Series A Fundraising Data: What Carta's Q3 2025 Shows

Headlines scream record-breaking AI rounds, but Carta's data tells a different story. Seed deal volume is falling, the median Series A now closes nearly three years after the seed, and founders in the bottom valuation quartile graduate at half the rate of everyone else.

Key Facts

• Median Series A pre-money valuation: $50M pre-money, ~$60M post (Peter Walker)
• Median seed valuation: $16M pre-money, ~$20M post (Peter Walker)
• Median time from seed to Series A: Just over 2 years (24+ months) (Peter Walker)
• Seed-to-A graduation rate (within 2 years): 20% of seed companies (Peter Walker)
• Founder secondaries at seed: Twice as common in 2025 as in 2021 (Peter Walker)

Seed Valuations and the Graduation Gap You Need to Know

Median seed stage in 2025 is $16M pre-money with ~$3.5M raised, putting the post-money around $20M — the most expensive in Carta's history. But raising too little is the real danger: bottom-quartile seed companies graduate to Series A at half the rate of all other cohorts.

The headline number for seed is striking: a $16M pre-money valuation and roughly $3.5M raised puts the typical post-money around $20M. Peter Walker called this 'historically very expensive — as expensive as it's ever been.' But the more actionable insight lives in the quartile breakdown.

Walker and his team sliced seed-stage companies into valuation quartiles and tracked which ones reached Series A within 24 months. The top three quartiles performed similarly — within a few percentage points of each other. The bottom quartile was the outlier. 'Those companies made it to Series A about half as frequently as the others,' Walker said. The implication for founders: raising too little at seed is a meaningful structural disadvantage, not just a short-term cash problem.

Pablo Srugo raised the causation question — do low-quartile rounds fail because they're underfunded, or because underfunding is itself a signal of weaker companies? Walker acknowledged both are likely true, but the data point still holds: if you have the option to raise more at seed, the graduation data suggests you should.

"The companies that had a big change were the lowest valued companies. If you're raising a Seed round in the zero to 25th percentile of valuation, those companies made it to Series A about half as frequently as the others." — Peter Walker

"Seed stage is about $16 million pre-money and $3.5 million or so raised in the round. That is historically very expensive — as expensive as it's ever been, really." — Peter Walker

• Median seed: $16M pre-money, $3.5M raised, ~$20M post
• Top three quartiles graduate to Series A at similar rates
• Bottom quartile graduates at roughly half the rate
• Seed deal volume on Carta is declining despite rising capital

How Long It Actually Takes to Raise a Series A Right Now

The median time from seed to Series A is just over two years in 2025. Only 20% of seed companies hit that milestone within 24 months, down sharply from 40% at the 2021 peak. The A-to-B gap is even longer — close to 1,000 days, or nearly three years.

One of the clearest signals from Carta's Q3 data is that the time between rounds is stretching — every stage, every quarter. Walker put it plainly: 'Every quarter, like clockwork, the time between these early-stage rounds is getting longer.' Seed to A now sits just above the two-year mark. A to B is close to a thousand days — roughly two years and nine months.

The graduation rate compounds this. At the 2021 fundraising peak, 40% of seed companies made it to Series A within two years. That number has fallen to 20%. Walker noted this creates a tale of two cities: high-growth AI darlings like Ramp raise multiple rounds in a single calendar year, while the median company grinds through a much longer, quieter path.

For founders, the practical implication is clear: plan your runway accordingly. Walker advised, 'Pick your investors wisely — it's going to be a long road.' Bridge rounds are increasingly common, and reserved follow-on capital from seed investors is becoming a standard part of fund construction rather than an exception.

"Every quarter, like clockwork, the time between these early stage rounds is getting longer. Seed to A is just above two years. A to B is close to a thousand days." — Peter Walker

"It used to be forty percent of companies did Seed to A in two years or less. Now it's twenty percent." — Peter Walker

Team Size and Compensation: What's Changed at Series A

The median Series A company in 2025 has roughly 17 full-time employees — down from 25 in 2021 and still falling. Salaries aren't keeping pace with inflation for most roles, and early employee equity grants have slightly decreased. The exception: AI/ML engineers, whose pay rose 25–40% in 18 months.

Smaller teams are one of the most concrete structural shifts in early-stage startups. Walker noted that on the day of a Series A close, the typical company now has about 17 employees, trending toward 15 by year-end. In 2021, the same milestone came with 25. Round sizes are going up while headcounts are going down — which means capital is flowing to compute, distribution, or simply sitting as runway.

