> ## Content Index
> Fetch the complete content index at: https://longtalkshort.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Seed Investor Who Won't Fight the Giants: David Tisch on Why Small Has to Be a Choice
- URL: https://longtalkshort.com/david-tisch-boxgroup-20vc-seed-investor-wont-fight-the-giants/
- Published: 2026-10-31T19:36:56.000Z
- Updated: 2026-10-31T19:36:56.000Z
- Description: BoxGroup's David Tisch says seed funds shouldn't fight the giant firms head-on, that 90% win-rate claims are "total BS" and that low ownership is an emotional worry, not a business one. Here is his case, with what we checked and his conflicts.
- Author: Truman
- Tags: Economics, David Tisch, 20VC

**If you have ever started something small and felt pressure to make it big, David Tisch has an argument for you.** He runs BoxGroup, a New York seed firm that writes small, early cheques into many startups. On 20VC, Harry Stebbings pushes him on whether that model still works with the giant funds moving into seed investing.

Tisch's answer: **the mistake isn't being small.** It is pretending you can beat the giants at their own game, or quietly hoping to become one.

*In short: Tisch argues that seed investors should stop competing head-on with multi-stage firms, stop obsessing over ownership and decide early what they actually want to become.*

## 1\. Don't fight the best firms head-on

> I've never understood trying to compete head to head against who I would deem to be the best VCs in the world

The large multi-stage firms have spent decades building brands and attracting founders, and he thinks they have become "really, really good at seed investing," more focused and more efficient. The market has shifted with them. **An early round that might have drawn three to five term sheets can draw 10 to 15,** and seed rounds of $10 million to $20 million are normal.

To him this is a fact to accept, not a grievance. "You can complain about it or just accept the market conditions," he says. BoxGroup has lost deals to big firms such as General Catalyst offering a much higher price and a much bigger cheque. What he won't do is **build a business that depends on beating those firms in a straight contest.**

So how do the funds that do fight claim to be winning?

## 2\. The win-rate myth

Tisch's sharpest claim is aimed at "concentrated" funds, the ones that back a few dozen companies and insist on owning a large slice of each:

> Every concentrated manager will say that we win 90% of the deals we want to win. And I think that's total BS. I think it's like 20%.

If he's right, the problem is **adverse selection.** A fund that plans to own 10% of 30 companies won't end up with the 30 it most wanted. It ends up with the 30 that accepted its terms. The founders with the most options are the least likely to give a big stake to a small fund. **A fixed ownership target can steer a fund away from the very companies that would make it.**

![A funnel diagram. At the top, a crowd of companies a seed fund most wants. A filter marked 'ownership demands' lets only some through. At the bottom, the companies the fund actually gets, which are not the same set it most wanted. No people are shown.](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/10/fig1-69.png)

Our sketch of the adverse-selection argument Tisch makes against fixed ownership targets. A concept diagram, not data. Made for Long Talk Short.

## 3\. Ownership is an emotional thing

BoxGroup's own stakes range from about 1% to 10%. "There's no too low," Tisch says. "That's an emotional thing, not a business thing." He admits he is sad when the firm writes a tiny cheque, but the real question is simply yes or no. When it does get squeezed, the usual cause is arriving late, when a round is already under way and the firm is "begging for scraps." **His fix is to meet founders earlier,** ideally before there is a company.

He also defends the arithmetic. Two things make a small stake worth more than it looks: **the biggest outcomes have grown much larger,** and the best companies dilute their early investors less than people assume. He tells Stebbings, a sceptic of the model, that he "didn't do the math right," and Stebbings concedes he hadn't seen outcome sizes expanding.

## 4\. How the firm decides

BoxGroup doesn't vote on deals. "Somebody here has to want to do a deal," and colleagues can't "throw a grenade on someone else's deal" unless there is an ethical or reputational issue. Most startups fail, so explaining why one will fail adds nothing. **The job is to see why it might work.**

Follow-on rights, he says, are "earned," not owed. He doesn't sell shares early in secondary deals, because "I work for our LPs, and my customers are founders." And he dismisses the lean startup movement as "a total fake": companies that want venture funding need a lot of money, even if being well funded shouldn't mean being undisciplined.

## 5\. Ask what Fund 4 looks like

His advice to new fund managers is to ask what their fourth fund will look like, because "Fund 4 is actually your vision." Most new firms, he thinks, start small but **really want to become multi-stage firms,** the most crowded part of the market. "Be careful what you wish for," he says; many people running big multi-stage firms "wish they were the smaller firms." BoxGroup, he says, has "no ambition to do anything else."

![A fork in a path after Fund 1. One branch climbs through Funds 2 to 4 towards a crowded cluster of multi-stage towers. The other stays level as early-stage funds, on purpose. A note reads 'decide which path before you raise Fund 1.' No people are shown.](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/10/fig2-57.png)

Our sketch of Tisch's 'what does Fund 4 look like' question. A concept diagram, not data. Made for Long Talk Short.

## Weighing it: what holds up

**Holds up:** BoxGroup raised a $275 million seed fund in October 2025, planned for about 150 companies. Tisch cites Cursor as a big win; SpaceX's $60 billion all-stock purchase of its maker, Anysphere, completed in August 2026, is the largest acquisition of a venture-backed startup on record.

**Disputed: his own win rate.** The win-rate figures are opinions, not data. Tisch puts BoxGroup's own rate at roughly 70% to 90%, close to the 90% he calls "BS" in others. Our transcript has a gap just before that answer, so the exact question is unclear.

**Not checked:** the term-sheet counts and his dilution claim. The transcript garbles how much more the bigger firm offered, so we haven't given a multiple.

**Conflicts:** Tisch is talking his own book. His case favours BoxGroup's model, and the companies he praises, including Cursor, Ramp, Clay, Baseten and Rogo, are ones the firm has backed. We didn't check for links to a16z, which he praises, or any fund ties of 20VC.

**Bottom line:** the adverse-selection logic is sound; the numbers around it are his estimates.

**Related reading:** for another 20VC debate on where the money comes from, see [Who Pays for the Chips?](https://longtalkshort.com/20vc-who-pays-for-ai-build-out-meta-microsoft-anthropic/); for judging decisions apart from outcomes, see [Judge the Bet, Not the Result](https://longtalkshort.com/annie-duke-thinking-in-bets-judge-the-bet-not-the-result/).

## Long Talk Short

- Multi-stage firms are good at seed; competing with them head-on is a losing game.
- Claims of a 90% win rate hide adverse selection: a fixed ownership target filters out the best deals.
- Low ownership is "an emotional thing"; arriving late is what costs you.
- No voting and no grenades: the job is to see why something works.
- Decide what Fund 4 looks like before you raise Fund 1.

**The one thing to remember:** staying small is a strategy only if you choose it on purpose.

Source: ["Why No Manager Wants to Stay Small & Why Competing with Mega Platforms is a Losing Game"](https://www.youtube.com/watch?v=BsvXF-0%5FG-Y&ref=longtalkshort.com), 20VC with Harry Stebbings and David Tisch (YouTube). We worked from a machine transcript and quote only clear passages. This is an independent summary and is not affiliated with 20VC, BoxGroup or David Tisch.

*This article is for general information only and is not investment advice.*