Roughly 9 out of 10 companies fail, meaning only a handful of bets can offset the losses.
The first few years will feel awful; losers shut down first, and winners can take 10+ years to show how big they’ll get.
Round names are unreliable: AI companies raising $100m+ can still be pre‑product.
I liken early‑stage investing to playing the claw machine in Mtn View, CA at Aladdin’s Castle, where most tries yield nothing but occasional jackpots.
I usually recommend building a portfolio of 50 to 100 companies.
Benchmark invested $6.7M in eBay in 1997 for ~22% of the company (roughly a $30M post-money valuation).
eBay was already at a ~$4M annual revenue run rate AND profitable.
I spread investments over 5–10 years, which translates to about 5–20 companies per year, allowing exposure to both up‑ and down‑markets.
Because early years feel painful – losers shut down first and winners can take 10+ years to show how big they'll get – I stay the course and finish building my portfolio.