What Seed Funding Is and How Early Startups Raise It

Seed funding is the first outside check a startup cashes, not the first dollar that ever touches the business. Pre-seed, seed, and Series A each answer a different question, and mixing them up is how founders end up pitching the wrong investor with the wrong ask at the wrong time.

Pre-seed gets the team and the idea into shape. Seed pays for proof: does anyone actually want this thing, and can the team show enough evidence to convince a Series A investor to write a much bigger check? Series A then scales whatever got proven. Confusing what each round is for is the single most common way founders waste a fundraise, and it's an entirely avoidable one.

Carta puts the median seed round at $2.5 to $3.2 million, with founders giving up 12 to 15% of the company. Median post-money seed valuation hit $24 million in the fourth quarter of 2025. None of that is a target to hit. It's the water level, and everything below is about what pushes a round above or below it.

More than half of all seed dollars in 2025 went into deals of $10 million or more, according to Crunchbase. That single fact explains almost everything else happening in this market: bigger checks, fewer of them, and a much higher bar for who gets one. Deal count dropped since the 2021 to 2022 peak even as check size climbed, and that split the market into two lanes. A small number of startups raise large seed rounds backed by real traction. Everyone else fights over smaller checks with a lot less certainty attached. There's no third lane forming to catch them, and pretending otherwise just wastes a founder's runway on the wrong pitch.

The gap to Series A has widened to match, and this is the part most founders still haven't priced in. Only 24 to 27% of seed-funded startups that raised $1 million or more in 2023 and 2024 made it to a Series A. Investors at that stage now want $2 million to $4 million in annual recurring revenue, not the $1 million bar that used to clear the room, and the median time to hit that number after seed runs 774 days. The average gap between seed and Series A stretched past two years in 2024, up from 1.7 years in 2019. That's exactly why bridge rounds stopped being an occasional patch job and became their own category: in the third quarter, 40% of all seed and Series A rounds raised were bridge rounds, near a historic high, per Carta. That's not a market catching its breath. It's the runway between rounds getting longer than most founders' budgets ever accounted for.

AI has carved its own track inside all of this, and it's not a minor subplot. Funding to AI companies doubled in the second quarter of 2024, and Carta's data on 775 primary Series A rounds shows AI companies and non-AI SaaS companies sitting on entirely separate valuation curves. "AI-native" is shorthand investors use to sort decks before they even open them.

Seed isn't harder to raise because investors have less money sitting around. It's harder because they expect proof before they sign, and that proof bar moved up a full flight of stairs in three years.

Who writes seed checks and what each type of investor wants

Four kinds of investor write seed checks, and each one is grading a different exam. Confuse them, and the pitch you give one is the pitch that should've gone to another.

Angel investors typically write checks ranging from $25,000 to $750,000 out of their own pocket, often as former operators or sector specialists who move fast because there's no investment committee standing between their gut and the wire transfer. Micro-VCs run smaller funds than institutional players; they bring more process than an angel and far less than a multi-stage fund, and often lead rounds under $3 million. Multi-stage funds (the Sequoias, the Greylocks, the Lightspeeds of the world) write bigger checks at higher valuations and bring brand plus follow-on capital, but they run heavier diligence before committing to anything. Accelerators sit apart from all three: Y Combinator writes a standard check paired with Demo Day exposure to hundreds of investors at once. Techstars runs a similar model built around network access and hands-on mentorship.

Beyond those four sit AngelList syndicates and angel groups (useful when a specific network matters more than a specific check size), corporate venture arms active in particular industries, and the alumni capital that circles YC's Demo Day like moths around a very well-funded porch light.

Angels move on story and domain credibility before there's any traction to point to. Micro-VCs want early signals, a team that looks like it can execute, and a market big enough to return the fund. Multi-stage funds want the same three things at a much higher bar, since they're anchoring a seed round they plan to follow into Series A themselves. Accelerators bet on coachability, often before the product even exists.

