Research Question
Are current late-stage private market valuations sustainable in a structurally higher rate environment?
Research Summary
Late-stage multiples have compressed ~35% from 2021 peak but remain ~20% above 2015–2019 averages. AI-driven revenue growth is masking what would otherwise be a deeper repricing.
Evidence For
- No strong supporting evidence found in current literature.
Evidence Against
- No notable contradicting evidence found.
Relevant Data
12.4x
Late-stage NTM revenue multiple
vs 2015-19 avg 10.1x
9.2 years
Median time to liquidity
Final Assessment
Repricing is partial, not complete. Another leg down likely if AI growth narratives soften.
58/100
Confidence reflects strength of evidence weighted by source quality.
For the Geeks · sources & method▾
Statistical method
Structured expert elicitation (IDEA protocol)
4 independent domain experts, Delphi-style two rounds, aggregated via geometric mean of probabilities.
n = 4 experts · 2 rounds
Agent chain
GPT-5.1 (reasoning)Claude Sonnet 4.5 (verification)Zerve Retriever v3
Data sources
- Acemoglu & Restrepo (2023) — Tasks, Automation & Wage Effects ↗NBER Working Paper 28257.
- PitchBook — Global Venture Report Q1 2026 ↗Stage-by-stage deal counts, median round size, dry powder.
- OECD AI Policy Observatory ↗Cross-country AI adoption + labour exposure indices.
Last agent run · 2026-06-08