Pre-seed valuation 2026: SAFE caps, ranges, and dilution math

Pre-seed valuation in 2026 is not a value you defend, it is a SAFE cap you negotiate. Carta's Q2 2025 data puts the median post-money cap at $7.5M for raises under $250K and $10M for $250K to $500K. Treat the number as a dilution decision, not a DCF exercise.

Contents

• Pre-seed valuation 2026 at a glance
• Your pre-seed valuation is a SAFE cap, not a number you defend
• Pre-seed SAFE cap benchmarks by raise band
• How much to raise at pre-seed (and why size sets the cap)
• Value backward from your pre-seed dilution target
• Pre-seed post-money valuation: how the cap converts to equity
• What a too-high pre-seed cap actually costs you
• The AI premium at pre-seed: does it flow through?
• Get your number, then create the competition that moves it
• FAQ

Most pre-seed valuation guides hand you a range, "$3M to $10M depending on the market," and tell you to find your spot in it. That framing is wrong at the root. At pre-seed you have almost nothing to price on: no revenue, often no shipped product, sometimes no incorporated entity. There is no defensible model, so the number is not a valuation anyone argues from a spreadsheet.

It is a SAFE cap, and a cap is a negotiated dilution decision. Two things set it, and neither is your pitch: the prevailing post-money cap conventions for your raise size, and how much ownership you are willing to give up before a priced seed. This guide gives you the 2026 cap ranges by raise band, the "value backward from your dilution target" method to replace copying a friend's number, and where the AI premium actually shows up.

Pre-seed valuation 2026 at a glance

The number you negotiate is a post-money SAFE cap, and it tracks your raise size, not your ambition. Here are the 2026 anchors from Carta's primary data, the exact figures to bring into a term conversation.

Read it this way: your band, not the headline range, is what matters. A founder raising $200K and a founder raising $450K are in two different cap markets even though both call it "pre-seed." Find your raise band first, then read the median off the row that matches it.

Your pre-seed valuation is a SAFE cap, not a number you defend

Stop treating the pre-seed number as a valuation and start treating it as a cap you trade against dilution. A priced round has a real pre-money and post-money valuation because there is a share price. A pre-seed SAFE does not: the cap is a ceiling on the price your money converts at later, nothing more.

That distinction changes what you optimize for. You are not defending a value, you are choosing how much of the company to sell for the check you need. The standard 2025 structure is a post-money SAFE with a cap and no discount, and among the minority of SAFEs that do carry a discount, 63% use a 20% rate. Cap-only is the market, so the cap is the whole negotiation.

The dominance of the instrument is not a style choice, it is the entire pre-seed market. In 2025, 93% of pre-seed rounds under $500K were done on SAFEs, and 97% of all sub-$500K rounds on Carta were pre-seed rounds. If you are raising under half a million dollars, you are almost certainly setting a cap, not pricing equity. For the priced-versus-SAFE decision at larger sizes, see the breakdown in SAFE vs priced round.

Pre-seed SAFE cap benchmarks by raise band

Caps sort cleanly by raise size, and the 2025 medians are the anchor to bring to the table. The two published Carta bands are the most useful numbers a pre-seed founder has, because they are broken out at the granularity you actually raise at.

Both figures come from Carta's State of Pre-Seed Q2 2025.

The pre-seed market is going barbell-shaped, which is why one median hides two realities. Carta reports the smallest rounds under $500K and the largest at $3M+ are both growing while the middle disappears. A tiny angel round and a $3M institutional pre-seed are priced by different investors with different math, so do not average across them.

Caps also lag the priced market, so do not read a hot seed headline as your ceiling. Carta observed that SAFE caps at pre-seed stayed flat in Q1 2024 even as priced seed valuations moved, which suggests caps trail the priced market by one to two quarters. If seed valuations spiked last quarter, your pre-seed cap has probably not caught up yet, and pricing as if it has will read as greedy.

How much to raise at pre-seed (and why size sets the cap)

Decide the raise first, because the raise sets the cap, not the other way around. Carta's own pre-seed guidance is explicit: size the round to the milestone you need before a priced seed, then work backward to the cap. Raise size is the primary anchor for the number, so anchoring on the number first is backwards.

