Winner-Take-All Has Taken Over Venture. The Top 1% of AI Start-Ups Are Valued at 3-10x ‘Normal’ Multiples
Based on fresh Carta data from 3,001 primary rounds raised by US startups (Sept 2024 – August 2025).
The venture capital market has officially split into two radically different universes. And this is super important for B2B founders to understand:
- In Universe A, solid B2B and SaaS companies raise at reasonable multiples following traditional venture math.
- In Universe B, AI-powered startups command valuations that defy every historical precedent.
The data from Carta’s latest analysis reveals something unprecedented: we’re not just seeing a “hot market” — we’re witnessing the complete takeover of winner-take-all economics in venture capital.
Numbers That Will Make Your Head Spin
Let’s start with the headline: Seed rounds in the 99th percentile are hitting $161M valuations. That’s not a typo. That’s +123% higher than even the 95th percentile at $72M.
But it gets wilder as you move up the stack:
- Series A: The gap between “great” (95th percentile at $250M) and “legendary” (99th percentile at $716M) is a +186% jump. We’re talking about Series A companies valued higher than most Series C rounds just 24 months ago.
- Series B: The 99th percentile hits $2.068B — that’s unicorn territory in what used to be a “growth” round. The jump from 95th to 99th percentile? A casual +200%.
- Series C and beyond: We’re now seeing Series C companies in the 99th percentile valued at $7.169B and Series D at $8.104B. These aren’t outliers anymore — they’re the new “wow” category.
You can see it play out live in our new, SaaStr VC Valuation Calculator here:
What’s Actually Happening Here?
This isn’t market irrationality. This is the mathematical result of winner-take-all economics reaching venture capital. When the difference between #1 and #3 in any AI category means the difference between a $200B and a $1B outcome, traditional valuation models break down entirely.
1. AI Has Created Super-Category Winners
The companies hitting 99th percentile valuations aren’t just “AI companies” — they’re companies building AI moats so deep that competition becomes nearly impossible. When your model gets better with every customer interaction, when your data flywheel compounds daily, when your AI creates switching costs that didn’t exist in traditional SaaS, you’re not competing in the same market as everyone else.
With OpenAI raising at $500 Billion, and Anthropic at $170 Billion, the potential returns for VCs for investing in a category winner seem almost unlimited today.
2. Capital Efficiency Meets Infinite TAM
The deadliest combination in today’s market: AI companies that can show both massive TAM expansion AND improving unit economics. Traditional SaaS companies might 3x their addressable market through international expansion. AI companies are 10x-ing their TAM by automating entire job functions. When you combine that with AI-driven operational efficiency, investors will pay almost any multiple.
3. The “Never Raise Again” Premium
The top 1% of AI companies raising at these valuations have a credible path to never needing external capital again. Their AI advantages compound, their margins improve with scale, and their competitive moats widen automatically. VCs aren’t just paying for growth — they’re paying for the chance to own a piece of a business that might generate cash for decades without dilution.
The Three-Tier Market Structure
What we’re seeing is the emergence of a three-tier market:
Tier 1: Traditional SaaS (50th-75th percentile). These are well-built software companies following the old playbook. Series A at $58-97M, Series B at $136-262M. Good businesses with predictable SaaS metrics, but they’re competing in a world where AI isn’t central to their value proposition.
Tier 2: AI-Enhanced SaaS (90th-95th percentile). These companies have successfully integrated AI into existing SaaS models. Series A at $160-250M, Series B at $450-688M. They’re getting the “AI premium” but haven’t achieved true AI-native differentiation.
Tier 3: AI-Native Category Creators (99th percentile). This is where the magic happens. These aren’t SaaS companies with AI features — they’re AI companies that happen to have software interfaces. They’re creating entirely new categories, automating entire workflows, and building moats that didn’t exist in the pre-AI world.
Bessemer Venture Partners also has a good take here: Great B2B Start-Ups vs. Shooting Stars … vs the new Category of AI Supernovas:
Almost everyone wants to fund the supernovas and shooting starts.
What This Means for You
If you’re a founder:
If you’re building traditional B2B/ SaaS, you’re competing for Tier 1 valuations no matter how good your execution. If you’re building AI-enhanced SaaS, you need to prove your AI creates genuine differentiation, not just feature parity. But if you’re building something truly AI-native — something that couldn’t exist without AI at its core — there’s more capital available at higher valuations than any founder in history has ever seen.
The question isn’t “should I add AI to my product?” It’s “can I build something that’s impossible without AI?”
If you’re an investor:
The cost of missing an AI-native category creator has never been higher. The companies you passed on at $50M pre-money in their Series A aren’t just raising Series B at $500M+ — they’re potentially building the next $100B+ businesses. But the risk of paying Tier 3 valuations for Tier 2 companies has also never been higher.
The due diligence question has fundamentally changed: it’s not “is this a good SaaS business?” It’s “does this AI advantage compound automatically?”
If you’re an operator:
The talent war is about to get even more intense. Companies raising at these valuations can afford to pay top-of-market compensation and offer equity packages that are genuinely life-changing. If you’re at a Tier 1 or Tier 2 company, you need to be thinking about your next move strategically.
The Dangerous Part
Here’s what keeps me up at night about this data: When 99th percentile Series A companies are trading at 7x median valuations, we’re not seeing market dynamics — we’re seeing market separation.
The gap suggests that investors believe the top 1% of AI companies will capture disproportionate value compared to everything else. Historically, when this kind of valuation divergence happens in venture, it means one of two things:
- The top tier really has discovered a new form of value creation that justifies infinite multiples
- We’re in the late stages of a bubble where “AI” has become the new “blockchain” or “metaverse”
Here’s why I think it’s mostly #1: The AI companies hitting these valuations aren’t just growing faster — they’re growing more efficiently while simultaneously expanding their addressable markets. They’re not just better businesses; they’re different businesses entirely.
But that “mostly” is doing heavy lifting. Because if I’m wrong, the correction will be unlike anything we’ve seen in venture capital.
The Bottom Line
Winner-take-all economics haven’t just influenced venture capital — they’ve completely taken over. We’re not living in a world where the best companies get 2x valuations anymore. We’re living in a world where AI-native companies trade in a completely different universe.
For founders, this means the difference between building a “good SaaS business” and building an “AI-native category creator” isn’t just better outcomes — it’s 10x different outcomes. The middle ground is disappearing rapidly.
For investors, this means the cost of being wrong about which companies are truly AI-native versus AI-enhanced has never been higher. Miss the next AI category creator, and you don’t just miss good returns — you miss generational returns.
The math is simple: In a winner-take-all world, you either build something that becomes the winner, or you compete for scraps. The valuation data just makes that reality impossible to ignore.
Choose your game accordingly. Because in venture capital, there’s now the AI game and everything else.
Data source: Carta analysis of 3,001 primary rounds raised by US startups, September 2024 – August 2025. Want more startup data insights? Subscribe at carta.com/data
