Why the AI premium poisons non-AI comp sets.
In 2026, AI-labelled startups raise at materially higher valuations than non-AI peers at the same stage. The premium peaks at Series A — AI deals price +85% over non-AI — and stays meaningful through Series C. For founders building non-AI companies, this matters operationally: if you include AI deals in your comparable benchmark set, you produce inflated valuation anchors that VCs reject in negotiation.
The premium by stage
What this looks like in dollars
Series A non-AI median pre-money: $48M. AI Series A median pre-money: ~$89M ($48M × 1.85). A founder pulling 10 Series A comps from Crunchbase or PitchBook in 2026 will get a mix — say 4 AI deals at $80-$120M and 6 non-AI at $40-$55M. The naive median across all 10 is approximately $65M. Using that as your benchmark when you're non-AI overstates your defensible valuation by 30%+.
How to filter AI from comp sets
- Read each comp's positioning. If the company describes itself as “AI-native”, “AI-powered”, “LLM-first”, or includes generative-AI as core product, it's an AI deal even if the underlying SaaS category isn't.
- Check the round announcement language. AI deals consistently lead with model-capability claims, evaluation benchmarks, or training infrastructure. Non-AI deals lead with ARR, growth rate, or customer metrics.
- For ambiguous deals (a SaaS company that added an AI feature), check whether the marketing positioning treats AI as the core product or a feature. Core-AI = AI premium; AI feature = non-AI premium.
- If your comp set ends up with 6 AI and 4 non-AI deals, separate them. Report “AI median: $X. Non-AI median: $Y.” Then state which group you're benchmarking against.
If you ARE AI-native
The +85% premium is the headline but it's not uniformly distributed. Foundation-model companies (OpenAI, Anthropic, Mistral peers) command the highest multiples. Application-layer AI (vertical SaaS with AI features) commands lower premiums — closer to +30-50% at Series A. AI infrastructure (training, inference, evaluation) sits in between. Pull comp sets from your specific AI tier, not the aggregate.
Practical recommendation:When VCs ask for your benchmark, lead with the AI/non-AI split. “Our 8 comps split 3 AI ($65M-$110M) and 5 non-AI ($35M-$58M). We're non-AI, benchmarking to the non-AI subset at $46M median.” This pre-empts the VC's likely objection about AI deals being in your set.