How early-stage startup valuation works

The short version:

• Most valuation advice hands founders a formula, scorecard, step-up, checklist methods, as if an early-stage valuation is computed. At pre-seed and seed, it mostly isn't. There are no meaningful financials to plug in, so the number is negotiated, not calculated.
• What actually sets the number: what investors in your stage and category are currently paying (comparables), how much leverage you have (competing interest), and the round's structure, most early rounds are SAFEs or notes with a valuation cap, which isn't a valuation at all.
• The "methods" still have a use, but it's narrow: they help you arrive at a defensible range to anchor a negotiation, not a true value. Treating their output as fact is how founders talk themselves into a number the market won't support.
• The part you control isn't the valuation. It's your credibility going into the room: clean books, a clean cap table, a burn-and-runway number you know cold, and a round structure you understand. That's what lets you hold a number rather than hope for one.

Search "how to value an early-stage startup" and you'll find the same thing everywhere: a set of methods with reassuring names, the scorecard method, the step-up method, the checklist method, risk-factor summation, each promising to turn your team, market, and traction into a number. It's comforting, and it's mostly a polite fiction. At pre-seed and seed, your startup doesn't have a value that a formula can discover. It has a price, and price is set by what someone will pay. 

That distinction sounds academic until you're in the room. A founder who believes their valuation was calculated walks in trying to defend a number as if it were a fact, and gets confused and defensive when an investor simply doesn't agree. A founder who understands their valuation is negotiated, walks in knowing the number is a meeting point between what they'd like and what the market will bear, and spends their energy on the things that actually move it. This piece is about what those things are, and about the one input a founder fully controls.

Why the formulas don't really work at the earliest stage

Valuation methods work by plugging real inputs into a model: revenue, growth, margins, comparable multiples. That's how a Series B gets valued, and it's reasonably rigorous, because there's data. Run the same exercise on a pre-seed company and you're plugging in guesses. The "market size" is an estimate, the "traction" is a handful of early signals, the "team score" is a judgment call. The formula gives you a number, but the number inherits all the uncertainty of the inputs, precise-looking, and no more reliable than the assumptions underneath it. 

This is why the four famous early-stage methods disagree with each other, sometimes wildly, on the same company. They aren't measuring a hidden true value. They're each a structured way of guessing. Useful as a sanity check, useless as a source of truth. Any founder who has run their startup through two of these methods and gotten two different answers has already discovered this. The methods don't fail because you did them wrong. They fail because there's no fact there to find yet.

So if the number isn't computed, where does it actually come from?

What actually sets the number: comps, leverage, and structure

Three forces do almost all the work in an early-stage valuation, and none of them is a formula.

Comparables and stage norms

The strongest gravity on your number is simply what companies like yours, same stage, same category, same geography, same rough quality, are raising at right now. Investors price against a live market they see every day. They know the going range for a seed AI company or a pre-seed fintech this quarter, and your number will land inside that range unless there's a strong reason otherwise. Most of "valuation" is really "where do I sit within the current band for companies like me."

Leverage

Within that band, where you land is decided by demand for your round. One interested investor and you take roughly the number they offer. Several competing for the allocation and you move to the top of the band or beyond. This is why founders talk so much about creating a process, running conversations in parallel, generating real competition. The valuation isn't a property of your company in isolation. It's a function of how many people want in at once. Nothing lifts a number like another term sheet.

Structure

Here's the part the formula-based framing misses entirely: most early rounds don't set a valuation at all. They're SAFEs or convertible notes with a valuation cap, a ceiling that converts to equity later, not a price paid today. A $10M cap is not a $10M valuation. It's a negotiated ceiling on a future conversion, and the interplay of the cap, any discount, and how much you raise on the instrument determines your real eventual dilution. Founders who fixate on the cap as if it were a valuation often miss that the structure, not the headline number, decides what they actually give up.

Put together: your number is roughly the current market band for companies like yours, moved up or down by how much competition you've created, expressed through a structure that may not even be a "valuation." That's a negotiation, not a calculation.

