Founder Resource · Q3 2026 Update
Early-Stage Valuation Cap Guide & Fundraising Benchmarks
What valuation cap should you set? How much dilution should you expect? Where does your round fit? Benchmarks from Carta and AngelList, updated through August 2026.
Carta State of Pre-Seed Q2 2026 · 370K+ instruments
Carta State of Private Markets Q1 2026
AngelList live data May 2026 + State of Venture H1 2025
SVB State of Markets H1 2026
US Startups · Post-money SAFEs + Priced Rounds
💡
How to use this guide
Start with the Overview tab to understand where each round fits in the fundraising journey. Then select your specific round type for detailed val cap benchmarks by amount raised and sector. Use this to benchmark your cap, not as a fixed target. Every deal is different.
Want your own numbers? The
free SAFE calculator runs on these exact benchmarks: enter your amount and cap, and it returns your ownership, your dilution through to Series A, and how your cap sits against the median for your round size and industry.
Where the numbers come from: every benchmark here is real market data, not estimates, drawn from
Carta and
AngelList, the cap-table platforms thousands of US startups use to manage equity and run their funding rounds, so the figures reflect actual closed deals. Revenue (ARR) figures come from
Silicon Valley Bank.
⚖️
Round stage vs. instrument, two different things
A funding stage, pre-seed, seed, Series A, describes how mature your company is and roughly how much you raise. An instrument, SAFE, convertible note, or priced equity, is the legal document used to take the money. They are independent: a seed-stage round can be raised on a SAFE (pre-priced, simple, no term sheet) or as a priced equity round (term sheet, lead investor, formal valuation). That is why this guide has both a “Seed on SAFEs” and a “Seed (Priced)” tab, same stage, different paperwork.
Want to dig deeper? Tier framework, Bessemer growth archetypes, and PitchBook AI premiums.
Where do you fit? Three tiers of the 2026 market.
Winner-take-all economics have split venture capital into three distinct categories, with growth rate as the key multiplier. Find your tier to know which benchmarks to use.
50th–75th percentile
Traditional SaaS
Well-built software companies following the proven SaaS playbook with predictable metrics.
Base multiples
- 8× to 15× ARR
- Growth-adjusted: up to 20× ARR max
Growth impact (per Bessemer)
- 150%+ growth: +30% premium (capped)
- 100–149% growth: +20% premium
- <50% growth: −30% discount
Carta H2 2025 · PitchBook Q3 2025
90th–95th percentile
AI-Enhanced SaaS
Successfully integrated AI into existing SaaS models, capturing the “AI premium” but not true AI-native differentiation.
Base multiples
- 15× to 30× ARR
- Growth-adjusted: up to 48× ARR
Growth premiums (per Bessemer)
- Shooting Star (200%+): +60% premium
- Supernova (300%+): +100% premium
- Below 100%: discount applies
PitchBook Q3 2025 · Bessemer State of AI 2025
99th percentile · step function
AI-Native Category Creators
Core AI products. Bessemer's Supernova (300%+ YoY) and Shooting Star (200%+ YoY) trajectories. Includes companies like Perplexity, Cursor, and Abridge.
Step function logic
Only companies hitting Bessemer's 200%+ growth bar get top 1% Carta + PitchBook multiples. Without extreme growth, AI-native companies drop to the AI-enhanced tier.
Multiples by growth tier
- Supernova (300%+): 99th percentile, up to 120× ARR
- Shooting Star (200%+): 98th percentile, up to 90× ARR
- <200% growth: drops to 90th–95th percentile (AI-enhanced tier)
Growth requirement is absolute. No growth rate = no top tier access. Carta + PitchBook top 1% requires Supernova trajectory. Winner-take-all dynamics demand extreme scaling.
Bessemer State of AI 2025 · PitchBook Q3 2025
The multiples shown are top-end observations at each percentile, not medians. The 120× and 90× figures represent the 99th and 98th percentile multiples for companies hitting Bessemer's growth bars. Median AI-Native multiples are closer to 30–50× ARR. Use these for upper-bound modeling, not central-case projections.