Size to a milestone, not to a runway you can brag about. The milestone that matters is whatever unlocks a priced seed: a working product, early usage, a first cohort of paying customers, a technical breakthrough. Everything you raise should buy the shortest credible path to that proof point and nothing more.

Use YC's cost math to floor the number:

• Engineer cost rule: YC pegs a fully loaded engineer at roughly $15K per month, so an 18-month runway for a five-person team implies a raise floor around $1.35M at the seed stage. At pre-seed you are usually funding a smaller team for a shorter proof window, so your floor is lower.
• The YC benchmark: YC invests $500K for 7% plus an MFN, which implies a post-money near $7.1M. Downstream investors price against that number, so it is a useful reference point even if you never go through YC.
• Do not over-raise to justify a bigger cap. A bigger raise at a flat cap sells more of the company. If you cannot tie the extra $300K to a specific milestone, cut it. For the equivalent decision one stage later, see how much to raise at seed.

Value backward from your pre-seed dilution target

The correct method is one line of arithmetic: cap equals raise divided by your target dilution. This is the single most useful move in this guide, and it replaces copying a cap you saw a friend close at. Pick the dilution you are willing to eat, divide the raise by it, and you have the cap that market data should sanity-check, not dictate.

YC's dilution guidance gives you the target:

• Aim for ~10%, the level YC recommends founders target.
• Accept up to 20% if the round genuinely requires it, per the same YC guidance.
• Do not cross 25% of the company sold before your priced seed, or you enter Series A with too little founder equity to survive the next two rounds.

Worked example: if you need $500K and you are willing to give up 5%, your cap is $500K ÷ 0.05 = $10M, which happens to sit right on the Carta median for that raise band. If you can hit your milestone on $300K, the same 5% target gives a $6M cap, and you keep the difference.

Dilution also rises with raise size at a predictable clip:

Both figures are from Carta's State of Seed 2025 data. The lesson: bigger raises cost proportionally more of the company, so a larger round has to buy a proportionally bigger milestone to be worth it. For the full stage-two picture, see dilution at seed.

Pre-seed post-money valuation: how the cap converts to equity

On a post-money SAFE, ownership sold equals check divided by cap, which is why the cap only means anything next to the raise. This is the mechanic that round-number caps hide. A $10M cap sounds identical whether you raise $250K or $1M against it, but the equity you part with is four times larger in the second case.

Here is the conversion at common pre-seed combinations:

A cap in isolation tells you nothing about what you sold. The $10M cap costs 5% on a $500K raise and 10% on a $1M raise. Selling the same 5% while raising $1M would require a $20M cap, which is far above the 2025 median for any pre-seed band. When someone quotes you their cap, it is meaningless until you also know their raise.

Stacked SAFEs at different caps compound this quietly. Two or three SAFEs on similar terms are clean. Five SAFEs at caps from $6M to $18M convert into a tangle of ownership percentages that your priced-seed lead will force you to reconcile, sometimes painfully, when the round prices.

A pre-seed cap set too high is not a trophy. It is a loan against your priced seed round, and the interest is paid in your own dilution.

What a too-high pre-seed cap actually costs you

A cap above what your seed can support is an implicit loan against that seed round. This is the frame no lookup-range guide gives you, and it is the one that saves founders from a bad negotiation win. If you push the cap to a number your priced seed cannot clear, you have not banked value, you have created a repayment problem with a due date.

The mechanics are unforgiving. Your SAFE converts at the seed, and a sky-high cap forces the seed price to absorb the implied step-up. If your cap implies more ownership than your seed metrics justify, one of two things happens: your seed valuation gets compressed to make the math work, or your SAFE converts at a lower effective price and you eat the dilution you thought you had dodged. Either way, the bill comes due at the round that actually prices your equity.

