
When I took my first CFO role, I was shocked by how fragmented our software spend data was. The silos made total spend management impossible.
That's why Tropic's approach is so refreshing. Their 2025 Software Buying Trends report, gives a glimpse of what they do best (and why I use them).
It doesn't just show you that average software spend rose 10% this year or which categories cost most per employee. It gives you the framework to unify your spend data for visibility across your entire software ecosystem.
Grab their free report to benchmark your software spend, prepare for emerging AI pricing models, and build a strategic approach to software management.
CoreWeave S-1 Breakdown
CoreWeave, the AI Hyperscaler, has filed their S1 to go public. Here’s the breakdown.
(If you’d like to listen to this in audio format you can find it on Apple, Spotify, and Youtube.)
What is CoreWeave? (Business Overview)
CoreWeave is basically a supercharged cloud provider built specifically for AI.
Instead of offering general-purpose cloud computing like Amazon Web Services (AWS), it rents out super-powerful GPUs (the high-performance chips AI models need to run) to companies like Meta, Microsoft, and AI startups.
Think of it like this: If AWS is a giant all-you-can-eat buffet for cloud computing, CoreWeave is a high-end sushi bar that only serves AI workloads.
“For AI to reach its full potential, it needs a purpose-built AI cloud platform with infrastructure and managed cloud services that are delivered in an efficient, automated, and highly performant way. Enter CoreWeave, the AI HyperscalerTM.”
Key Metrics

Revenue Growth:
2022: $16M
2023: $229M (+1,346% YoY)
2024: $1.9B (+737% YoY) (not a typo, lol)
Profitability: Still deep in the red
2022 Net Loss: ($31M), 194% of revenue
2023 Net Loss: ($594M), 259% of revenue
2024 Net Loss: ($863M), 45% of revenue
Losses are still really large, but declining as a percentage of revenue
Revenue Model: 96% of revenue comes from long-term contracts (2–5 years), providing high revenue visibility.
“We generate revenue by selling access to our AI infrastructure and proprietary managed software and application services through our CoreWeave Cloud Platform. Access to our platform, including compute, networking, managed software services, and application software services, is currently priced on a per GPU per hour basis. Storage is sold separately on a per gigabyte per month basis.”
Remaining Performance Obligations (Contracted Backlog): $15.1B (+53% YoY)
YO!!! That’s a LOT of RPO (as the pro’s call it)
RPO is cool because it gives companies a way to show off the multi-year contracts they've landed with customers, and provide additional confidence on the company’s ability to deliver on revenue expectations.
Instead of just looking at what's in the bank right now, RPO gives you the full picture of what's in the pipeline.
It's like seeing the whole iceberg, not just the tip.
“We purpose-built our CoreWeave Cloud Platform to be the infrastructure and application platform for AI.”
7 Key Themes
1. The AI Compute Land Grab
There’s a global shortage of high-end GPUs, and CoreWeave is one of the few players with access to them. As AI models grow larger and demand more compute, CoreWeave is cashing in by renting out the most valuable resource in AI right now: NVIDIA GPUs.
It’s kinda like chip arbitrage.
CoreWeave buys GPUs from NVIDIA and rents them out at a markup—essentially playing middleman in the AI compute economy. But as more companies try to secure their own GPUs (or build custom chips like OpenAI's rumored “Stargate”), does this model have long-term staying power?
CoreWeave’s entire business depends on getting the newest and best GPUs from NVIDIA. If NVIDIA decides to prioritize other customers or sell directly to AI companies, CoreWeave is in trouble.
2. “Take-or-Pay” Contracts = Revenue Visibility
96% of CoreWeave’s revenue comes from multi-year, locked-in contracts, meaning customers pay whether they use the compute or not. This isn’t a typical cloud business with unpredictable on-demand usage—it’s more like a gym membership for AI companies… that lasts more than one year… and requires large prepayments…
At the same time, we’re in a period where AI startups are raising billions and spending freely on compute. If AI funding cools off, will CoreWeave’s backlog hold up? Or is there a risk that some of its long-term contracts don’t actually get paid as companies go belly up?
“Committed contracts generally have a fixed price for their duration, which is measured on a dollar per contracted GPU per hour basis, and are billed monthly based on the customer’s reserved usage commitments. Committed contracts generally have a predetermined term and start either on a fixed date or when we deliver the capacity specified in the contract. As of December 31, 2024, our committed contracts had a weighted-average contract duration of approximately four years.”