Matching investor type to stage isn't optional polish. It's the whole strategy. Pitching a multi-stage fund with nothing but a founder story and a slide deck wastes an afternoon nobody gets back, and it burns a relationship that might've actually worked at the next stage.

The SAFE note: why it became the default early-stage instrument

A SAFE is a contract, not a share of stock. It gives an investor the right to shares later, converting automatically into preferred stock once the company raises a priced round.

Y Combinator built the SAFE in 2013, and it now dominates early-stage deals so thoroughly that calling it "popular" undersells it: 89% of all pre-priced investments in the third quarter of 2024 were SAFEs. That's not a trend. That's a monopoly.

The reason is mechanical, not fashionable, and it's worth understanding exactly why. A convertible note is debt: it carries an interest rate and a maturity date, and if the priced round takes longer to show up than planned, that clock keeps running whether the company is ready or not. A SAFE has neither. No interest, no maturity date, no repayment obligation hanging over anyone's head. Fewer terms to fight over, standardized paperwork that keeps legal bills down, and founders keep full control since SAFE investors don't get board seats or veto rights just for signing.

SAFEs also let founders run rolling closes: taking money from different angels across several weeks without reopening the whole negotiation every time a new check lands. That matters because most seed rounds get built angel by angel, not signed all at once around one table. And since early-stage companies rarely have the data to defend a real valuation, a SAFE just postpones that argument until a priced round exists to ground it in actual numbers instead of guesses.

The terms inside a SAFE that determine what founders actually give up

Two levers decide what a SAFE investor ends up owning: the valuation cap and the discount. At conversion, the investor takes whichever gives them the better price, full stop. Founders who don't model both before signing are negotiating blind.

The cap sets a ceiling. If the next priced round comes in above it, the SAFE investor still converts as if the company were only worth the cap amount, which hands them more shares for the same dollars. The discount works differently: a straight percentage off what new investors pay at that next round. A 20% discount means the SAFE holder buys in at 80 cents on the dollar relative to the fresh money.

Post-money SAFEs with a cap are now the default. In the first half of 2025, 96% of SAFEs issued included a cap, up from 86% in 2024, and roughly 81% carried only a cap with no discount attached.

Here's the part founders routinely get wrong, and it costs them more equity than any other mistake in this whole process: the post-money structure dilutes founders the moment a new SAFE gets signed, not when the priced round eventually happens. Post-money ownership math folds in that new SAFE immediately. Stack three or four SAFEs across a raise and cumulative dilution can land well past what any single term sheet suggested on its own. Model that math before signing the next SAFE, not after realizing ownership shrank by more than expected.

The cap does quiet work later too. A low cap set early can anchor Series A conversations in ways that help or hurt depending on how the company's value moved since. What it should reflect has nothing to do with hope: where the company actually stands right now, what comparable deals are pricing at, and how much runway this specific check needs to buy.

When a priced seed round makes more sense than a SAFE

Roughly half of seed deals at $3 million or more in 2025 were priced rounds instead of SAFEs. Below that line, SAFEs still run the table. That split isn't a coincidence, and founders raising above $4 million who still default to a SAFE out of habit are usually making a mistake.

A priced seed round issues Series Seed Preferred Stock: actual equity changes hands at a negotiated valuation, with ownership defined from day one instead of deferred to some future conversion event. Size is the biggest factor pushing a round this direction. At $4 million and up, investors want clean, immediate ownership rather than SAFE conversion math that gets messier every time another SAFE stacks on top. A lead investor demanding a board seat is another trigger, since a SAFE simply cannot grant one. And some founders just want a clean cap table walking into Series A instead of a pile of SAFEs converting all at once and creating a mess someone has to untangle by hand.

The trade-off is real, and it should be stated plainly instead of hedged: priced rounds cost more in legal fees, take longer to close, and force an actual valuation conversation instead of postponing it. What they buy back is clarity, and clarity gets more valuable as the check size grows. That's the whole story behind the shift: as seed rounds get bigger, the instrument used to raise them moves upmarket right along with them.