What the current market band actually looks like

Because comparables do most of the work, it's worth knowing roughly where the bands sit, not to pin your number, but so you can tell a defensible ask from a fantasy. The most reliable public source is Carta, which processes equity for tens of thousands of startups and publishes quarterly benchmarks. A snapshot from their 2025 data (which shifts, so check the latest before you raise): 

At pre-seed, there usually isn't a priced valuation at all. You raise on a post-money SAFE with a valuation cap, now the standard instrument. In 2025, median caps ran around $10M for rounds in the $250K-$1M range and roughly $15M for rounds in the $1M-$2.5M range, with meaningful variation by sector (crypto, biotech, and healthtech tend to command higher caps). At seed (roughly $2-5M raised), median pre-money valuations were around $16M and post-money around $20-24M. Sector matters a lot: AI-native companies carried noticeably higher medians than the broader market. And a structural fact worth internalizing: the majority of early rounds under about $4M are now done on SAFEs or notes, and the "cap-only, no discount" post-money SAFE has become the default template. 

Two things to take from the numbers. First, the bands are wide and move quarter to quarter, which is exactly why a formula can't pin your value, the market itself is a moving range. Second, knowing the band is most of what "a defensible number" means: a $6-8M cap on a small pre-seed is easy to defend with the data behind it. A $30M cap on the same round is a conversation you'll lose.

So why compute a valuation at all?

Because a negotiation still needs an anchor, and "I don't know, what do you think?" is a terrible one. 

This is the real, narrower use of the methods and the comps: not to find your true worth, but to arrive at a defensible range you can walk in with. You want to be able to say, credibly, "companies at our stage and traction in our space are raising around here, and here's why we're at the upper end of that," and have the numbers behind it hold up. The valuation exercise is preparation for a negotiation, not a substitute for one. Its output is a range and a rationale, not a verdict. 

Which reframes what "getting your valuation right" even means. It doesn't mean finding the highest number a method will justify. It means walking in with a number you can defend, that sits sensibly against real comparables, that leaves room to grow into the next round, and that you have the credibility to hold.

The trap on the other side: the number that's too high

It's worth saying plainly, because the incentive all points one way: the highest valuation you can get is often not the one you want. 

A number inflated above what your fundamentals support doesn't disappear. It becomes the bar you have to clear next time. Raise at a valuation you can't grow into, and your next round is a down round, new money at a lower number, which is bruising for morale, triggers anti-dilution provisions that can savage your ownership, and signals trouble to the market. Founders reaching for a headline number also tend to trade away on terms what they won it on price, accepting liquidation preferences or provisions that cost them more than the higher valuation gained. A slightly lower number from a better investor, with room to grow into it, beats a trophy valuation that becomes an anchor. The goal was never the maximum. It was the right number, defensibly held, on clean terms.

The one thing you actually control

Notice that almost everything above, the market band, the competition, the comps, is outside your direct control on the day. There's one input that isn't, and founders underrate it: your credibility in the room. 

You can't dictate the market's range, but you can be the founder who clearly deserves the top of it rather than the bottom. That comes down to unglamorous things. Numbers you can speak to fluently, a burn-and-runway figure you can state instantly, because command of your own finances reads as command of your business. Records that hold up, clean, current books so any figure you cite ties out and nothing surfaces later to spook an investor. A cap table that's accurate, with your existing SAFEs and notes properly accounted for, so the ownership math an investor runs actually works and there are no undocumented surprises. And a real understanding of the structure you're offering, cap, discount, dilution, so you're negotiating rather than nodding along. 

None of that changes the market's range. All of it changes where in the range you land, and whether the number survives diligence once it's agreed. A founder who walks in with messy books and a confused cap table hands the investor every reason to negotiate down or walk. A founder whose numbers are airtight removes those reasons and can hold a higher number with a straight face. The valuation is negotiated, but the strength of your position in that negotiation is something you build in the months before, in the quiet work of keeping your house in order.