3. Cloud Giants Are Watching
AWS, Google Cloud, and Azure could easily decide to go harder into AI compute and squeeze CoreWeave out. But right now, they’re spread too thin across multiple industries, while CoreWeave is 100% focused on AI.
The big question: Can CoreWeave stay ahead?
AWS, Azure, and Google Cloud can all buy NVIDIA chips too. Money be green. And they have even more of it.
CoreWeave argues that its networking, storage, and Kubernetes-based orchestration make its offering more than just a GPU rental shop. Will customers see the difference, or will they just chase the lowest cost?
Generalized clouds operated by hyperscalers were not built to serve the specific requirements of AI. These clouds were created over a decade ago and were designed for general purpose use cases such as search, e-commerce, generalized web-hosting, and databases, and relied on CPU-based web-scale compute, and thus are not optimal for the high compute intensity requirements of AI.
4. The Private Funding Boom Leading to IPO
CoreWeave has raised billions from firms like Magnetar, Blackstone, and Coatue—with NVIDIA itself investing $100M+. This IPO is as much about giving early investors an exit as it is about raising capital for expansion.
Not to put on my tinfoil hat or anything, but what is this a bridge to?
Going public helps CoreWeave raise capital, but it also puts pressure on them to prove profitability. Do they scale to profitability, or is the real goal to give existing investors liquidity and then position themselves for a merger with Microsoft, Google, or Oracle (all of whom need more AI compute capacity)?
5. Expansion Beyond AI
Right now, AI training and inference dominate demand, but can CoreWeave expand into other compute-heavy markets like high-frequency trading, biotech simulations, or video rendering? Or will it remain tied to the boom-and-bust cycle of AI funding.
“Our CoreWeave Cloud Platform is an integrated solution that is purpose-built for running AI workloads such as model training and inference at maximum performance and efficiency.”
Their sector expertise is a double edge sword - it makes them the best at what they do, by remaining hyper specific, while also exposing them to AI specific shocks.
6. Microsoft’s CoreWeave Play: A Call Option on GPU Demand
In 2024, Microsoft accounted for 62% of CoreWeave's revenue, highlighting a heavy reliance on a single customer.
Microsoft’s deal with CoreWeave is the ultimate low-risk, high-upside move—they’re essentially leasing AI compute capacity instead of building it outright, keeping their options open as the AI arms race unfolds.
Minimal Commitment, Maximum Flexibility – By outsourcing AI workloads to CoreWeave, Microsoft secures compute without the CapEx burden of building its own infrastructure upfront.
Hedging Against Supply Chain Risk – If GPU shortages persist, Microsoft retains access to high-end AI compute without having to overcommit on hardware purchases today.
Easy to Walk Away – If AI compute pricing normalizes or Microsoft scales its own internal capacity, they can dial down their CoreWeave usage with minimal downside. Or they could resell it.
For Microsoft, this is a call option on GPU demand—cheap to hold, but with huge upside if AI workloads keep exploding. For CoreWeave, though, it means their biggest customer can leave the moment it’s more cost-effective to go in-house.
7. WeWork Vibes?
CoreWeave has $15 billion in long-term lease commitments tied to data centers, power contracts, and equipment rentals. These obligations don’t appear as traditional debt on the balance sheet, but they function as long-term liabilities.
Why this raises red flags:
Similar to WeWork’s model, CoreWeave is signing long-term leases to scale quickly while keeping the liabilities off the books.
If AI compute demand weakens, CoreWeave could be locked into expensive, underutilized capacity.
A shift to owning data centers would change the financial profile, requiring billions in CapEx instead of lease agreements.
For now, investors need to treat this $15B in leases as hidden leverage—if market conditions shift, these commitments could become a major financial burden.
Speaking of financials…
Financials
$1.4B in cash on the balance sheet
$7.9B in debt
Massive facilities they can tap into for capex needs
A large convertible note to Magnetar, which we’ll discuss below
As mentioned, there’s $15B missing… the Off-Balance-Sheet Leverage Masks True Debt Load
These obligations aren’t classified as debt, which makes CoreWeave’s balance sheet appear less risky than it actually is.
Revenue grew at 737% last year, while cost of revenue grew at 617%, showing some early leverage in their model
Cost of Revenue includes data centers, utilities for power, and the people involved in data center operations and customer success. There’s also a big depreciation and amortization charge in here that gets added back in an EBITDA bridge
Technology and infrastructure expense consists of costs associated with servers, switches, networking equipment and internally developed software, and the engineers required to build their products
The sales organization is focused on a direct named account strategy to drive demand from the world’s leading AI labs and AI enterprises. They supplement this with a product-led growth (“PLG”) motion serving individual users and developers working at AI labs and AI enterprises.
