Seed Rounds Are Getting Bigger. The Odds of Reaching Series A Are Getting Worse.

The median seed round is now $3 million. Five years ago it was $1 million. The graduation rate to Series A has gone in the opposite direction.

That is the blunt summary of data published by Crunchbase in May 2026, and it should recalibrate how every angel investor thinks about portfolio construction, expected loss rates, and the actual odds of seeing a meaningful return from an early-stage check. The Crunchbase analysis, authored by Gene Teare and published May 26, 2026, documents a structural shift in the seed market that has been building since 2023 and accelerated sharply through the AI boom. The numbers are striking on their own. The implications for angels writing $25,000 to $100,000 checks are more striking still.

The Data: What Has Actually Changed

Start with round sizes. The median U.S. seed round stood at approximately $3 million in 2025, roughly three times its 2018 level, according to Crunchbase. The upper quartile median reached $5.6 million -- more than double the 2018 figure. Andy McLoughlin, managing partner at Uncork Capital, told Crunchbase that his firm now sees rounds ranging from $3 million to $5 million at the inception stage all the way to $8 million to $10 million and above. His own average check has nearly doubled in 18 months, from under $2.5 million to $4.5 million.

A separate Crunchbase report from April 2026 added context: in 2025, more than half of all seed dollars went into deals of $10 million or above. Roughly 350 deals fell in the $10 million to $50 million range, with another 20-plus at $50 million or higher. The largest seed round of 2025 was the $2 billion raised by Mira Murati's Thinking Machines Lab -- a number that would have been an enormous Series C a decade ago.

Carta's State of Private Markets data shows that median seed pre-money valuations hit $16 million in Q4 2024, the highest on record going back to at least 2016. By Q1 2025, with the number of seed rounds falling 28% year-over-year, that median had risen another 18%. Fewer deals, higher prices.

Now look at the graduation side. Through 2020, companies that raised a seed round of $1 million or more had a typical graduation rate to further funding or an exit of 55% or higher, per Crunchbase. That reflected a historical norm many angels had internalized: somewhere around one-third to one-half of seed-backed companies would eventually reach Series A.

Since 2021, those rates have fallen sharply. Of companies that raised a $1 million-plus seed round in 2023, only 24% have progressed further, according to the May 2026 Crunchbase report. For the 2024 cohort, the figure is 16%. Analysis from Incisive Ventures published in June 2025 corroborated these trends using Carta cap table data, finding that while Q1 2018 seeds saw 30.6% reach Series A within two years, Q1 2022 seeds saw just 15.4% do so in the same window. The conversion rate roughly halved in four years.

The time dimension compounds the problem. Since 2023, the median U.S. startup has been taking more than two years to raise a Series A after an initial seed. The ARR bar has risen accordingly. McLoughlin noted that the threshold for a successful Series A is no longer $1 million in annual recurring revenue. Investors now expect $2 million to $4 million in ARR. That is a 2x to 4x increase in required traction to access what has also become a more expensive round -- the median U.S. Series A reached $15 million in 2025, with an upper quartile at $25 million.

Why Seeds Inflated

Three structural forces drove this inflation.

The first is the disappearance of the line between pre-seed and seed. What was called a seed round in 2018 -- a $500,000 to $1.5 million check from angels to prove a concept -- is now labeled pre-seed or not labeled at all. The "seed" designation has migrated upmarket. Carta's 2025 annual State of Pre-Seed report found that 95th percentile seed valuations reached $80.5 million, nearly three times higher than six years earlier. As Euclid Ventures noted in February 2026, the word "seed" is now applied to everything from $500,000 SAFEs to $15 million equity rounds.

The second force is the AI premium. Carta's 2025 in Review report documented that at every stage from Series A onward, AI startups raised larger rounds and commanded higher valuations than non-AI counterparts. At Series A, the median AI valuation was 38% higher than the median non-AI valuation. At Series E+, the AI premium reached 193%. That pricing pressure cascades backward. Mercedes Bent, formerly of Lightspeed Venture Partners, told Crunchbase that "seed today is basically what Series A was seven years ago."

The third force is capital concentration. Crunchbase data shows that in 2025, deals below $3 million for seed-stage startups declined roughly 20% year-over-year in both count and dollars. Growth occurred only in the $10 million and above tier. The Bay Area captured a third of all seed funding deals. A startup with a Stanford pedigree or a name-brand lead investor raises at terms that would have been aggressive Series A terms in 2019. Everyone else competes in a different market with materially worse odds.

The Series A Cliff

McLoughlin put the competitive dynamic plainly in the Crunchbase piece: when a company goes to raise a Series A, it is not competing against its direct competitors. It is competing against every deal floating in the venture ecosystem simultaneously -- the quality of every team, the growth trajectory of every other company those partners are watching.

Katie Stanton, founder of seed fund Moxxie Ventures, described it as a binary market: you are either an AI elite team growing fast enough to raise from one of the big firms, or you are everybody else.

For everybody else, the math is brutal. Carta's Q4 2024 data showed the median startup raising a Series A waited 774 days -- more than two years -- from its previous primary funding round. The 75th percentile wait exceeded three years. Nearly 40% of all seed-stage financings in 2024 were bridge rounds, meaning a large portion of capital labeled "seed" on cap tables represents companies scrambling for time before a Series A they may never close.