The takeaway

Stop trying to calculate your early-stage valuation and start preparing to negotiate it. At this stage there's no true number for a formula to find. There's a market band set by comparables, a position within it set by how much demand you create, and a structure, usually a SAFE cap, not a valuation, that decides what you actually give up. Use the methods to build a defensible range, not to manufacture a verdict. Resist the trophy number that becomes next year's anchor. And pour your controllable energy into the one input that's yours: clean books, a clean cap table, and total fluency with your own numbers. The valuation is set in the room. Your strength walking into the room is set long before.

Frequently asked questions

How do you value an early-stage or pre-revenue startup? 

At pre-seed and seed there's rarely enough data for a formula to produce a meaningful value, so the number is negotiated rather than calculated. In practice it's set by what comparable companies at your stage, sector, and geography are currently raising at, moved up or down by how much investor competition you've created, and often expressed as a valuation cap on a SAFE or note rather than a priced valuation. Methods like the scorecard or step-up are useful for building a defensible range to anchor the negotiation, not for finding a true value.

Which startup valuation method is most accurate for early stage?

None is truly accurate at the earliest stage, because they all rely on inputs (market size, traction, team quality) that are estimates for a pre-seed company, which is why the methods often disagree on the same startup. Their real value is as a structured way to arrive at a defensible range, not a precise figure. Use one or two to sanity-check where you sit against real comparables, then treat the output as a negotiating anchor rather than a verdict.

What actually determines a startup's valuation at seed stage? 

Three things do most of the work: comparables (what similar companies are raising at right now), leverage (how much competing investor interest you've generated), and structure (most seed rounds use SAFEs or notes with a valuation cap rather than a set valuation). Your financials matter less as inputs to a formula and more as credibility, clean books and a clean cap table let you hold the top of the market range and survive diligence.

What's the difference between a valuation cap and a valuation? 

A valuation is a price agreed today in a priced equity round, setting exactly how much of the company the investment buys. A valuation cap, used in SAFEs and convertible notes, is a ceiling on the valuation at which that instrument will later convert to equity, a negotiated future limit, not a price paid now. A $10M cap is not a $10M valuation, and the interplay of cap, discount, and amount raised determines your actual eventual dilution, which is why founders should model the conversion rather than fixate on the cap number.

Is a higher valuation always better for founders? 

A valuation above what your fundamentals support becomes the bar you must clear next round, and missing it forces a down round, raising at a lower number, which triggers anti-dilution provisions, damages morale, and signals weakness. Founders chasing a headline number also often concede costly terms to get it. A slightly lower valuation from a strong investor, with room to grow into it and clean terms, is usually better than an inflated one that becomes an anchor.

How do I prepare my startup's financials before raising?

Get your books clean and current so every figure ties out, keep your cap table accurate and your SAFEs or notes properly recorded, and know your burn and runway cold. This won't change the market's valuation range, but it determines where in that range you land and whether the number survives diligence, a founder whose numbers are airtight can hold a higher figure than one explaining messy records. Setting this up early, ideally from incorporation, makes it effortless when a raise arrives.

Further reading (worth your time)

• Real, current benchmarks. Carta's quarterly State of Pre-Seed and State of Seed reports are the best public data on what founders actually raise at, median valuation caps, dilution, deal terms, by round size and sector. Carta Data & Insights. Free to read, updated every quarter.
• The standard instrument, from the source. Y Combinator created the post-money SAFE that most pre-seed rounds now use. Their explainer and templates are the primary reference. Y Combinator, Understanding SAFEs and Priced Equity Rounds and the SAFE documents.
• How dilution actually compounds. A clear walkthrough of pre-money vs post-money and how caps convert. Carta, Post-money vs pre-money SAFEs and any SAFE conversion calculator to model your own cap table.
• The valuation methods, explained. If you want to understand the scorecard, step-up, and risk-factor methods in depth (as range-setting tools), the Corporate Finance Institute's overview is solid. CFI, Startup Valuation Methods.
• Founder-side fundraising strategy. For how leverage and process actually move a number, the writing at First Round Review and Paul Graham's essay How to Raise Money are worth reading before you start a raise.