Benchmark's GP, Everett Randle on Why Mega Funds Will Not Produce Good Returns
Benchmark GP Everett Randle argues mega funds like a16z and GC cannot return 5x net to LPs and that AI companies require an entirely new valuation framework beyond SaaS metrics.
- Randle’s current Benchmark fund includes Mercor and Sierra at 60x and 30x MOIC respectively on invested capital.
- Mega funds (a16z, GC, Lightspeed) writing billion-dollar checks cannot credibly promise LPs 5x net returns; fund math defies physics at $8-10B scale.
- Tiger Global will likely perform far better than critics expected, given early large positions in OpenAI and Databricks plus liquidation preference protection.
- AI app companies with heavy inference costs may show 50% gross margins but generate $500K gross profit per customer vs. $200K for SaaS — margin % is the wrong metric.
- Code generation grew from near-zero to $6-7B ARR in ~2.5 years; Randle expects $4-5B net new ARR added in a single year, qualifying it as a multi-winner golden category.
- Randle missed OpenAI’s $32B round at KP by over-indexing on nonprofit structure and dilution risk, calling it his biggest miss.
- OpenAI likely reaches $1T valuation by end of 2025; Randle prefers OpenAI at $500B over Anthropic at $350B due to ChatGPT’s consumer dominance.
- Founders Fund’s conviction-testing mechanism: investors personally co-invest alongside the fund, creating a live accountability check on every deal they sponsor.
2025-11-10 · Watch on YouTube
