How much should a seed round raise?

Enough to reach the milestones a Series A investor will pay for, plus the months the next raise takes and a buffer. Many founders plan for 18 to 24 months of runway. For reference, Carta reported a median seed of 4.5 million at a 23.9 million post-money valuation in Q2 2026. Your milestones, not the median, should set the number.

How much dilution is normal in a seed round?

Carta reported median seed dilution of 19.4 percent in Q2 2026. Its analysis of 2025 priced rounds put the software median at 20 percent and deep tech at 21 percent, with fewer than 10 percent of software seed rounds selling 30 percent or more. Count the option pool and converting SAFEs when you calculate your own number.

What seed round terms should founders avoid?

Liquidation preferences above 1x or with participation, a board the founders no longer control, protective provisions that reach into budgets or hiring, and large pre-closing option pool increases. Also watch side letters that stack pro rata and most favoured nation rights. Any one can cost more than a difference in valuation. Have counsel review terms before signing.

What makes a seed investor a good one?

Reserves to participate in your next round, a record of supporting companies through slow periods, quick decisions, and help that portfolio founders confirm exists. Ask founders they have backed how the investor behaved in a bad quarter. A lead that will not support the Series A can send a negative signal to the investors you approach next.

What Is a Good Seed Round?

A good seed round raises enough to hit the milestones that make a Series A raisable, usually 18 to 24 months of runway, sells near typical dilution (Carta's Q2 2026 median was 19.4 percent), uses clean standard terms, and comes from investors able and willing to support the next round.

The market reference points

Benchmarks are a starting point, not a target. The most widely used are Carta's, drawn from US companies that manage their cap tables on its platform.

Size and price (Carta, Q2 2026): the median seed round raised 4.5 million at a 23.9 million post-money valuation.
Dilution (Carta, Q2 2026): median seed dilution was 19.4 percent.
By sector (Carta, January 2026, 2025 priced rounds): median seed dilution was 20 percent in software and 21 percent in deep tech, and fewer than 10 percent of software seed rounds sold 30 percent or more of the company.
The tail (Carta, Q2 2026): the 95th percentile seed valuation reached about 200 million, against 72.2 million a year earlier.

Two cautions. Medians come from different distributions and do not combine neatly: 4.5 million divided by 23.9 million is 18.8 percent, not the 19.4 percent median dilution. And the tail is concentrated in AI companies with unusual demand. A round that looks small next to the headline numbers can still be good for the company raising it.

Size the round against milestones, not the median

A good seed round is the amount that reaches the evidence a Series A investor will pay for, plus room for the raise to take longer than planned.

Work backwards. Name the milestones: revenue level, retention, a repeatable acquisition channel, the team that delivers them. Estimate monthly net burn to get there. Multiply by the months required, then add the months a Series A process takes and a buffer for things going slower than plan. Many founders plan for 18 to 24 months of runway from a seed for this reason.

Illustrative: milestones need 16 months at an average net burn of 220,000, the raise needs 5 months, and a 3 month buffer is prudent. That is 24 months, or about 5.3 million. If the company can raise only 3.5 million, it needs to cut the plan or the burn now, not discover the gap in month 14.

Raising more than the plan needs is not free either. Extra money at the same price is extra dilution, and extra money at a higher price raises the bar for the next round.

Dilution: why both ends hurt

Selling too little. Very low dilution usually means a very high valuation. That commits the company to growing into the price before the next round. If it does not, the next round is flat or down, and down rounds bring anti-dilution adjustments and renegotiated terms.

Selling too much. Selling 30 percent or more at seed is uncommon in software, per the Carta data above. It leaves less room for Series A, the option pool, and later rounds, and it can leave founders with ownership that later investors view as too low to keep them motivated.

Watch the pool. An option pool created or expanded before closing comes out of existing holders' ownership, not the new investor's. A round with a lower headline valuation and a smaller pre-closing pool can leave founders with more than a higher headline with a large one. Calculate founder ownership after the round, including the pool and any SAFEs that convert, before comparing offers.

Terms that make a round good or bad

Price gets the attention. Terms outlast it.

Liquidation preference. A 1x non-participating preference is the market norm at seed. Multiples above 1x, or participation, reduce what common holders receive in modest exits.

Board. A seed board is often two founders and one investor, or founders only on a SAFE round. Watch for a board the founders no longer control.

Protective provisions. Standard lists cover new senior securities, sale of the company, and changes to the investors' rights. Unusual additions, such as approval over budgets or hiring, are operating control.

Instrument. A round on standard post-money SAFEs is fast and simple. Check side letters for pro rata rights, most favoured nation clauses, and information rights, because they accumulate.

This is general information, not legal advice. Preference, board, and protective provision terms are negotiated in specific documents, so have counsel review the term sheet before you sign.

Investor quality

The same money from different investors is a different round. A good seed investor has reserves to participate in the next round, a record of supporting companies when progress stalls, and partners who answer quickly. Ask founders they have backed how they behaved in a bad quarter.

Watch signalling. If a large multi-stage fund invests at seed and then declines to lead or join the Series A, other investors notice. That does not rule such funds out. It means the relationship needs to be understood before the cheque clears.

A test for any seed offer

A seed round is good if you can answer yes to each of these:

Does the money reach the Series A milestones with at least a few months to spare?
Is founder ownership after the round, the pool, and any conversions at a level you can live with through two more rounds?
Is the valuation one the milestones can justify at the next raise?
Are the terms close to standard, with no unusual preference, board, or approval rights?
Will the lead investor support the next round, and do their portfolio founders say so?

Frequently asked questions

How much should a seed round raise?: Enough to reach the milestones a Series A investor will pay for, plus the months the next raise takes and a buffer. Many founders plan for 18 to 24 months of runway. For reference, Carta reported a median seed of 4.5 million at a 23.9 million post-money valuation in Q2 2026. Your milestones, not the median, should set the number.
How much dilution is normal in a seed round?: Carta reported median seed dilution of 19.4 percent in Q2 2026. Its analysis of 2025 priced rounds put the software median at 20 percent and deep tech at 21 percent, with fewer than 10 percent of software seed rounds selling 30 percent or more. Count the option pool and converting SAFEs when you calculate your own number.
What seed round terms should founders avoid?: Liquidation preferences above 1x or with participation, a board the founders no longer control, protective provisions that reach into budgets or hiring, and large pre-closing option pool increases. Also watch side letters that stack pro rata and most favoured nation rights. Any one can cost more than a difference in valuation. Have counsel review terms before signing.
What makes a seed investor a good one?: Reserves to participate in your next round, a record of supporting companies through slow periods, quick decisions, and help that portfolio founders confirm exists. Ask founders they have backed how the investor behaved in a bad quarter. A lead that will not support the Series A can send a negative signal to the investors you approach next.