They somehow only spent $18M on S&M last year (or 1% of revenue)
And this was only a 42% change y/y despite revenue growing 737%
Interest expenses go to paying down their debt facilities which are used to fund capex investments
They spent $332M on interest last year, representing 19% of revenue. This is up from $28M the year prior
Last year they spent 18x more on interest payments than on S&M expenses
Net loss margin was cut from 259% of revenue to just 45% of revenue year on year
Demonstrates that while they are investing a massive amount in build out from an absolute dollar perspective, the losses look better on a relative basis to revenue
Investors, Ownership, and Valuation
Founders, Investors, and Ownership
Founder Control and Voting Power
CoreWeave’s three co-founders—CEO Michael Intrator, CTO Brian Venturo, and Chief Strategy Officer Brannin McBee—hold 83% of the company’s voting power through a dual-class share structure, ensuring they maintain control. However, this dominance hinges on Intrator’s position—if he leaves, their special Class B shares convert to common stock, reducing their grip.
Pre-IPO Cash-Out
Before the IPO, the founders sold 11–16% of their stakes, collectively cashing out nearly $500 million. This de-risks their personal exposure while still leaving them with multi-billion-dollar stakes based on CoreWeave’s target ~$35B valuation.
Key Institutional Investors
NVIDIA owns about 1% of CoreWeave—small but strategically significant given CoreWeave’s reliance on NVIDIA GPUs.
Blackstone led a $7.5B debt facility to fund data center expansion.
Magnetar Capital co-led multiple financing rounds and structured a convertible note deal giving them discounted IPO shares (more on this “interesting” set up later)
Coatue Management led CoreWeave’s $1.1B Series C, betting heavily on AI infrastructure. They’re a known cross over investor who can exit their existing shares and / or double down in the public markets.
DigitalBridge backs CoreWeave’s real estate and data center expansion.
Other Investors with Minority Stakes
Altimeter Capital, Assured Asset Management, Claridge Venture Partners, Interplay, and IronArc Ventures hold smaller positions, likely looking for a post-IPO exit.
Who Stands to Gain the Most?
Founders still hold billions in equity, even after partially cashing out.
Magnetar structured a sweetheart deal, getting shares at a steep discount.
NVIDIA’s small stake signals strategic interest rather than deep financial commitment.
Private equity players like Blackstone and Coatue may seek an exit post-lockup, adding potential selling pressure.
How High Can the Valuation Go?
Revenue Growth Is Insane – From $16M in 2022 → $1.9B in 2024 (+737% YoY).
Massive Contract Backlog – $15.1B in remaining performance obligations (RPO).
Market Loves AI Infrastructure (At the Moment) – If investors see CoreWeave as an “AI infrastructure pure play,” it could command a premium.
Right now, AI compute is being valued like a scarce commodity. If investors treat CoreWeave more like an AI hardware supplier (NVIDIA) than a cloud provider (AWS), the stock could pop.
The AWS vs. NVIDIA Valuation Debate
How the market values CoreWeave depends on what bucket investors put it in:
IPO Valuation Math: The Ballpark Estimate
Last private valuation was at $23 billion in November of 2024
They’ve raised a total of $4B of preferred capital to date and are looking to raise another $3 to $4B at a $35B valuation
If they grow revenue next year even just 100% (I know, it’s wild to say doubling is the “lower expected range of growth”) to $4B that would still be less than a 10x forward revenue multiple
Likely Market Multiple: 15x forward (depending on hype vs. fundamentals)
Implied Valuation: $30B – $38B
For reference:
If CoreWeave leans into AI scarcity hype, we could see $30B+ valuation territory. If the market starts worrying about cash burn, we’re probably closer to their last round in the high $20B’s.
Miscellaneous Stuff of Note
Payback period
CoreWeave’s customers prepay 15–25% of their contract value upfront, allowing the company to fund infrastructure without tapping as much outside capital. The company estimates that GPUs reach full payback within 2.5 years, meaning that from year three onward, they generate pure cash flow. They’ll want to keep their contracts above that point, right now they’re at 4 years.
The Shift to Owning Data Centers Could Reshape the Business
CoreWeave currently rents data center space, but the S-1 suggests it may begin building its own infrastructure. This would be a massive shift in capital allocation, bringing it closer to the AWS model—but also requiring billions in upfront investment.
