20VC: You Need $1BN To Do Seed Today: The Untold Truths of Venture | Who Wins and Who Loses in the Next Decade of VC | Why Margin and Price Matter Less in a World of AI with Sameer Dholakia, Bessemer

Bessemer's Dholakia Says AI Rewards Paying Up for Growth

Sameer Dholakia argues AI is a super cycle on the scale of the cloud, and that it rewards investors who pay up for category winners. A few labs will consolidate while hundreds of applications grow on top, so the market is large enough for growth to outrun a high entry price. That also reshapes venture, where capital becomes a moat and funds must be giant platforms or specialists.

Takeaways

• Frontier labs will consolidate while applications multiply. Sameer's cloud-redux thesis holds that compute-heavy labs shrink to a handful of players while hundreds of applications are built on top. Bessemer joined Anthropic's Series D at a 15bn dollar pre-money valuation and wishes it had written a bigger check.
• Growth at scale matters more than entry price. Sameer defends paying about 150 times current ARR for Anthropic and Legora by using next-twelve-month revenue. He calls 10 to 15 times NTM fair for hypergrowth, looks no further than 18 to 24 months, and weighs retention, margins and a five to seven year hold.
• Capital is now a competitive moat. The breadth of competition is unchanged, but the capital behind competitors has grown. Sameer says capital helps win customers and talent, though a great product is still required. Triple-triple-double-double growth alone no longer draws capital, yet those companies can still return well at lower prices.
• Venture's middle is squeezed between platforms and specialists. Seed checks of 3 to 5m dollars on 20 to 40m valuations are now factors higher. Sameer says funds must be giant global platforms or true specialists, and he does not see how a small seed fund gets enough shots on goal.
• Bessemer concentrates growth capital in high-conviction bets. Bessemer may write 200 to 400m dollar checks, and average late-stage checks run 5 to 10 percent of the fund. The last growth fund put 20 percent into Anthropic. One group of four general partners covers early and late growth, targeting at least 5x.
• Open and closed models can both build huge businesses. Sameer rejects choosing sides because the market is large enough that the token mix is irrelevant, citing Red Hat. Hosting by Fireworks removes the Chinese label, enterprises will want to own some intelligence, and sovereign models such as Sarvam and Mistral suit governments.
• Companies stay private longer, so liquidity matters. Sameer says DPI matters, firms should take some chips off the table, and founders should sell some secondary. He expects the IPO window to reopen next year and in 2028, and says public-company discipline makes better companies.

Team culture Sameer carries from SendGrid