Revenue, last 12 months$7.6BQ2 2026: $2.58B, up 112%
Revenue backlog$103.7B30 Jun 2026, plus $25B+ signed in early Q3
Net loss, Q2 2026-$0.63Binterest of $0.64B a quarter
Largest customer, Q2 202636%of revenue; 71% a year earlier
Net debt$29.5B30 Jun 2026
Value over time
Total market value
$20B$40B$60B$80B
Nov 25Feb 26May 26Aug 26
$30B -55% since Oct 2025
$20B$40B$60B$80B$100B
Apr 25Jul 25Oct 25Jan 26Apr 26
$30B +71% since 2025
Total value, not price per share: share counts change with fundraising, stock pay and splits.
What today's market value implies
If CoreWeave is worth $30B today, our model says this has to happen by 2036:
Revenue in 2036$97B9.5x today's $10B, growing 25% a year
Net profit in 2036$14B0.1x what Apple earned in 2025 (about $110B)
Operating margin in 203618%our middle-case margins for each business line
Value today / revenue2.9xthe S&P 500 average is about 3x
Growth vs our middle case+4 ptsa year, on every business line
Assumption Solved from our model: middle-case margins, 18x earnings in 2036, a 12.0% required return and 3% new shares a year. The lab below opens on this "market's view"; change any number there.
Assumption Starting cases are Stonkometer values, not forecasts. The needle reads your inputs. Not sure what to assume? Answer 5 quick questions.
The market price needs about +4 points more growth a year than our middle case, on every business line.
Revenue by business line
Scenarios vs today
Share your valuation
Your link opens your exact model: every slider, scenario and moonshot.
Download
Defaults and conventions
Fact Revenue, cash, debt and share counts come from the filings or reports listed under Sources (data as of 2026-10-30).
Fact Currency: nominal US dollars. All four companies report in US$, so there is no FX conversion.
Assumption Inflation is not a separate input: growth rates and the risk-free rate are nominal, so they already include it.
Assumption Risk-free rate 5% (the 10-year US Treasury yield was 4.99% on 26 Oct 2026); risk premium 7%; required return 12% a year.
Assumption Tax 21% on profits (US federal rate); dilution 3% a year; 10-year horizon; magic revenue earns 30%.
Assumption Today's operating margins per business line (AI cloud -2%) are Stonkometer estimates from the latest filings.
Assumption Bear, middle and bull values and bet sizes, margins and chances.
Change any of them in Advanced mode. Full list on the method page.
Affiliate links: we may earn a commission if you sign up or buy, at no extra cost to you. It never changes our numbers. Affiliate disclosure
Advertisement
Build your assumptions: 5 questions
Not sure what to put in the sliders? Answer in plain terms: prices, customers, and how much the big risks worry you. Each question starts with a Fact (or Our view for rough benchmarks); your answer is your Assumption. The answer choices are ours, and skipped questions use the middle one.
0 of 5 answered
Moonshot explorer Our view
Markets that don't exist yet but could one day pay CoreWeave. Each one is valued like an option: the chance it happens, times the profit if it does, discounted back to today. The market sizes, shares, chances and dates are our guesses: move them, and switch on the ones you believe in to add them to your model.
The default cloud for AI inference
Most AI answers, agents and video generation run on rented GPUs, and CoreWeave becomes a default place to run them.
Fact CoreWeave set MLPerf inference records in 2026 and launched tools for agents and reinforcement learning.
What would have to be true:
AI use keeps growing fast
Big clouds do not undercut it on price
If it happens: CoreWeave earns $20B a year from 2031, worth $95B then at 20x profit. Weighted by a 20% chance and discounted to today: $9.3B.
Owning the power and the buildings
Owning more data centres and grid power instead of leasing them, and renting space to others.
Fact Contracted power grew to about 3.7 GW in mid-2026, with 1.5 GW active.
What would have to be true:
Cheap enough financing
Power connections arrive on time
If it happens: CoreWeave earns $7.5B a year from 2030, worth $47B then at 20x profit. Weighted by a 25% chance and discounted to today: $6.7B.
Switch on the moonshots you believe in to add their value to your model above.
Value today = chance x (market x share) x margin x (1 - tax) x 20x profit, discounted from the start year at your required return and shrunk for dilution. Margins are our view: The default cloud for AI inference 30%, Owning the power and the buildings 40%.
The largest "neocloud": rents out NVIDIA GPU clusters to AI labs, Microsoft, Meta and other big tech, mostly on multi-year contracts. Below is a valuation lab built on its latest SEC filings. It opens on what today's price implies; pick another scenario or move the sliders, and the Stonkometer compares today's price with your numbers. Assumption Our middle case puts CoreWeave at $20B against a market value of $30B. It is a starting point, not a forecast; yours will differ, and that is the point.
Fact from filings or reports, sourced below · Assumption a Stonkometer default you can change · Our view opinion
Big claims, checked
Fact Each claim is quoted with its source. The maths shows what would have to be true; the slider defaults are assumptions.
CoreWeave Q2 2026 results (Aug 2026)
“Revenue backlog of about $104 billion, plus more than $25 billion of new commitments in early Q3.”
About ten years of revenue at the latest quarter's pace; customers pay as capacity is delivered.
CoreWeave Q2 2026 results (Aug 2026)
“Revenue up 112% to $2.58 billion in the quarter.”
From $1.21B in Q2 2025.
CoreWeave Q2 2026 results (Aug 2026)
“About 3.7 GW of contracted power, 1.5 GW of it active.”