IPO Ratchets Protect Insiders at the Expense of Retail Investors
Series B and C preferred stockholders have built-in downside protection if the IPO prices below a certain threshold. If that happens, they automatically receive extra shares—effectively diluting common stockholders. This structure shields institutional investors from risk while shifting it to retail buyers.
Lock-Up Expirations Could Lead to a Sell-Off
Insiders are generally locked up for 180 days post-IPO, but certain conditions allow early selling. Expect significant selling pressure after six months, particularly from private equity investors who are likely looking to exit.
Magnetar’s $230M "Cloud Deposit" Looks More Like Financial Engineering
In August 2024, CoreWeave received a $230M refundable deposit from Magnetar under a pre-negotiated cloud services deal. If Magnetar’s portfolio companies don’t use the services, the deposit is refunded with a 1.1x to 1.7x markup. This structured deal raises the question: is this real revenue, or a form of synthetic financial engineering?
Magnetar’s Special IPO Discount Puts Retail Investors at a Disadvantage
Magnetar also bought convertible notes that converted into CoreWeave stock at the IPO price. However, they also received a "Penny Warrant," allowing them to buy even more shares for $0.01 each. This gives Magnetar a better deal than retail investors, who are buying in at full price.
Final Thought: The IPO Is Structurally Built for Insiders
Between IPO ratchets, early liquidity options, and structured deals with key investors, the CoreWeave IPO is designed to protect insiders while leaving public investors exposed to more risk. While the company is growing rapidly, its financial structure raises the question of whether this IPO is a long-term public play—or just a bridge to an eventual acquisition.
🚀 Bull Case: The AI Compute Supercycle Continues
Valuation: $40B+ (+20x forward revenue multiple)
AI model sizes and compute needs keep growing, fueling GPU demand. CoreWeave remains a key player, benefiting from scarcity pricing.
Revenue skyrockets beyond expectations as AI startups and enterprises lock in long-term contracts.
Microsoft and other hyperscalers continue outsourcing AI workloads, rather than fully internalizing their compute.
Gross margins expand as CoreWeave improves efficiency and potentially builds its own data centers to cut costs.
Public markets reward AI infrastructure stocks, treating CoreWeave more like NVIDIA than a commodity cloud provider.
Stock price surges post-IPO, and CoreWeave becomes a long-term AI infrastructure leader.
What it looks like: A “Snowflake for AI Compute” narrative, commanding a premium valuation as the go-to hyperscaler alternative.
🐻 Bear Case: AI Compute Demand Normalizes & Hyperscalers Eat Its Lunch
Valuation: $10B–$15B (5–7x forward revenue multiple)
Microsoft, AWS, and Google ramp up their own AI compute offerings, reducing reliance on CoreWeave.
GPU supply chain constraints ease, leading to lower pricing power and eroding margins.
Customer concentration risk materializes—Microsoft or a major AI startup pulls back, shrinking CoreWeave’s backlog.
Cash burn accelerates as CapEx commitments for GPUs and data centers outstrip revenue growth.
Stock struggles post-IPO as investors realize CoreWeave’s margins resemble traditional cloud businesses, not AI SaaS.
CoreWeave pivots or gets acquired at a discount by a cloud giant looking for additional compute capacity.
What it looks like: A high-growth, capital-intensive business that struggles to achieve profitability and loses its competitive moat.
📉 Base Case: High Growth, High Cash Burn, with an Uncertain Endgame
Valuation: $20B–$30B (10–15x forward revenue multiple)
CoreWeave continues growing fast, but AI infrastructure becomes more competitive.
Revenue climbs steadily, but profitability remains elusive due to constant reinvestment in GPUs and data centers.
Microsoft stays a key customer but negotiates better pricing, squeezing CoreWeave’s margins.
IPO pops, but stock remains volatile, as investors debate whether this is a long-term business or an acquisition target.
M&A rumors emerge as hyperscalers or private equity explore buying CoreWeave to absorb its customer base.
What it looks like: A highly valuable but capital-hungry AI infrastructure provider that either matures into a niche hyperscaler or gets rolled up into a larger cloud player.
Which case seems most likely? If AI demand remains white-hot, CoreWeave’s upside is massive—but if hyperscalers start competing more directly, the story changes fast.
None of this is investment advice. I wrote this with my dog on the couch. Do your own homework.

TL;DR: Multiples are DOWN week-over-week.
