Startup Dilution per Round: 2026 Benchmarks by Stage

Short answer for 2026: founders sell about 18% of the company at seed, roughly the same again at Series A, and a shrinking slice in every round after that. Those are the headlines. What matters is what you own after the option pool top-up and the SAFE conversions land in the same round, usually 5–10 points worse. Here are the benchmarks by funding stage, the math, and the levers.

Median dilution on Carta is 18% at seed and again at Series A, then falls with every later round. Founders typically hold about 56% after seed and 36% after Series A. Our healthy band for each round, in the table below, sits around the Carta median but is measured after the pool and the SAFEs, not before.

How much dilution per round is typical in 2026?

Typical dilution falls with every round. On Carta's latest benchmarks, seed and Series A both sit at 18%, and each later round sells a smaller slice, down to single digits from Series C on. Pre-seed is different: it's a stack of SAFEs, and a median-sized raise on the median cap sells about 13%. That's lower than any point in recent years, but only if your valuation clears the median.

The 2026 picture is unusually founder-friendly at the median. The round figures in the table come from more than 1,000 priced rounds closed by software startups on Carta in the six months to July (Carta). The ownership column comes from Carta's founder ownership report. The backdrop is a barbell: megadeals and AI companies took the bulk of US venture dollars in the first half of the year, and the table below has the split. Medians describe the middle of that barbell, not your deal.

If you're not an AI company, read the medians as the optimistic case. Carta's Q1 report puts the non-AI Series A median at a $55M valuation, a fraction of what AI foundational model companies command (Carta, State of Private Markets). Raise the median Series A round on a $55M post-money and you've sold 26%, not 18%. The market didn't get generous in 2026. It got bifurcated.

What is startup dilution, and what actually causes it?

Dilution is the drop in your ownership percentage when the company issues new shares. Your share count never changes; the denominator does. New investors are the visible cause. The invisible ones are the option pool top-up investors require before they price the round, SAFEs converting at their caps, and anti-dilution clauses that reprice earlier investors if you ever raise a down round.

The formula is one line: dilution = new shares issued ÷ fully diluted shares after the round, or money raised ÷ post-money valuation when nothing else moves. Raise $4.1M at a $24.3M post-money and the new investors own 16.9%; every existing holder keeps a proportional share of the rest. Two complications. Investors quote pre-money and you're diluted on post-money; our post-money valuation explainer walks the conversion. And four other things move in the same round:

• Option pool top-up. Most term sheets require the pool to be created or expanded before closing, inside the pre-money, so you pay for those shares, not the new investor. Carta calls it the option pool shuffle, and the next section puts numbers on it (Carta).
• SAFE conversion. Every post-money SAFE converts in the priced round at a percentage locked when the investor wired, so later SAFEs dilute you, not them. Mechanics in our SAFE note guide; the stacking math is below.
• Pro-rata rights. An existing investor's right to buy into the next round to keep their percentage. Not dilution itself, but pro-rata money is part of the round size and shrinks the allocation left for the new lead.
• Anti-dilution clauses. Raise at a lower price later and weighted-average anti-dilution (the common form) hands earlier investors extra shares; a full ratchet reprices their whole stake. With the down-round rate at 11.4% in Q1 2026 the clause bites less often, but it never bites investors (Carta).

The option pool shuffle, in numbers

The median seed-stage employee pool on Carta is 11.8%, and most companies don't use all of it before the next round (Carta, Peter Walker). Term sheets most often land in the same range, and the median gap from seed to Series A is about 20 months (Carta); the table below has the benchmarks. So a 20% ask at seed isn't a market term; it's a valuation discount dressed as generosity to your future hires. Counter with a hiring plan for those months, then ask for post-money sizing so the investor shares the dilution. Pool size also drives equity compensation per hire.

How is dilution calculated? A 4-round worked example

Work in post-money percentages. For each round: new investors get raise ÷ post-money; the option pool is reset to its target inside the pre-money; SAFEs convert at their caps on the pre-round capitalization; every existing holder is scaled by what's left. Run it four times on current medians and a founding team ends near 40% by Series B, with an 18% headline round costing 24% of founder equity.

Here is a static dilution calculator on Carta's medians. Assumptions: founders own everything at the start; the pre-seed is raised on post-money SAFEs at Carta's median cap for a round that size; seed, A and B are at the medians above; the pool is created at seed and refreshed at A and B, sized in the pre-money as term sheets require. Every input is in the table.

The seed row is where founders get surprised. The headline is what the new investor takes, but the pool and the SAFE holders, who convert into 9.5% of the post-seed table, land in the same round, so founders drop from 86.7% to 61.6%. That is a fall of 25 points. The headline said 18%. Series A repeats it: the pool top-up comes out of existing holders, so the headline round costs founders 24% of their equity. Rule of thumb: real dilution per round is the headline plus the pool top-up, then re-based.

Now compare the last two columns. The clean model leaves founders several points above Carta's median founding team after both seed and Series A (Founder Ownership Report). That gap of 5–11 points is the dilution nobody models: a second SAFE tranche, a bridge before the A, advisor shares, a departed co-founder's vested stock, a bigger pool than you planned for. Model it before outreach; the investor's associate will.

How do stacked SAFEs dilute founders?

Each post-money SAFE sells a fixed percentage: investment divided by its cap. Three SAFEs that each look small can add up to 20% of the company committed before a lead ever prices a round, and every later SAFE dilutes you, not the earlier holders. In our work, the most common Series A surprise is founders discovering the stack after the term sheet, not before it.

