The question every Series A founder asks: "What ARR do I actually need?" The honest answer is that the benchmark has bifurcated. AI startups are closing rounds at ARR levels that would have been seed checks in 2022, while traditional SaaS founders are being held to $2-3M+ ARR thresholds.
This piece synthesizes 2025 round data from Crunchbase, Carta, PitchBook, and OpenView's annual benchmark report, plus our own observations from running diligence support on 30+ Series A rounds in the last 18 months.
The headline numbers
| Sector | Median Series A ARR | 25th–75th percentile | Notes |
|---|---|---|---|
| Traditional SaaS | $2.2M | $1.5M – $3.5M | Growth rate matters more than absolute ARR |
| AI-native SaaS | $1.1M | $400K – $2.5M | Wide distribution; growth rate is everything |
| Fintech (B2B) | $3.0M | $2M – $5M | Compliance overhead pushes Series A later |
| Fintech (consumer) | varies | revenue often not primary metric | GMV, MAU, contribution margin matter more |
| Marketplaces | $4M GMV / $800K take-rate | $2-8M GMV | Take-rate + cohort retention drive valuation |
| Biotech / hardware | n/a | n/a | Milestone-based; ARR not the gating metric |
| Climate / energy | varies | $1.5M – $5M | Project-based; ARR mixed with project revenue |
These are 2025 closing-round benchmarks. The 2026 environment looks similar at the high end, with continued compression on the low end for traditional SaaS as AI-enabled competitors push down margins.
Why ARR alone is a misleading benchmark
If you take only one thing from this report: investors don't price on ARR. They price on the combination of:
- ARR magnitude: the absolute number you have today
- Growth rate: what your YoY or T3M-annualized growth looks like
- Net revenue retention (NRR): whether existing customers expand or churn
- Burn multiple: how much you spent to generate that ARR
- Path to credible $100M ARR: whether the trajectory math works at all
A $1.5M ARR business growing 300% YoY with 120% NRR and a 0.8x burn multiple will get priced higher than a $4M ARR business growing 60% YoY with 90% NRR and a 2.5x burn multiple. The first one is a Series A; the second one is a seed extension.
The AI exception
AI-native startups have rewritten the benchmark in 2025. We've seen four patterns:
Pattern 1: Sub-$1M ARR, extreme growth
Many AI-native startups closed Series A in 2025 with $300K-$1M ARR but were doubling MoM. The thesis: by the time the round closes and announces, they'll be at $3-5M ARR and a $50M+ post valuation.
This works because:
- Inference costs are dropping fast → margins improving
- Distribution is product-led → CAC is near-zero
- TAM is genuinely large (every SaaS gets re-imagined)
Pattern 2: $1-3M ARR but 200%+ NDR
AI tooling startups where existing customers double their seat count or usage every quarter. The Series A pitch isn't "this is a $3M ARR business." It's "this is a $30M ARR business in 18 months at current expansion rates."
Pattern 3: Vertical AI applications
Industry-specific AI products (legal AI, healthcare AI, manufacturing AI) often raise Series A at $1.5-2M ARR with strong unit economics + a clear path to becoming the category leader. Closer to traditional SaaS benchmarks but with better growth.
Pattern 4: AI infrastructure
Foundation model orchestration, evaluation, agent frameworks: many raising Series A on $500K-$1M ARR + strong design partner traction. Investors are pricing on category dominance potential, not current revenue.
What changed from 2022-2023 benchmarks
The 2021-early-2022 era had Series A closing at $500K-$1M ARR for traditional SaaS. That window slammed shut in mid-2022 and the benchmark has reset higher:
- 2021 median: ~$800K ARR
- 2022 median: ~$1.5M ARR
- 2023 median: ~$1.8M ARR
- 2024 median: ~$2.0M ARR
- 2025 median: ~$2.2M ARR (traditional SaaS)
For traditional SaaS, the floor is meaningfully higher than 2-3 years ago. Investors learned the hard way that $500K ARR seed extensions weren't durable.