A flat or slightly down seed is a real risk when the pre-seed cap outran the story. Investors read a pre-seed cap that is 40% above your band as a signal you will fight valuation over substance, and it kills momentum before diligence starts. The clean move is a cap you can defend at the next round, so the seed prices up and everyone in your cap table is aligned to raise it.

Do not optimize the cap in isolation from the seed you will actually raise. The number that maximizes your pre-seed ownership on paper is often the number that caps your seed valuation in practice. Anchor the cap to the priced-seed post-money you are realistically targeting, then leave room above it.

The AI premium at pre-seed: does it flow through?

The AI premium is real at priced seed rounds, but it does not drop cleanly onto your pre-seed cap. This is the question every AI founder asks, and the honest answer is that the premium is a seed-stage priced-round phenomenon that reaches pre-seed only indirectly and on a lag.

The seed-stage numbers are large and well-documented:

At pre-seed the premium shows up as a slow lift in small-raise caps, not a multiplier you can quote. Carta attributed the rising median cap on sub-$250K raises partly to investors "now foreseeing higher upside for very early-stage startups", which it linked to AI productivity gains. That is a general tide lifting the $6.5M-to-$7.5M median, not a 42% badge you get for saying "AI."

Do not price a seed-stage AI premium into your pre-seed cap unless a lead is validating it. The premium crystallizes when a priced seed round competes for your equity, not when you write "AI-native" on a SAFE. Bank it at the stage where the market actually pays it, and compare your seed expectations against seed valuation benchmarks rather than pre-seed anecdotes.

Get your number, then create the competition that moves it

Setting the right cap is half the job; the other half is creating the competition that lets you raise it. Once you have valued backward from your dilution target and sanity-checked it against your raise band, the cap on paper is fixed. What actually moves the number off that anchor is not a better deck, it is a second interested investor.

The single biggest lever on your cap is the count of real conversations you can run in parallel. A cap is a negotiation, and negotiations move when there is a credible alternative in the room. One motivated investor with no competition sets the floor. Two who both want in set a higher one, because now saying no has a cost for them.

Knowing your number is half the job. The other half is running enough of the right conversations in parallel that a second term sheet appears, which is what actually moves the price. Causo matches you to the investors most likely to fund your stage and sector and drafts the outreach, so you can run that many conversations without spending three weeks list-building.

FAQ

What is the average pre-seed valuation in 2026? There is no single average, because pre-seed "valuation" is a SAFE cap set by raise size. Carta's Q2 2025 data puts the median post-money cap at $7.5M for raises under $250K and $10M for raises of $250K to $500K. Larger pre-seed raises of $1M+ carry higher caps, and AI companies with a lead can push above these medians.

What is a normal SAFE cap for a pre-seed round? For a sub-$250K raise, the 2025 Carta median post-money cap was $7.5M; for $250K to $500K it was $10M. The default instrument is a post-money SAFE with a cap and no discount. Anchor your cap to the medians for your raise band, not to a round number you saw on Twitter.

How much equity do you give up at pre-seed? On a post-money SAFE, the equity sold equals your raise divided by the cap. A $500K raise on a $10M cap sells 5%, and a $250K raise on a $7.5M cap sells about 3.3%. YC's guidance is to target roughly 10% dilution, accept up to 20% if the round demands it, and avoid giving up more than 25% before your priced seed.

How do you value a pre-seed startup with no revenue? You do not value it, you price the round backward from raise size and dilution. Size the raise to the milestone you need before a priced seed, then divide by your target dilution to get the cap. Carta's own pre-seed guidance treats raise size, not projections, as the primary anchor for the number.

What happens if my SAFE cap is too high? A cap set above what a priced seed can support becomes an implicit loan against that round. Your seed investors have to absorb the step-up in implied ownership, which either compresses your seed valuation or forces a down-conversion that dilutes you anyway. A cap that is too high is a repayment problem you postpone, not a win you bank.

Related on the hub

• Seed valuation 2026: fair ranges, SAFE caps, and dilution math — Related valuation guide.
• The H1 2026 Pre-Seed Funding Report — Related fundraising basics guide.
• Seed round valuation 2026: the benchmark report — Related fundraising basics guide.