Y Combinator's post-money SAFE, the standard since 2018, makes the percentage explicit: ownership sold equals investment divided by the post-money cap, and YC's own examples are in the table below (Y Combinator). The catch is the word post-money. The cap is 'post' all the SAFE money you raise but not 'post' the priced round's new money or its new option pool, so each SAFE's percentage is protected against later SAFEs and diluted only by the priced round. Convenient for investors. Expensive for founders who sign several of them.

Take a founder who signs three SAFEs over 18 months, each a modest cheque on a rising cap. Together they've committed 20.1% of the company. At a median seed the SAFE holders convert into 14.3% of the post-seed cap table, and the founders are left at 56.8%: almost exactly Carta's median founding team after seed, without a single 'big' round. It's why we tell pre-seed founders to model the cap table after every SAFE. The stacks are getting taller: in Carta's latest pre-seed review, the larger rounds typically involve 10 or more instruments (Carta, State of Pre-Seed).

How do you give up less equity per round?

Six levers, in order of impact: raise what the next milestone needs rather than the maximum on offer; price the round on evidence, because valuation is the output of cash raised and dilution; size the option pool from a hiring plan and push for post-money sizing; sequence smaller rounds against milestones; use non-dilutive capital for what it fits; and accept more dilution only when it buys a bigger company.

1. Raise the right amount, not the max. Carta's read of its own benchmarks: valuations are mostly the output of cash raised and dilution (Carta). Work backwards from the milestone that unlocks the next round, add six months of buffer, and stop.
2. Anchor the valuation on numbers you can defend. In roughly 60% of the models we review, first-year revenue projections are too aggressive by a wide margin, and founder valuation expectations run just as far ahead of the market (table below). An overpriced round that stalls costs more dilution than a fair one that closes: the reset arrives with a lower cap and a bigger pool. Valuation methods and a model that ties out are the cheapest dilution insurance.
3. Negotiate the pool from a hiring plan. Bring the roles, the grant sizes and the months to the next round; counter a 20% ask with the low-teens pool the data supports; ask for post-money sizing. Every point of pool you don't need is founder equity.
4. Sequence smaller rounds against milestones. Selling a smaller slice now and another at double the price in 18 months raises more money per point of equity than selling the whole slice today, if you hit the milestone in between. Raise twice and miss once and you've done worse than either.
5. Use non-dilutive capital where it fits. Grants, R&D credits, revenue-based financing and venture debt fund specific line items without touching the cap table; hardware founders using non-dilutive debt before they raise now arrive at institutional rounds with less dilution than prior generations (Carta).
6. Know when more dilution is the right trade. A smaller slice of a company that reaches $200M is worth far more than a bigger slice of one that stalls at a fraction of that; the table below has the arithmetic. If the extra money buys the growth, the smaller slice wins. If it buys six more months of the same burn, it doesn't.

Can you raise capital without dilution?

Yes, for parts of the plan: government grants and tax credits, revenue-based financing repaid from a percentage of sales, venture debt alongside or after an equity round, and customer prepayments. None of it funds a pre-revenue company's losses; lenders and grant bodies pay for defined things, not runway. So: equity for the business model, non-dilutive money for equipment, R&D, working capital and inventory. Our non-dilutive funding guide ranks the options by stage, and venture capital vs venture debt covers when debt is a bridge and when it's a trap.

How much of the company should founders own at Series A?

The median founding team owns 36% after Series A, a little more in digital industries and less in physical ones (Carta Founder Ownership Report). By Series B the median is lower again, with AI teams keeping more than non-AI teams, and by Series C the employee pool is bigger than the founders' stake; the first table has each figure. A clean path through our bands lands founders near 47% after the A, and that's the number to aim for, because every later round takes its share of whatever you kept. Arrive at a Series A well below the median and the conversation shifts from your growth to your cap table.

What dilution mistakes cost founders the round?

Five we see repeatedly: the investor's cap-table model doesn't tie to the founder's, which kills the second meeting; no dilution scenarios prepared before outreach; a valuation built on first-year projections that are 2–3× too aggressive; counting only the new investor's percentage and forgetting the pool and the SAFEs; and re-raising after a failed attempt without re-anchoring the valuation or cleaning up the cap table.

• The second-meeting kill. The associate rebuilds your cap table from the data room and gets a different founder percentage than your deck; the partner reads it as sloppiness or spin. Across our recent client cohort, cap-table and dilution scenarios prepared before outreach prevented exactly this. Fix: one fully diluted table with every SAFE and side letter in it (cap table software helps), and the same numbers everywhere.
• Scenarios built after the term sheet. By then the lead has anchored the pool and the pre-money. Fix: pessimistic, expected and stretch valuations, each with pool and conversions, before the first meeting; it also sets your walk-away price.
• A valuation the model can't carry. First-year revenue in most of the models we review is far too aggressive (the levers table has the numbers), and the ask inherits the error; when the market reprices the round, the reset brings a larger pool and a lower cap, two hits for one mistake. Fix: a driver-based model that survives a stress test, then a valuation range built from it.
• Headline-only math. '18% dilution' in the deck, 24% in reality once the pool top-up lands. Investors don't correct you; they adjust their view of all your numbers. Fix: quote post-pool, post-conversion ownership, always.
• Re-raising on the old story. After a failed attempt, founders go back out with the same cap and cap table. Re-raises that work usually need a narrative reset, valuation re-anchoring, cap-table hygiene and a fresh investor list. Fix all four, or you're paying for the same no twice.

None of this needs a calculator you don't have. It needs one clean fully diluted cap table, three valuation scenarios, and the discipline to quote post-pool, post-conversion numbers before an investor does. Founders running raises solo typically take 9–12 months. The raises we run through our fundraising advisory closed 70% faster on median, and the cap table is one of the first things we tie out.