For AI-native categories, the benchmark is set by growth velocity, not ARR floor. Investors who insisted on $2M+ ARR missed the highest-momentum AI companies of 2023-2024.
Growth rate as the dominant signal
A 2024 OpenView analysis of Series A rounds found that growth rate is the strongest single predictor of round size + valuation, more so than ARR magnitude:
| Growth rate (T12M) | Typical Series A outcome |
|---|---|
| <50% YoY | Hard to raise; pitch as seed extension instead |
| 50-100% YoY | Need higher ARR ($3M+) to compensate |
| 100-200% YoY | Sweet spot; $1.5-2M ARR sufficient |
| 200-400% YoY | Premium round; $1M ARR sufficient if other metrics solid |
| >400% YoY | Investor competition; pre-empts at any ARR |
If your growth rate is below 100% YoY at the Series A stage, your pitch needs to lead with something other than growth — durable margins, defensible moat, exceptional NRR, or a category-creating wedge.
Burn multiple at Series A
The benchmark that became dominant in 2023-2024 and remains so in 2026:
| Burn multiple | Interpretation |
|---|---|
| <1.0x | Best-in-class. Pre-empt likely. |
| 1.0-1.5x | Excellent. Easy raise. |
| 1.5-2.0x | Good. Need a growth story to justify. |
| 2.0-3.0x | Pushing the limit. Better have a moat. |
| >3.0x | Hard to raise without exceptional growth (>200% YoY). |
Burn multiple = net burn / net new ARR. So if you burned $500K last quarter and added $400K of net new ARR, your burn multiple is 1.25x.
For deeper coverage of burn multiple + Rule of 40 dynamics, see Burn Multiple, Rule of 40, Net New ARR.
NRR thresholds at Series A
| NRR (T12M) | Interpretation |
|---|---|
| <90% | Logo churn problem. Will gate the raise. |
| 90-100% | Acceptable but not exciting. |
| 100-110% | Healthy. Industry standard. |
| 110-130% | Strong. Premium pricing justified. |
| >130% | Exceptional. AI tooling territory. |
NRR at $1-3M ARR is often noisy due to small denominator. Investors will look at quarterly cohort retention curves more than the headline NRR number.
What investors will diligence
If you're approaching Series A in 2026, expect investors to drill into:
- Cohort retention curves: show every quarterly cohort's gross retention over time
- ARR composition: % from top 10 customers, average contract value, contract length
- Pipeline → close-rate analysis: sales velocity, win rate, cycle time
- Burn vs net new ARR by quarter: burn multiple trend, not just average
- Unit economics by acquisition channel: CAC payback by channel
- Headcount plan vs revenue plan: when does each function scale?
This is the work that the diligence-readiness quiz tests. If you'd struggle to produce any of the above in 48 hours, you're not Series A diligence-ready.
Bay Area / SF-specific dynamics
For SF/Bay Area-based startups raising from tier-1 venture firms (a16z, Sequoia, Greylock, Founders Fund, etc.):
- ARR thresholds run ~10-20% lower than the national median due to proximity / pattern recognition
- Pre-money valuations run 20-40% higher
- Burn multiple tolerance is similar; Bay Area VCs are not more generous on efficiency
- Series A round sizes trend larger ($15-25M vs $10-15M national median)
These reflect the warmer dealflow + valuation environment in SF, not a different underlying benchmark.
Companion data + tools
- Founder Salary Report 2026: what CEOs actually pay themselves at seed through Series B
- Burn Multiple Benchmarks by Stage 2026: burn multiple targets at each stage
- Startup Valuation Benchmarks: pre-money by stage
- Revenue Benchmarks for Fundraising: ARR thresholds across stages
- Runway Calculator: model your runway against revenue scenarios
- Diligence Readiness Quiz: 8-question score on Series A readiness
Sources
- Crunchbase Q4 2025 round data
- Carta 2025 State of Private Markets
- PitchBook NVCA Venture Monitor 2025
- OpenView 2025 SaaS Benchmarks Report
- StartupCFO internal data, 30+ Series A diligence engagements 2024-2025
