Sameer Dholakia says Bessemer's Anthropic Series D at $15B pre was the lowest-voted deal in firm history at the time. Byron Deeter and Sameer pushed it through, and Bessemer's empowerment culture let conviction override consensus. He wishes they had written a bigger check. That is where alpha comes from.
Sameer Dholakia argues intelligence will be a larger market than oil, so the precise split between open and closed tokens is not worth predicting. He says even the common '99% of tokens open, 99% of dollars closed' ratio is wrong, but both will be enormous.
Sameer Dholakia says Bessemer paid roughly 150x current ARR for Anthropic and about 100x for Legora, which would make many VCs flinch. He focuses instead on the next 12 months of forecast gross profit, where 10-15x NTM can be defensible. The rear-view multiple is less important.
Sameer Dholakia says competition breadth has not changed, but capital intensity behind competitors has, creating new intensity. He believes capital is a moat and kingmaking matters because compute is scarce and expensive, yet it is necessary but not sufficient without a great product. Founders still must deliver.
Sameer Dholakia says seed checks used to be $3-5M on $20-40M valuations, but AI deals now price factors higher. With so many shots on goal needed to find winners, he does not see how a small seed fund succeeds. He says the math pushes seed funds larger.
Sameer Dholakia compares venture to other maturing industries: as it professionalizes, winners consolidate into giant global platforms or narrowly specialized firms. He says being a mid-sized generalist in between is a rough place to make money, and he is glad not to run a billion-dollar fund caught there.
Sameer Dholakia says Bessemer's late-growth funds average 5-10% per company, but they flexed to roughly 20% in Anthropic. He argues concentration is still the goal even at $4B scale, because rare generational outcomes justify breaking the normal sizing rule when conviction is high. The fund can still be concentrated.
Sameer Dholakia says Bessemer invested in Fireworks at a $15B valuation and he sees a potential 20x return. He argues outcome scenarios are bigger than they have been in a long time, so de-risked late growth can produce venture-style multiples. That changes late-stage underwriting.
Sameer Dholakia says SendGrid flew the entire company to Mexico each January and dedicated a day to shared non-work activities like kayaking or mountain biking. He argues teammates solve problems together while colleagues point fingers or leave when things go wrong. That is a strategic advantage.
Sameer Dholakia says companies will stay private longer because private venture capital is larger and LPs want AI gains before public markets. He still expects an IPO unlock later next year and in 2028, and says going public imposes valuable operating discipline. That discipline makes stronger companies.
Sameer Dholakia says democratized access is hard to argue against, but retail investors should only back giant platforms or true specialists because venture is high risk. Otherwise they risk adverse selection into weaker funds, and many would be better off compounding capital in public markets.
Sameer Dholakia says Fireworks' pitch is that enterprises will not want to rent all intelligence from a model provider. He pairs that with Lynn's technical depth and George Hu's go-to-market strength, which is why he sees a very large business. That combination underpins his bull case.
Sameer Dholakia says Bessemer's early fund needs a fund-returner at roughly 50x while the growth fund needs about 10x. He says early-stage hit rates are lower so the biggest winners must pull up the misses; growth is more de-risked. The blended portfolio still has to clear a 5x fund target.
Sameer Dholakia says investors who compare Waymo only with Uber miss personal car ownership, trucking and logistics. He imagines buying a car with a Waymo driver subscription and argues Waymo could license its stack to every OEM, making it a trillion-dollar business. He says that is easy to see.
Sameer Dholakia says therapeutics is the greatest potential impact of AI, with automated wet labs and models reasoning over massive datasets. He cites Dario Amodei's view that the next decade could deliver biological progress that would otherwise take fifty years, and he believes that compression is underappreciated.