What happens to companies that raise large seeds but cannot convert? The case of Yupp is instructive. The company raised $33 million at seed in 2024, backed by prominent investors including Google DeepMind's Jeff Dean and Twitter co-founder Biz Stone. It had 1.3 million users and several AI labs as paying customers. It shut down in March 2026, nine months after closing its round. Size of check is not a proxy for quality of outcome.

What This Means for Angels: Dilution, Timeline, and Loss Rate

Dilution exposure is rising. When the median seed round is $3 million and the median Series A is $15 million, an angel writing a $50,000 check at seed watches their ownership compress significantly before the company reaches the stage where institutional value creation accelerates. Carta data shows median seed dilution held around 20% per round through 2024. If a company raises a seed, a seed extension, a Series A, and a Series B before a buyer appears -- a timeline that could span six to eight years -- an early angel may hold a fraction of their original percentage.

Timeline expectations need to lengthen. The old model of seed-to-Series A in 18 months is gone. The current median exceeds two years, and the 75th percentile is over three. That affects not just when you might see a markup, but how long capital is illiquid and how many other opportunities you pass on while waiting. Angels who underwrite a two-year mark-to-market cycle will be systematically wrong.

Expected loss rates are higher than most models assume. If only 16% to 24% of recent seed cohorts have progressed further, the implicit loss rate on the remainder is very high. AngelList's Fund Benchmarks Report 2025 shows that 2021-2023 vintage seed funds track near 3% median IRR -- a stark break from the 15%-plus medians of the 2017-2020 vintages. Angels who wrote checks in 2021 and 2022 should not assume paper value reflects what those positions will ultimately return.

The Honest Math on Portfolio Construction

Traditional guidance for angel portfolio construction -- built during an era when graduation rates ran 30% to 40% -- said an angel needed at least 20 to 25 investments to have a reasonable probability of capturing one outlier that returned the portfolio. That logic is still correct. But the inputs have changed.

If graduation rates to Series A have fallen from 30-40% to 16-24%, and if the time to meaningful outcome has stretched from four to six years to potentially eight to ten, then the diversification requirement goes up, the reserve requirement for follow-on goes up, and the patience requirement goes up. An angel planning to write 15 checks and call it done is likely to be disappointed by the math.

There is also a selection problem the averages obscure. The 16% to 24% graduation rates are averages across all seed-backed companies. Companies backed by name-brand lead investors, headquartered in the Bay Area, operating in AI, and led by repeat founders are graduating at rates well above average. Companies backed by angel syndicates in secondary markets, outside the AI narrative, and led by first-time founders are graduating at rates well below it. If your deal flow primarily comes from the latter category, your effective graduation rate may be closer to 10% than 20%.

A Framework for 2026

First, get honest about check size relative to round size. If a company raises $5 million at a $20 million post-money valuation and you write a $25,000 check, you own 0.125%. Even a $100 million acquisition returns $125,000 before dilution and taxes. Either write larger checks into fewer deals or accept that you are building optionality and network access, not financial returns, from small positions in large rounds.

Second, reserve capital for follow-on. Companies you believe in will come back for extensions. If you have no reserves, you will be diluted without the ability to participate, watching the ownership you paid for shrink precisely when the company most needs your conviction.

Third, calibrate portfolio size to actual math. If your expected graduation rate is 20% and you need at least one company to reach Series A and beyond, you need at least 20 companies for a coin-flip chance at that outcome. If your effective rate is 10%, you need 40 or more. Most angels I talk to hold between five and twelve investments. That is not a portfolio; it is a collection of lottery tickets with no statistical basis for expecting a return.

Fourth, think carefully about AI exposure. Not because AI is guaranteed to win, but because the market is currently bifurcated in a way that makes non-AI seed deals structurally harder to finance at Series A. A company without an AI narrative is competing for a smaller pool of institutional attention than it would have commanded in 2020. That is not permanent, but it is real right now.

The data is not telling us that seed investing no longer works. It is telling us the game has changed in ways that require updated assumptions. Bigger rounds, lower graduation rates, longer timelines, and a more bifurcated market demand a more deliberate approach. The angels who treat 2026 like 2019 are going to be surprised by the results.

---

This article is for informational and educational purposes only. Nothing in this article constitutes investment advice or a recommendation to buy or sell any security. Past performance of venture asset classes does not predict future results. Angel investing involves substantial risk of loss, including loss of the entire amount invested. Consult a qualified financial advisor before making investment decisions.

Angel Investors Network is not a licensed broker-dealer, investment advisor, or funding portal. Content published on this site is editorial in nature and does not constitute an offer to sell or a solicitation to buy securities. Any investment involves risk and is suitable only for investors who can afford to lose their entire investment.

Author Disclosure: Jeff Barnes, MBA has no personal position in any company, fund, or platform named in this article. Angel Investors Network has no current commercial relationship with any party mentioned. AIN provides marketing and education services, not investment advice. Past performance does not guarantee future results. All investments involve risk, including loss of principal.