How do you value a startup?

“Most VCs price you backward from an ownership target, not forward from your spreadsheet: they need 15–20% ownership, so your valuation is often check size divided by that target. Growth rate sets the multiple, not ARR. Two companies at $5M ARR can trade 5x apart if one is growing 50% and the other 150%. Top decile today is roughly 10x ARR, more for AI-native, less if growth is soft.”

“I don't think there is a true value for an early-stage startup; the number is implied by what an investor pays for a real share, then checked against comparable deals and the return they need to hit. The valuation is mostly a negotiation over potential, risk, and dilution, not a spreadsheet output.”

“I don't value a startup by what it's worth today. At SurePayroll we raised $17M, and at MATH we looked for early signals before they showed up on any spreadsheet. Pre-revenue, valuation is a calculus of risk, reward, and whether the next round can clear the bar this one sets. A $3M raise on a $15M pre feels like a win until you realize you must build enough to justify north of $20M next time, or the game ends.”

“There is no algorithm for valuing a company. In early-stage deep tech, I work backwards from the outcome: if Lux needs a $2 billion return on a $500 million venture fund, we need a handful of companies that can exit at multi-billion-dollar valuations, and we need to own enough at entry to survive dilution.”

“I care most about the exit math: can this company plausibly return my fund if everything works. That means I start with the founder and the market, then work back to price.”

“I don't price deals off comparables or DCFs; I pay for the risk I'm taking and the size of the outcome if it works. The best founders make me believe the biggest risk is already smaller than it looks.”

“I don't price rounds off a spreadsheet. At the earliest stage I'm asking whether this founder is the one I'd bet on if the model changes three times, and whether the market is big enough that a win matters. The number is downstream of conviction, not the other way around.”

“I mostly don't value startups the way spreadsheets do; at seed, I'm trying to answer whether this is the rare company that could return the fund, and that comes down to the founder and the market they might own. Later-stage valuation is more about comps, growth, and efficiency, but even then the best deals are the ones where the future is underpriced because others can't yet see it.”

“I don't anchor on a spreadsheet; I anchor on whether this founder can move mountains, then check if the price lets the round still make sense for the fund. Conviction cannot be price-dependent, but ridiculous is still ridiculous.”

“So I think valuation is mostly a byproduct of the story the founder is selling and the market they're chasing, not a spreadsheet exercise. At Greylock, I'm really trying to figure out if this can be a $10 billion-plus company, because that's what drives our returns.”

“I don't price a startup from a spreadsheet. I start with the founder and the problem: is this a hair-on-fire issue, do they have the tenacity and edge to own it, and is the market big enough if it works.”