Startup funding - understanding rounds, dilution and terms#

If you are building a company, there are two different games you can play. I learned this late, and it matters before you take any money. The choice changes what fundraising means, what your cap table will look like, and how much control you keep.

1. The two kinds of companies#

Venture-scale company#

This kind is built to raise large rounds at high valuations. Investors put money in because they expect an acquisition or IPO that returns 10 to 300 times their investment. Everything in the company is tuned toward that outcome. The trade-off is direct. You give up control, ownership, and flexibility.

Durable profitable company#

This one makes money, pays the founders well, and grows at its own pace. It spends less than it makes, so it does not need outside investors. The founder keeps most of it. The trade-off is slower growth and a smaller market to chase.

The mistake I see most often is wanting both at once. You cannot take a $5M check from a fund that needs outlier returns and then decide to run a calm, profitable business at your own pace. Those two ideas pull in opposite directions.

2. The funding rounds#

Funding does not come all at once. It comes in stages, because no investor wants to pay for years of work before they have seen progress.

Pre-seed#

This is idea plus maybe a prototype. The goal is to build and get first users. About 92% of these rounds use SAFEs, not priced rounds. Size splits into two worlds. Small rounds are under $250K. Large rounds average about $1.4M. Cap is usually $10M to $15M. When you pitch here, you are selling who you are and what you are building.

Seed#

You have a working product and early traction. Median valuation hit a record $24M post-money in late 2025, up from about $18M the year before. AI startups see about 38% higher valuations at Series A, and the best AI seed rounds clear $40M. Pitch is product and early traction. One thing to know, the old $1M ARR rule for Series A is outdated. Most investors now look for $2M to $3M ARR.

Series A#

This is proven product-market fit and strong growth. The bar is now $2M to $3M ARR, not $1M. Exceptions happen with exceptional growth, a very strong technical team, or a competitive AI deal. Pitch is revenue, growth, unit economics, and how well you spend. If you pitch Series A like it is still pre-seed, investors pass.

Series B and beyond#

Revenue is in the tens of millions. The pitch is mostly numbers on the dashboard, story matters less. Median founding team ownership by Series B is about 23%. At this point the founder often does not fully control the company anymore.

3. How dilution actually works#

This is the part most founders misunderstand. It helps to see it with small numbers.

1. You and a co-founder start with 100 shares, 50 each.
2. An investor does not buy your shares. The company creates new shares.
3. The company prints 20 new shares and gives them to the investor.
4. Total is now 120. You still have 50, but 50 out of 120 is 41%.
5. You went from 50% to 41% without selling a single share.

The Facebook example makes the same point on a larger scale. In 2004 Peter Thiel put in $500K for about 10% of Facebook. The company created new shares for him, so the founders' percentages shrank. By IPO the original founders owned a much smaller fraction of their starting percentage, but the company was worth about $100B, so that fraction was still worth billions.

You are trading percentage for a chance to build something larger than you could fund alone. That is the deal.

4. Key term sheet terms#

Valuation gets the headline. Terms decide who controls the company when things get hard.

Liquidation preference#

This decides who gets paid first if the company sells.

Founder-friendly is 1x non-participating. Investors get their money back, and that is it.

The one to watch is participating. Investors get their money back and then take a share of what is left. Example, $10M raised, company sold for $30M. With participating preference, investors take $10M back first, then also take a share of the remaining $20M.

Protective provisions#

These give investors veto power over selling the company, raising another round, changing the budget, issuing new shares, or taking on debt. You can be CEO on paper and still need investor approval for these.

Pro rata rights#

Early investors can put more money into later rounds to keep their percentage. This can limit your ability to bring in new investors.

Drag-along rights#

If a majority of shareholders want to sell, the minority must sell too, even if the founder does not want to.

These terms are not written for good times. They are written for the down round, the bridge round, the missed milestone, the board fight, the month when payroll is due and the next check is not.

5. Runway and the clock#

Once money hits the bank, runway is the number to watch.

Runway is how many months you can keep operating before cash runs out. Founders who track it every week tend to survive. Founders who check once a quarter get surprised.

How much to raise has changed. The old rule was 18 months. In 2025 and 2026 the practical range is 24 to 30 months. Fundraising takes longer than it used to. Companies that raised Series A in late 2024 had waited about 2.1 years since the previous round on average.

There is also a hidden clock. Investors expect growth milestones in about 12 to 14 months. Miss them and you get difficult conversations, worse terms, a bridge round, or worse. The most common way venture-backed companies die is simple. They run out of money between rounds.

Selling a piece of your company for runway also sells a deadline. You cannot decide to slow down. You cannot be patient.

6. The 2025-2026 market split#

Since 2022 the market has split.

Regular startups have to prove more. Investors want paying customers, sustained growth, retention, and a clear path to the next milestone. Raising on an idea and figuring it out later does not work now.

AI startups are raising at levels that do not look normal. AI Series A valuations are about 38% higher than non-AI, AI valuations hit decade highs in 2025, and AI companies raise larger rounds at every stage from Series A onward.

And a third path has opened up. Building without outside funding is a real option in 2026 in a way it was not in 2022. AI has cut the cost of building software. Products that once needed ten engineers and a $1M seed can be built by one person with AI tools. Solo-founded companies went from about 25% of new startups in 2019 to over 33% in 2025 and 2026. One-person companies hitting $1M ARR actually exist now.

7. The no-funding alternative#

The economics have shifted.

What used to take ten engineers and a seed round can be done by one person with the right tools. Solo founders are now over a third of new companies. The phrase lifestyle business has been recast as durable profitable company, which is a better description anyway. There are also lean AI-native specialist funds that did not exist three years ago.

Not raising money is a real option in 2026 in a way it was not in 2022. If you can build without it, that option deserves a serious look.

8. Cautionary tale: Travis Kalanick and Uber#

Travis Kalanick built Uber. By IPO he owned less than 9%.

In 2017, major investors pushed for his resignation as CEO during months of controversy. He had built the company, but the investors he had taken money from years earlier had real say over who ran it.

This is not unusual. Stewart Butterfield owned about 8% of Slack at IPO. The pattern is consistent.

Once you take venture money, governance is not theoretical. Board seats, voting rights, and approval rights matter a lot when the company hits a rough patch. The terms are not written for when things are going well. They are written for those moments.

9. Key takeaways#

Concept	Key point
Two company types	Venture-scale needs 10-300x exits vs durable profitable that stays founder-owned
Dilution	Company creates new shares for investors, your percentage shrinks while share count stays the same
Valuation vs control	Headlines talk valuation, control is decided by liq pref and protective provisions
Runway	24-30 months is the new normal, track it weekly
2026 market	AI at a premium, regular startups need more proof, bootstrapping is viable
SAFEs	About 92% of pre-seed, they delay valuation, not avoid it
Series A bar	Moved from $1M ARR to $2-3M ARR
Kalanick lesson	Raising puts you on a deadline with outside governors
Uber and Slack founders	Owned under 10% at IPO despite building the companies

"Experienced founders don't just ask 'what valuation did I get?' They ask 'what did I give up to get it?'"