Haystack's Semil Shah on the broken unicorn factory, emerging fund manager traps, and 12 years of seed investing

Summary

Semil Shah built Haystack out of necessity. Unable to land a job at any of the VC firms he was consulting for, he started with a $1 million fund and in the first eight months wrote seed checks into Instacart, Envoy, DoorDash, and HashiCorp. He took no fees across the first three funds — on a $1 million fund, that meant roughly $200k spread over 10 years — and describes the early years as surviving on luck as much as skill.

Emerging manager traps

Shah's sharpest critique is aimed at new managers who treat the fund as a salary vehicle before they've earned LP trust. His line: nobody owes you a paycheck to deploy other people's capital. He's watched managers go to market targeting $20–30 million when LPs are only willing to fund $6–8 million, and walk away rather than start smaller. AngelList gave Shah a way around that dynamic early on — SPVs, rolling funds, and a private SMA arranged through Naval Ravikant, who whitelisted four emerging managers to receive capital from an undisclosed sovereign fund.

On portfolio construction, Shah argues the math is no more complex than basic algebra, but it was still the hardest concept for him to fully internalise. He credits time around larger funds for building that intuition, and warns that LPs scrutinise every wire in and out. Inconsistent deployment patterns become visible by fund two or three, and LPs will ask for them to be smoothed out.

How long before you hang up the cleats

For Series A investors doing two to four deals a year at $7–8 million a check, a weak portfolio becomes apparent quickly — fund leadership tracks follow-on rates and board metrics, and AI is making that monitoring more systematic. Seed investors get more runway because they can take more shots at smaller check sizes, but Shah thinks five years without a standout is actually too long. The check-size ceiling matters too: he says the competitive line for seed investors tightens somewhere between $250k and $500k, where the best rounds become harder to access.

He flags two archetypes that consistently overestimate their edge. The first is the casual angel who backed 15 companies, got lucky on two unicorns, and now wants a $50 million fund. Shah's own experience cuts against that logic — his two best angel investments were capped at $25k each; pushing to $50k would have gotten him cut. The second is the big-fund investor who mistakes brand for judgment. His counter-example is Matt Kohler's 12-year run at Benchmark, where the hit rate at Series A was exceptional on a rifle-shot basis — a standard very few will replicate.

The broken unicorn factory

Shah agrees with Sam Lessin's framing that capital pressure has cascaded down every stage: crossovers squeezed growth, growth investors pushed into Series A territory, and that compression has inflated seed pricing. But he's heard the same complaint every year for 12 years, with brief exceptions around COVID and other dislocations, and it has never resolved into the correction people predict. His answer is to ignore hot deals priced to perfection and stay disciplined about meeting founders consistently — his view is that most investors waste too much time at conferences and bad meetings, leaving genuine access gaps that a focused seed investor can still exploit.

On X versus old Twitter, Shah says the platform's content is still strong but harder to surface. The removal of TweetDeck broke his workflow and he hasn't found a replacement rhythm. His observation that the best tweets now circulate through private group chats rather than the public timeline points to a real discovery problem, even if the underlying content quality hasn't fallen.