🤖 The AI premium is plateauing at the bottom of the market, still concentrating at the top
These three tiers assume an AI premium that keeps widening. At pre-seed, it no longer does. AI startups took 48.6% of all pre-seed dollars in H1 2026, essentially flat against 2025's full-year 49.8% after climbing from about 32% in 2021 (Carta, State of Pre-Seed Q2 2026). Carta calls it a plateau: the past six months showed no continued rise.
Higher up the market the concentration continues. Carta's State of Private Markets Q1 2026 puts more than 60% of all venture capital into AI companies, with 14.2% going to foundational-model companies alone, and 83% of SaaS capital flowing to AI-labelled startups. The two findings do not contradict each other, they describe different floors of the same building: at pre-seed the AI share has found its ceiling near half the market, while at growth stage the winners keep absorbing a rising share. Practically, being an AI company is table stakes at pre-seed rather than a differentiator, and the premium you can actually negotiate sits in the tiers above.
Bessemer growth benchmarks: what “great” looks like
Bessemer studied 20 high-growth AI startups including Perplexity, Cursor, and Abridge to define two archetypes for the AI era. These benchmarks replace the SaaS-era “T2D3” framework with “Q2T3.” Source: Bessemer State of AI 2025, the most recent edition. The 2025 date is the vintage of the study, not of this page, and no 2026 edition has been published as of August 2026.
Supernovas
Explosively scaling AI startups with unprecedented growth and adoption.
| Metric | Year 1 | Year 2 |
| Annual Recurring Revenue | ~$40M | ~$125M |
| Gross Margin | ~25% (often negative) | ~25% |
| ARR per FTE | $1.13M | 4–5× typical SaaS |
The high revenue often comes with fragile retention and thin margins. “Thin wrapper” labels could be thrown. Margins stretched to zero or negative as startups fight for winner-take-all prizes.
Examples studied by Bessemer: Perplexity, Cursor, Abridge.
Shooting Stars
Fast-growing, capital-efficient AI startups with strong PMF, solid margins, and loyal customers, scaling like stellar SaaS.
| Metric | Y1 | Y2 | Y3 | Y4 |
| ARR | ~$3M | ~$12M | ~$40M | ~$103M |
| Gross Margin | ~60% across the trajectory |
| ARR per FTE | ~$164K |
The Q2T3 pattern (quadruple, quadruple, triple, triple, triple) replaces SaaS-era T2D3. These businesses look more like stellar SaaS companies, with strong gross margins and durable retention.
“While we love Supernovas, we believe this era will be defined not by a few outliers, but by hundreds of Shooting Stars.”
Bessemer Venture Partners
Bottom line for founders: if your trajectory doesn't match one of these two patterns, you'll be benchmarked against Traditional SaaS multiples. The AI premium is real but earned, not assumed.
PitchBook Q3 2025: AI vs Non-AI premiums
The market is rewarding AI with concrete valuation premiums at every stage. Here's the data.
| AI Valuation Premium | 56% higher at Series C · 230% higher at Series D+ |
| AI Step-Up Premium | Series B step-up at 2.1× for AI vs 1.4× for non-AI (PitchBook Q3 2025, Series B only). Refreshed, all stages pooled: 2.2× for AI vs 1.6× for non-AI in H1 2026, up from 1.9× / 1.4× in 2025 (PitchBook-NVCA Q2 2026). Different scope, so the two rows are not interchangeable: the gap is widening on both cuts. |
| Market Reality | Down rounds at 11.4% (Carta, State of Private Markets Q1 2026, lowest since 2019), back down from a peak near 22% in 2023. Q2 2025 IPOs averaged −30% from peak private valuations (PitchBook Q3 2025). |
| Capital Concentration | AI captures 65% of total funding with only 35% of deal count |
What this means: capital is concentrating sharply. AI-native and AI-enhanced companies are absorbing the majority of venture dollars even though they represent a minority of deals. For non-AI founders, the valuation bar is higher than it appears at face value, and the burden of proof shifts to growth and capital efficiency. For AI-native founders, the premium is real but conditional on hitting Bessemer's growth bars.
For informational purposes only · Not legal or financial advice · Data: Carta, AngelList, Silicon Valley Bank, Bessemer Venture Partners, PitchBook public reports · Compiled by BVJ Consulting · Updated Q3 2026