What investors are actually evaluating at seed

Team sits at the top of nearly every seed checklist, and for one blunt reason: products change, markets shift, but the people running the company usually don't.

Two-person founding teams make up 36% of new startup formations and raise a disproportionate share of the money, while solo founders pulled in just 20% of rounds last year. Carta's numbers show teams with more than one founder outperform solo founders by 163% and carry meaningfully higher valuations on average. Investors aren't counting heads for the sake of it. They're checking for complementary skills, a working relationship that predates the pitch, and a shared appetite for risk that hasn't been tested yet by an actual crisis.

Beyond team, investors test whether the problem is real and big enough to matter, whether the market can support a venture-scale return, and whether a working product proves the team can build, not just talk about building. Traction, where it exists, carries real weight, whether that's user engagement, early revenue, or a waitlist that's actually growing week over week. How much traction matters shifts by investor type and round size. At pre-seed, team and idea alone can close a check; by seed, some proof the market is responding has become close to mandatory.

AI has become its own filter inside all of this. It comes up in nearly every seed pitch meeting now, and the data shows AI-enabled startups commanding distinct, often higher, valuations at comparable stages to everyone else in the room.

Building the MVP that earns a seed check

An MVP is the smallest version of a product that lets a real user finish a real task. Not a prototype, not a demo reel. A working thing that does one job all the way through, start to finish.

The word "viable" is the part founders skip past too fast. The product has to let someone complete an entire task and walk away with something that feels finished, even if it's narrow in scope. A half-built interface with three buttons that don't do anything yet doesn't clear that bar, no matter how polished the deck looks around it.

Strategically, the MVP's whole job is generating feedback that turns a pitch into evidence. Seed investors aren't funding a slide about a concept; they're funding a team's ability to learn fast and change direction when the data says to. A study in the Journal of Business Venturing Insights found startups using MVPs were 2.3 times more likely to pivot successfully after gathering customer feedback than startups that launched a fully built product first. That gap matters directly, because the ability to pivot is exactly what a seed check is supposed to buy.

History backs this up consistently: the companies that scaled fastest typically started with the smallest version of the thing they were trying to prove, not the finished version they eventually built.

The mistakes that cost founders credibility at seed are consistent and avoidable. Ignoring early user feedback is the biggest one, since at this stage the feedback basically is the product. Building on a foundation that can't scale is another; when an investor asks how the MVP becomes the real platform, "we'll rebuild it later" isn't an answer, it's a red flag with a bow on it. Treating the MVP as a finished product instead of a hypothesis test misses the entire reason to build one at all. In practice, the handoff looks like this: MVP validation justifies the pre-seed raise, and by seed, investors expect a precise account of what that MVP taught the team and how the product changed because of it.

Finding and structuring the co-founder relationship before raising

Solo founders and two-founder teams each represent a significant share of new startups, but the fundraising market has already made its preference plain: two-founder teams raise far more successfully, and the data is clean enough that it isn't really a debate anymore.

Investors look for a specific pattern in a founding pair. Complementary skills top the list, usually one technical co-founder paired with one commercial one, though the exact split varies by business. A shared work history matters almost as much, since it lowers the odds the team cracks under the grind between seed and Series A. And investors want the hard conversations already settled before the pitch ever happens: who owns what, who decides what, and how the equity got split in the first place.

That split has moved meaningfully over the past decade. Equal equity splits between two-founder teams have become increasingly common, according to Carta. Investors ask about the split directly, and an arrangement that's still unresolved, or quietly contested behind the scenes, reads as a warning sign rather than a minor detail to sort out after the wire hits.

For founders building alone and hunting for a technical or commercial partner, YC's co-founder matching platform is among the most widely used resources of its kind for solving this problem before anyone sits down in front of an investor and gets asked who does what.