscal year (Yahoo Finance, 2026 guidance coverage).

That capex is consuming cash faster than the business generates it: free cash flow ran negative $5.74 billion in the second quarter alone, as $6.42 billion of capital spending swallowed $679 million of operating cash flow (247wallst.com, Aug 11, 2026), and the resulting borrowing shows up directly in the income statement. Quarterly interest expense climbed from $267 million in the second quarter of 2025 to $536 million in the first quarter of 2026 to $640 million in the second quarter of 2026, and CoreWeave has guided to $860–$940 million for the third quarter — more than tripling from that Q2 2025 starting point to the midpoint of the Q3 2026 guidance (qz.com, Aug 11, 2026).

Credit-rating agencies have priced that risk formally: Moody's affirmed a Ba3 rating with a stable outlook, and S&P assigned a B+ issuer rating while revising its outlook to positive from stable — both solidly sub-investment-grade readings that mean fresh CoreWeave debt gets priced at junk rates (S&P Global Ratings; cbonds.com, Apr 9, 2026), just as long-term Treasury yields climb toward multi-year highs.

Against that backdrop, the CEO's own trading pattern draws extra scrutiny, even accounting for the fact that two of the three 2026 sales were pre-scheduled under 10b5-1 plans rather than discretionary, opportunistic trades:

Date	Shares Sold	Approximate Value	Nature
Jul 7–8, 2026	~369,489	~$30.8 million	Reported stock sale
Aug 18, 2026	307,692	~$29.5 million	10b5-1 plan (est. Nov 2025)
Aug 20, 2026	13,129	~$1.2 million	Non-discretionary tax withholding

Source: SEC Form 4; Motley Fool, Yahoo Finance, Blockonomi, Jul–Aug 2026.

Meta: Customer, Competitor, and the Whole Contradiction in One Company

No single relationship captures CoreWeave's tug-of-war better than the one with Meta. In July 2026, reporting that Meta was building its own resale cloud service to compete directly with CoreWeave's core business knocked nearly 13% off the stock in a day. Three weeks later, Meta itself showed up as the company's largest incremental customer, expanding its commitment by $21 billion to a cumulative $35 billion — a deal CoreWeave management highlighted as central to the record backlog. Both facts are true simultaneously, and neither cancels the other out: Meta clearly believes it needs more external GPU capacity than it can build itself in the near term, which is bullish for CoreWeave's order book, while also building toward a future where it needs less of that capacity, which is exactly the long-run risk HSBC flagged back in July 2025 when it first raised the GPU-commoditization thesis. The same dynamic — a hyperscaler customer whose own AI capex decisions can swing a supplier's stock double digits in either direction — runs through this series' coverage of Meta's own capex-driven hardship from the other side of the relationship.

What the Multiple Says When There's No P/E to Read

CoreWeave isn't profitable, which means the P/E-based framework SPXScore builds its coverage around has no trailing or forward earnings multiple to work with here — trailing-twelve-month net income sits at negative $1.93 billion, or an EPS loss of $3.70 (stockanalysis.com, Aug 24, 2026). For a plain-English refresher on what P/E and its forward-looking variants actually measure — and why they go quiet for a company like this one — see SPXScore's explainer on P/E, PEG, forward P/E, and forward PEG.

In place of earnings multiples, the more useful numbers are revenue- and EBITDA-based. CoreWeave's trailing price-to-sales ratio sits at 6.38x, versus a forward price-to-sales ratio of just 2.48x (stockanalysis.com statistics, Aug 24, 2026) — a wide gap that is, in effect, the bull case expressed as a single number: the market is pricing in roughly two and a half times more revenue over the next twelve months than it collected over the trailing twelve, consistent with the backlog converting to cash on something close to schedule. Enterprise-value multiples tell a pricier story: EV/Revenue of 12.45x and EV/EBITDA of 24.94x are still rich readings for a company burning billions of dollars of free cash flow every quarter (same source, same date). It's the kind of gap SPXScore's trend tracking is built to surface across the names this series covers — a stock that looks expensive on trailing results and considerably less so on what the market is underwriting for next year.

Wall Street's Buy, Hold, Sell Verdict

The headline consensus leans bullish. S&P Global's polled panel of 38 analysts rates CoreWeave a "Buy" with an average price target of $144.17 — implying roughly 72% upside from the $83.93 closing price — though the individual targets range from a low of $39 to a high of $317, a nearly eightfold spread that is among the widest this series has documented (stockanalysis.com forecast, Aug 24, 2026). MarketBeat's separately tracked panel of 32 analysts reads a bit more cautious — a "Moderate Buy" made up of 2 Strong Buy, 16 Buy, 11 Hold, and 3 Sell ratings, with a lower average target of $127.64 (MarketBeat, Aug 2026). The gap between the two consensus reads is itself informative: which panel you poll changes the picture more than usual for a stock this contested.

Firm	Date	Action	Price Target
Deutsche Bank	Jan 27, 2026	Upgraded to Buy from Hold	$140 (from $100)
HSBC	Jan 28, 2026	Maintained Reduce	$41 (from $44)
DA Davidson	May 18, 2026	Downgraded to Neutral from Buy	$100 (from $175)
Goldman Sachs	Aug 21, 2026	Reiterated	$139 (from $121)
Truist	Aug 24, 2026	Reiterated Buy	$165 (from $155)
CRWV stock price	Aug 24, 2026	—	$83.93

Sources: CNBC, Insider Monkey, Yahoo Finance, Investing.com, Jan–Aug 2026.

