Pre-seed funds the founder. Seed funds the hypothesis. Series A funds the machine.
These aren't just bigger checks as you go — they're completely different investments. A pre-seed investor is betting on whether you can build something people want. A Series A investor is evaluating whether you've built a system that can scale without breaking. The evidence required, the dilution expected, and the metrics that matter are entirely different at each stage.
Here's the complete breakdown of what each stage actually looks like in 2025, with real data on check sizes, valuations, and what gets funded.
The Numbers at Each Stage (2025 Data)
| Stage | Check Size | Pre-Money Val. | Dilution | Key Milestone |
|---|---|---|---|---|
| Pre-Seed | $500K–$2M | $3–8M | 10–20% | Compelling founder + market thesis |
| Seed | $2–4M | $10–18M | 18–25% | Early traction, signs of PMF |
| Series A | $10–18M | $40–65M | 20–28% | $1–3M ARR, repeatable growth engine |
Source: Carta, PitchBook, AngelList 2025 data. AI-native companies often raise at 1.5–2x these pre-money benchmarks due to team and market premiums.
Pre-Seed: Betting on the Founder
Pre-seed exists because good founders need capital before they have evidence. The median pre-seed in 2025 is $750K–$1.5M on a $4–6M post-money valuation, almost always structured as a SAFE with a valuation cap. The product may not exist. Users may not exist. Revenue definitely doesn't exist.
Pre-seed investors are underwriting three things: the founder's ability to execute, the market's size and timing, and the quality of the insight that others have missed. The best pre-seed decks spend 60% of their content on the problem and the insight — not on a product demo that will look completely different in six months anyway.
Who invests
Angels, pre-seed funds, scouts, accelerators (YC, Techstars, First Round Capital)
What matters most
Founder background, insight clarity, market size, unfair competitive advantage
How much to raise
12–18 months of runway to reach seed-fundable milestones — not more
Common vehicle
YC SAFE (standard), sometimes convertible note with discount
Seed: Proving Product-Market Fit
The seed round has gotten bigger and more structured over the last five years. What used to be a $1–2M friends-and-family round is now a $2–4M institutional round led by dedicated seed funds. A strong seed deck needs actual traction: active users, early revenue, or at minimum a waitlist with clear demand signals that are hard to dismiss.
Median seed valuations in 2025 sit at $10–18M pre-money for B2B SaaS. AI-native companies frequently raise at $20–35M pre-money even pre-revenue if the founding team has exceptional domain credentials or prior successful exits. The round is usually priced preferred equity or a SAFE with a higher cap than the pre-seed.
The single most common mistake at seed stage: treating the seed as the end goal instead of a bridge to Series A. Raise enough to reach the metrics your Series A investor needs to see — not just enough to keep the lights on for six months.
Series A: The Machine That Scales
Series A is where the evaluation framework changes completely. Investors at this stage aren't asking "can this work?" They're asking "does this already work, and can it scale without breaking?" The median Series A in 2025 is $12–15M on a $45–55M pre-money for B2B SaaS — down from the 2021 peak of $15–20M on $60–100M pre-money, but holding relatively stable since mid-2023.
What Series A investors specifically evaluate in 2025:
- →ARR: $1–3M for traditional SaaS; AI-native can raise earlier on strong team signals and fast growth
- →Net Revenue Retention: 100%+ minimum, 110–120% strongly preferred as a signal of product stickiness
- →CAC Payback: Under 18 months for capital-efficient businesses; under 12 months for top-tier rounds
- →Revenue growth: 3x year-over-year minimum; 5x is needed for the most competitive rounds
- →ICP clarity: Can you describe your ideal customer in one specific sentence, backed by conversion data?
The Conversion Math: What Actually Gets Through
The venture funnel is more brutal than most founders expect. Based on Carta data covering 2020–2024 cohorts:
~45%
Pre-Seed → Seed
Of companies that close a meaningful pre-seed, roughly half reach a subsequent seed round within 24 months
~35%
Seed → Series A
The most brutal filter — most companies exhaust runway or fail to reach repeatable metrics before A-round timing
~55%
Series A → B
The funnel widens once you've raised an A with real metrics — execution risk dominates over market risk
The implication: most companies that raise pre-seed will not raise a Series A. This isn't purely a failure of execution — it reflects how the funnel narrows as evidence requirements rise. The funding benchmarks at Value Add VC confirm this pattern holds consistently across 2019–2024 vintage years.
The Lines Are Blurring — Especially in AI
One important caveat for 2025: the traditional pre-seed → seed → Series A progression is less linear than it used to be. AI-native companies are raising rounds that structurally look like Series As — $10–20M on $60–100M pre-money — but are functionally pre-product bets on exceptional founding teams. Meanwhile, companies that would have raised Series As in 2019 are doing extended seed rounds at lower dilution to preserve optionality as interest rate normalization keeps LPs more selective.
The category matters enormously. Enterprise SaaS companies face traditional ARR milestones. Developer tools companies raise on MAUs and usage depth. AI foundation model companies raise on compute access, research talent, and benchmark performance. Don't benchmark your round against a company building in a different category — the frameworks differ substantially and you'll price yourself wrong.
The question every founder should answer before each raise:
What specific, concrete evidence will I produce with this capital that the next-round investor needs to see — and is the amount I'm raising enough to get there?
If you can't answer that with a specific metric and timeline, you're not ready to raise.
Track funding round benchmarks on the Benchmarking Dashboard at Value Add VC. Originally published in the Trace Cohen newsletter.