Walker pointed to a cultural shift catalyzed by an unlikely source: 'Underratedly, the Elon example is a big deal. Looking at how X is actually running okay with fifteen percent as many people as they had before is a big shock to a lot of folks.' CEOs across the spectrum revisited their headcount assumptions and founders are now genuinely reluctant to hire.

For employees, the picture is less optimistic. Early-stage equity grants have edged down, not up. Salaries are rising but lagging inflation — except for AI/ML engineers, whose total compensation jumped 25–40% over 18 months on Carta's platform. Walker's conclusion: 'Your average data scientist, your average marketer — it's a tough environment right now trying to get hired at startups.'

"Series A, 2021, you had twenty-five people working for the company. Now you probably have seventeen, sixteen. And I think by end of the year that'll be more like fifteen." — Peter Walker

"AI, ML engineers over the last eighteen months on Carta — their pay packages have risen between twenty-five and forty percent. Over eighteen months, that's unheard of." — Peter Walker

• Series A team size: ~17 today vs. 25 in 2021
• Non-AI salaries rising but below inflation
• AI/ML engineer pay up 25–40% in 18 months
• Early employee equity grants have slightly decreased

Competing in Crowded AI Categories: How Investors Think About It

In a market where every obvious AI use case attracts dozens of competitors — including foundation models — the differentiating edge has shifted from 'is this real value?' to 'can you go to market and iterate faster than anyone else?' Vertical expertise, prior relationships, and existing customers matter more than ever.

Pre-ChatGPT, the primary question for seed investors was simple: is this a must-have or a nice-to-have? Today, that question is largely answered before the pitch ends — AI creates obvious, demonstrable value in almost every vertical. The harder question is competition. Srugo framed it: 'Competition has risen to the top because you've got big tech, you've got incumbents, you've got new entrants, and you've got the foundation models.'

The foundation model risk is a new and specific threat. Unlike traditional competitive dynamics where a two-year head start creates defensible moat, AI capabilities can render a startup's core advantage obsolete overnight when a new model ships. Srugo noted passing on companies 'where I don't see how you maintain that edge' when there aren't enough workflows, integrations, or domain-specific business logic to survive a foundation model update.

The playbook that appears to work: deep vertical focus combined with founder domain expertise. Srugo cited Decagon — an AI customer support company founded by ex-Palantir operators who'd sold to enterprise before — as an example of founders whose go-to-market speed and existing relationships created a durable edge. Getting a customer into production, even in AI, still makes them meaningfully harder to displace.

"Competition has risen to the top because you've got big tech, you've got incumbents, you've got new entrants, and you've got the foundation models. You've just got a lot of people that you have to worry about." — Pablo Srugo

"Will Google build this is now, will Gemini build this, or will OpenAI build this? It's a unique foundation model risk." — Peter Walker

Founder Secondaries at Seed: When They Make Sense

Founder secondary transactions at seed stage are twice as common in 2025 as they were at the 2021 peak. Srugo's framework: a quarter to half a million in secondary at seed is often worth encouraging. At Series A, taking $1M off a $10M raise is reasonable. Beyond $10M, the company should be on a clear trajectory to justify it.

Walker flagged the secondary data as one of the most surprising signals in recent quarters: 'The number of founders taking secondary after their Seed round is twice as high this year as it was in 2021, which was the highest we'd ever seen.' The trend raises questions about founder motivation, risk alignment, and what it signals about the current market.

Srugo's take is pragmatic. Economic pressure that goes unaddressed doesn't make founders more committed — it just consumes mental bandwidth. 'If your startup doesn't work out, no one cares that you've paid yourself little,' he said. Taking modest liquidity at seed or A to cover basic life expenses moves the founder closer to the investor's risk profile — they still have one company, but they're not betting their personal finances on the outcome.

The threshold matters. A $250K–$500K secondary at seed: reasonable. $1M at Series A on a $10M raise: fine. Beyond $10M in secondary before the company is clearly on a breakout trajectory: a signal worth scrutinizing. Walker flagged one area of concern: almost none of these seed-stage secondaries include employee participation — a gap that's widening as the late-stage market develops more employee liquidity options through tender offers.

"The number of founders taking secondary after their Seed round is twice as high this year as it was in 2021, which was the highest we'd ever seen." — Peter Walker

"If you're raising a $10 million A and you haven't yet made money before, take a million off the table. Why not? It's only going to align us more." — Pablo Srugo

Median Valuations by Stage — Carta Q3 2025