Read together, the pattern isn't indecision so much as genuine, unresolved disagreement: HSBC's Reduce rating and DA Davidson's downgrade both rest on the same balance-sheet math this piece has walked through, while Deutsche Bank, Goldman Sachs, and Truist are all underwriting the backlog conversion story instead. Both camps are looking at the same company; they're weighting the two halves of its story differently.

The Macro Backdrop: Why the Debt Story Keeps Getting Louder

CoreWeave's financing costs aren't rising in isolation. The Federal Reserve held its benchmark rate at 3.50%–3.75% for a fifth consecutive meeting on July 29, 2026, in a 9–3 vote that saw three regional bank presidents — Cleveland's Beth Hammack, Minneapolis's Neel Kashkari, and Dallas's Lorie Logan — dissent in favor of a hike, the most dissents in that direction since 2016 (CNBC; Bloomberg, Jul 29, 2026). The next FOMC decision is due September 16, 2026. Meanwhile, the 10-year Treasury yield climbed to roughly 4.71%–4.74% in mid-to-late August 2026 — its highest level since early 2025 — and two-thirds of respondents in a Bloomberg Markets Pulse survey expect it to top 5% before year-end (Bloomberg, Aug 19, 2026). That backdrop matters unevenly across the market — established, cash-generative hyperscalers can fund their own AI capex from operating cash flow, while a company like CoreWeave, financing $35–$39 billion of 2026 spending substantially with debt, is directly exposed to where long-term yields sit. HSBC's own modeling that fresh CoreWeave debt could carry rates above 10% given the company's widened credit spreads is a direct expression of that exposure (Insider Monkey, Jan 28, 2026). For more on what's driving the move in long-term rates and what it means for capital-intensive stocks broadly, see SPXScore's coverage of the Treasury's own effort to tame yields.

None of that changes the demand side of the equation, which is exactly why the tug-of-war persists rather than resolving cleanly in either direction: hyperscaler AI capex keeps climbing toward that roughly $725 billion combined 2026 figure at the same time CoreWeave's own cost of capital keeps climbing with it.

Buy, Hold, or Sell?

Neither half of CoreWeave's story cancels the other out, which is precisely why the stock's own analyst-target dispersion — an eightfold range from $39 to $317 — is the most honest answer available right now. The bull case is straightforward: a $104-billion-plus backlog, expanding hyperscaler relationships, and a chip supplier willing to co-invest in the buildout, all set against AI infrastructure demand that shows no sign of slowing. The bear case is just as straightforward: leverage north of 9x by the broadest measure, interest expense on pace to more than triple in a little over a year, free cash flow deeply negative, sub-investment-grade credit ratings, and a CEO who has sold stock three times in six weeks even as he calls the backlog historic. A reasonable investor doesn't need to resolve that tension to have a view — they need only decide whether they believe CoreWeave's revenue converts to cash faster than its debt compounds, and how much confidence they're willing to place in a management team that is, by its own trading activity, hedging that same bet.

Frequently Asked Questions

What is CoreWeave's revenue backlog, and why does it matter?

CoreWeave's revenue backlog — remaining performance obligations plus estimated future revenue under committed contracts — reached $104 billion as of June 30, 2026, up 246% year over year (CoreWeave Investor Relations, Aug 11, 2026). It matters because it represents contracted future revenue the company hasn't yet delivered or recognized — the clearest evidence for the bull case, provided CoreWeave can build capacity fast enough to convert it into actual cash flow.

Is CoreWeave profitable?

No. CoreWeave posted a GAAP net loss of $626 million in the second quarter of 2026 and negative trailing-twelve-month net income of $1.93 billion, though its Adjusted EBITDA margin was a positive 59% for the quarter (CNBC; stockanalysis.com, Aug 2026).

What is CoreWeave's debt-to-EBITDA ratio?

Stockanalysis.com puts it at 9.22x as of August 24, 2026, on total debt of $51.61 billion — a figure the site itself describes as significantly elevated leverage (stockanalysis.com statistics, Aug 24, 2026). A narrower reading based on principal debt disclosed in CoreWeave's Q2 2026 10-Q, $35.55 billion, would produce a lower ratio, but leverage has risen sharply under either measure.

When does CoreWeave report next?

CoreWeave's third-quarter 2026 results are estimated to arrive in early-to-mid November 2026, based on its past reporting cadence; the company had not confirmed a specific date at the time of writing.

Conclusion

CoreWeave in August 2026 is a genuinely split story rather than a settled one: a $104 billion backlog and a roster of hyperscaler and chipmaker partners betting directly on its buildout, sitting on top of debt north of $50 billion, interest expense on pace to more than triple in a little over a year, and a CEO who has sold stock three separate times in six weeks. The stock's own 54% drawdown from its 2025 peak — alongside a still-positive 110% gain since its IPO — reflects a market that hasn't decided which force wins, and the widest spread of analyst price targets this series has documented suggests Wall Street hasn't either. The next real test isn't a single headline; it's whether the $25 billion-plus of fresh commitments booked since quarter-end converts into delivered, cash-generating capacity before the next refinancing comes due, and whether Treasury yields cooperate while it does. CoreWeave's third-quarter report, expected in early-to-mid November 2026, is the next checkpoint on that question.

---

This article discusses CoreWeave's historical stock performance and the financial, competitive, and macroeconomic conditions surrounding it as of August 24, 2026, for informational purposes only. It is not investment advice or a recommendation to buy, hold, or sell any security. CoreWeave's third-quarter 2026 results, estimated for early-to-mid November 2026, had not been reported at the time of writing, and figures such as total debt, backlog, and leverage ratios are subject to revision as the company reports new results and as data providers update their calculations.