How Much to Raise at Seed: The Number Sets Your Next Round's Bar

Ask ten founders how much they are raising and most will give you a number they did not calculate. It came from a friend's round, a pitch they admired at demo day, or a feeling about what the stage deserves.

Y Combinator's seed guide gives the arithmetic instead. Raise what it takes to reach profitability. When a follow-on round is coming, raise what it takes to reach the next fundable milestone, which is usually 12 to 18 months out.

Its rule of thumb: an engineer costs about $15,000 a month all-in, so 18 months with five engineers is $1.35 million. First rounds land between $500,000 and $1.5 million.

Now the market. Carta's Q2 2026 benchmarks put the median seed round at $4.5 million at a $23.9 million post-money valuation, with median dilution at 19.4%. That median is roughly three times the top of YC's range. Both numbers are honest. Only one of them describes your company.

Raise size, valuation and dilution are one decision

Set any two and the third falls out, whether or not you do the arithmetic.

Raise $1.5 million at an $8 million pre-money valuation and you have sold about 16% of the company. Raise $2 million at a $6 million pre-money and you have sold 25%. Same stage, same pitch, one sentence apart, and twice the cost.

A $3 million round when the plan needed $1.5 million roughly doubles what founders give up: 25% against 18.75%. Dilution compounds too. Carta's cap table data has a founder raising at 2024 median dilution keeping 40.28% of the company after a Series D, against 32.73% for the same founder at 2019 medians.

The seed sets the bar for the Series A

Your first round sets the number you have to hit to raise again.

A $5 million seed sets a $20 million to $30 million ARR bar for the Series A. A $1.5 million seed sets $8 million to $12 million. You are choosing the hurdle, and the hurdle is public.

What it costs to be wrong

The math founders run is the good case. What does $3 million buy that $300,000 does not? More engineers, more runway, more go-to-market. Every line assumes the thesis works.

Run the other case. $300,000 lasts a year. You stay close to customers because there is nobody else to do it.

You ship, you test, and the signal stays cold. Twelve months and $300,000 buys a clear answer, with the prototype, the story and most of your savings intact.

$3 million on the same unproven thesis lasts two or three years. You hire the team the deck promised and build the broad version of the product.

If the signal stays lukewarm, the same lesson arrives three years later at the price of a company. A bigger round raises the price of being wrong.

What actually kills companies

CB Insights read public post-mortems from 431 venture-backed companies that shut down since 2023. Running out of capital tops the list at 70% of the sample, and CB Insights calls it the final cause of death more often than the only one.

The risk that ends companies is a round that bought a plan the company couldn't make true before the money ran out.

The number follows the milestone

Write the milestone first. One sentence: what has to be true in 18 months for the next round to look obvious. Then price it. Engineers at $15,000 a month, a designer, the customer work, and a buffer for the parts you got wrong.

Carta's median says the market likes $4.5 million. The same market funded more than a third of recent pre-seed rounds under $250,000, and SAFEs of $1 million or more averaged $1.4 million in 2025. Under the median, the market pulls apart into the lean and the loaded.

The round is a decision you can read back later. Write the milestone, write the dilution it costs, and write what you would do if the answer comes back cold. That paragraph describes your company in a way no benchmark can.