Revenue runs at about $6.9B a year per active GW today ($10.3B run-rate / 1.5 GW).
Who owns it, who controls it
Fact Class A: one vote per share, sold to the public. Class B: ten votes per share, held by the co-founders.
Economic interest
20.4%57.2%
Co-founders (Intrator, Venturo, McBee) 20.4%
Magnetar 13.9%
NVIDIA 8.5%
Other shareholders 57.2%
Shareholder votes
73.6%17.9%
Co-founders (Intrator, Venturo, McBee) 73.6%
Magnetar 5.3%
NVIDIA 3.2%
Other shareholders 17.9%
Who picks the board
all
Elected by all votes, where the co-founders have about 74% all
Michael Intrator (CEO) holds about 39% of the votes, Brian Venturo about 20% and Brannin McBee about 15%.
Magnetar, the hedge fund that backed CoreWeave early, and NVIDIA are the largest outside holders; NVIDIA's stake includes the $2B it invested in January 2026.
Economic stakes are rounded from the proxy statement's Class A and Class B holdings; new shares and converts since then dilute them.
What CoreWeave owes, what it holds, and the spending it has already signed up for. Contracts for compute and infrastructure are not debt on the balance sheet, but they must be paid like it.
Cash and investments$5.5B30 Jun 2026
Debt$35.1Bincluding finance leases
Net debt$29.6Bcash minus debt
Spending commitments$35.5Bdata-centre and equipment leases signed but not yet started; about $16B of leases already running come on top
Commitments / revenue3.4xrevenue run-rate (Q2 2026 x 4): $10.3B
The debt is a mix of GPU-backed loans (DDTL facilities), unsecured notes, a Eurobond and convertible notes.
Interest was $640M in Q2 2026 alone, while the business made an operating loss of $49M.
$4.4B of principal falls due in the rest of 2026 and $6.2B in 2027. In September it added $3.7B of 2.875% convertible notes due 2033.
Much of the debt sits in subsidiaries and is repaid from specific customer contracts; some of it does not have recourse to CoreWeave itself.
Revenue was $2.58B in the quarter to 30 June 2026, up 112%. The operating loss was $49M and the net loss $626M, after $640M of interest.
Revenue backlog was about $104B, not counting more than $25B of new commitments signed early in the third quarter.
Active power rose by nearly 500 MW to 1.5 GW; contracted power is about 3.7 GW.
Three customers each brought 10% or more of revenue in the quarter: 36%, 26% and 10%. A year earlier the largest alone was 71%. Microsoft and OpenAI are named as significant customers; Meta committed up to about $21B and Jane Street about $6B.
It spent $14.1B on property and equipment in the first half of 2026 and had $35.1B of debt. Data-centre leases signed but not yet started add $35.5B.
In September 2026 it sold $3.7B of 2.875% convertible notes due 2033, convertible at $97.85 a share. Michael Intrator, Brian Venturo and Brannin McBee hold ten-vote Class B shares.
Key risks Fact
Debt of $35.1B and about $2.6B of yearly interest: $4.4B of principal is due in the rest of 2026 and $6.2B in 2027.
A few customers bring most of the revenue; if one cuts back or cannot pay, there is no easy replacement.
GPUs lose value fast as NVIDIA ships new chips, so contracts must pay them back within a few years.
Big clouds (Microsoft, Amazon, Google, Oracle) are customers and competitors at the same time.
Power and data-centre delays push revenue out while interest keeps running.
The company's own risk factors, summarised in plain English. Read the originals below.
The business lines in the model
AI cloud (GPU compute rented by the hour or by contract)$7.6B last 12 months (to 30 jun 2026), $10B on the latest basis
$2.58B in Q2 2026, up 112%. The GAAP operating margin was -2%, before $640M of interest in the quarter; before depreciation (adjusted EBITDA) it was 59%. Margin in 2036 is our estimate.
Bet: AI software and inference platform
Weights & Biases (bought in 2025) and tools for training, inference and agents, sold on top of the GPUs.
Bet: Government and sovereign AI clouds
National AI data centres in the US, UK and Europe for governments and regulated industries.
Last 12 months (to 30 Jun 2026): H2 2025 from the 2025 10-K plus H1 2026 from the Q2 2026 10-Q. Latest quarter x 4 (Q2 2026): Quarter to 30 Jun 2026 x 4. Backlog of $103.7B is more than 10 years of revenue at that pace. Cash $5.5B, less $35.1B of debt (30 Jun 2026); $1.4B of restricted cash and about $16B of operating leases not included. 458.9M Class A + 92.7M Class B shares on 31 Jul 2026 (10-Q cover). Convertible notes could add more..
Questions people ask Fact
Is CoreWeave stock overvalued?
At $54.87 a share CoreWeave is worth about $30B, with about $29.5B of net debt on top. It has no profit yet, so the lab values the AI cloud on the margins you expect it to earn once the data centres are built.
What is a neocloud?
A cloud company that mainly rents out GPUs for AI, rather than the full range of services Amazon, Microsoft or Google sell. CoreWeave, Nebius and IREN are the best-known listed ones.
Who are CoreWeave's biggest customers?
Three customers each brought 10% or more of revenue in Q2 2026: 36%, 26% and 10%. The 10-Q names Microsoft and OpenAI as significant customers; Meta has committed up to about $21B and Jane Street about $6B.
Not investment advice. Stonkometer is an educational calculator, not investment advice. Valuations depend entirely on the assumptions you choose. We are not a broker or adviser, and nothing here is a recommendation to buy or sell any security. Do your own research and consider a licensed adviser. Full disclaimer