Top 10 Medians:
EV / NTM Revenue = 13.2x (DOWN 2.8x w/w)
CAC Payback = 29 months
Rule of 40 = 54%
Revenue per Employee = $392k
Data source: Koyfin

Figures for each index are measured at the Median
Median and Top 10 Median are measured across the entire data set, where n = 110
Population Sizes:
Security: 18
Database and Infra: 14
Backoffice: 16
Marcom: 16
Marketplace: 15
Fintech: 16
Vertical SaaS: 16
Revenue Multiples
Revenue multiples are a shortcut to compare valuations across the technology landscape, where companies may not yet be profitable. The most standard timeframe for revenue multiple comparison is on a “Next Twelve Months” (NTM Revenue) basis.
NTM is a generous cut, as it gives a company “credit” for a full “rolling” future year. It also puts all companies on equal footing, regardless of their fiscal year end and quarterly seasonality.

However, not all technology sectors or monetization strategies receive the same “credit” on their forward revenue, which operators should be aware of when they create comp sets for their own companies. That is why I break them out as separate “indexes”.
Reasons may include:
Recurring mix of revenue
Stickiness of revenue
Average contract size
Cost of revenue delivery
Criticality of solution
Total Addressable Market potential
From a macro perspective, multiples trend higher in low interest environments, and vice versa.
Multiples shown are calculated by taking the Enterprise Value / NTM revenue.
Enterprise Value is calculated as: Market Capitalization + Total Debt - Cash
Market Cap fluctuates with share price day to day, while Total Debt and Cash are taken from the most recent quarterly financial statements available. That’s why we share this report each week - to keep up with changes in the stock market, and to update for quarterly earnings reports when they drop.
Historically, a 10x NTM Revenue multiple has been viewed as a “premium” valuation reserved for the best of the best companies.
Efficiency Benchmarks
Companies that can do more with less tend to earn higher valuations.

Three of the most common and consistently publicly available metrics to measure efficiency include:
CAC Payback Period: How many months does it take to recoup the cost of acquiring a customer?
CAC Payback Period is measured as Sales and Marketing costs divided by Revenue Additions, and adjusted by Gross Margin.
Here’s how I do it:
Sales and Marketing costs are measured on a TTM basis, but lagged by one quarter (so you skip a quarter, then sum the trailing four quarters of costs). This timeframe smooths for seasonality and recognizes the lead time required to generate pipeline.
Revenue is measured as the year-on-year change in the most recent quarter’s sales (so for Q2 of 2024 you’d subtract out Q2 of 2023’s revenue to get the increase), and then multiplied by four to arrive at an annualized revenue increase (e.g., ARR Additions).
Gross margin is taken as a % from the most recent quarter (e.g., 82%) to represent the current cost to serve a customer
Revenue per Employee: On a per head basis, how much in sales does the company generate each year? The rule of thumb is public companies should be doing north of $450k per employee at scale. This is simple division. And I believe it cuts through all the noise - there’s nowhere to hide.
Revenue per Employee is calculated as: (TTM Revenue / Total Current Employees)
Rule of 40: How does a company balance topline growth with bottom line efficiency? It’s the sum of the company’s revenue growth rate and EBITDA Margin. Netting the two should get you above 40 to pass the test.
Rule of 40 is calculated as: TTM Revenue Growth % + TTM Adjusted EBITDA Margin %
A few other notes on efficiency metrics:
Net Dollar Retention is another great measure of efficiency, but many companies have stopped quoting it as an exact number, choosing instead to disclose if it’s above or below a threshold once a year. It’s also uncommon for some types of companies, like marketplaces, to report it at all.
Most public companies don’t report net new ARR, and not all revenue is “recurring”, so I’m doing my best to approximate using changes in reported GAAP revenue. I admit this is a “stricter” view, as it is measuring change in net revenue.
Operating Expenditures
Decreasing your OPEX relative to revenue demonstrates Operating Leverage, and leaves more dollars to drop to the bottom line, as companies strive to achieve +25% profitability at scale.

The most common buckets companies put their operating costs into are:
Cost of Goods Sold: Customer Support employees, infrastructure to host your business in the cloud, API tolls, and banking fees if you are a FinTech.
Sales & Marketing: Sales and Marketing employees, advertising spend, demand gen spend, events, conferences, tools.
Research & Development: Product and Engineering employees, development expenses, tools.
General & Administrative: Finance, HR, and IT employees… and everything else. Or as I like to call myself “Strategic Backoffice Overhead.”
All of these are taken on a Gaap basis and therefore INCLUDE stock based comp, a non cash expense.
This is an interesting development for CoreWeave. It will be fascinating to see how their public offering impacts the AI cloud computing landscape.
Don't sleep on my boy $nbis as an ai hyperscaler...but I like coreweave.