The Twenty Minute VCSemil Shah: Lessons Learned Scaling from a $1M to a $50M Fund | 20VC #951
CHAPTERS
- 0:15 – 3:02
Motivation and fear: what Semil is running from vs. running toward
Semil frames his career motivation as a push-pull between fear of not being able to “participate” in the Bay Area ecosystem and the aspiration to become an early investor-of-record in enduring companies. He reflects on the emotional pressure of staying in California and the desire to show up on an S-1 as a surprising but meaningful early partner.
- •Fear of being forced out of the Bay Area ecosystem (financially/career-wise)
- •Seeing peers leave involuntarily as a motivating signal
- •Belief shift around 2015–2016 that long-term participation was possible
- •Running toward being the early, trusted partner who appears on the cap table/S-1
- •Investing as proximity to exceptional founders and outcomes
- 3:02 – 4:35
Choosing fund size: why Haystack stayed at $50M (and what changes next)
Semil explains that early Haystack funds often missed targets, but Fund 5’s $50M size worked well—especially during COVID—so they repeated it to reduce friction. The $50M target was designed to enable meaningful early ownership while keeping the strategy simple and consistent for LPs.
- •First four funds came in under target; later funds standardized at $50M
- •$50M framed as enabling ~10% positions at pre-seed/seed (in theory)
- •COVID-era support made “rinse and repeat” easier and lower-risk to pitch
- •Avoiding variability in docs/process to reduce fundraising friction
- •Expectation that future funds may grow modestly but remain under $100M
- 4:35 – 7:45
Pre-seed pricing reality: from SAFEs to “three on fifteen” medians
They unpack how pre-seed/seed round definitions blurred, while round sizes and caps inflated over the last decade. Semil argues the market median often looks like “$3M on a $15M” and warns that the evidence bar for seed is rising because the seed-to-A gap widened.
- •Round labels are noisy; Semil “gave up” on strict taxonomy
- •Inflation doubled traditional pre-seed SAFE sizes/caps over ~10 years
- •Median deal often resembles $3M raised on a $15M cap/valuation
- •Owning 10% across ~25 names + reserves quickly implies >$100M funds
- •Seed entry increasingly expects de-risking (product in market) to bridge to Series A
- 7:45 – 12:34
Talent vs. entrepreneurialism: identifying the overlap (and avoiding false signals)
Harry describes seed losses tied to pedigreed product leaders who moved slowly despite hype. Semil argues talent and entrepreneurial drive often diverge; Haystack tries to test for hunger, adversity, and prior “entrepreneurial expression,” not just polish and gravitas.
- •Hot, pedigree-heavy seed rounds can lack urgency/direction despite strong resumes
- •Big funds can rationally take these shots; constrained seed funds can’t do many
- •Core thesis: talent ≠ entrepreneurialism; success sits in the overlap
- •Signals to test: grit, adversity, hustle, endurance, prior side-hustles/initiatives
- •Interviewing the “how did you get here?” story to uncover drive beyond credentials
- 12:34 – 13:23
Series A as a signal engine: why hot As tend to lead to hot Bs
Semil agrees with Harry’s observation that premium Series A rounds tend to create follow-on momentum into Series B. The mechanism is simple: large pools of capital are eager to follow validated signals from competitive A rounds, at least until recent market shifts.
- •Hot Series As often correlate with hot Series Bs (more than later rounds)
- •Capital seeks signal/validation and piles into “premium” A winners
- •Follow-on capital availability amplifies earlier competitive dynamics
- •Pattern held for years, potentially changing in the current year’s environment
- 13:23 – 19:33
Why resist scaling AUM: dilution math, promises, and operating model constraints
Harry pushes Semil to raise more AUM; Semil explains why larger funds require higher ownership and more consistent deal-leading, which would change Haystack’s collaboration-heavy style. He prefers flexibility—driver, passenger, or backseat—because it preserves access and long-run syndicate dynamics.
- •Bigger funds require materially larger ownership at entry to matter at exit
- •Scaling AUM can force leading every deal and changing the value promise to founders
- •Haystack prefers flexibility: lead/co-lead/meaningful participant—not party rounds
- •Collaboration and repeated games with co-investors are strategic advantages
- •Changing fund size can trigger a chain reaction in dealflow, operations, and outcomes
- 19:33 – 22:09
Ownership floors and seat positioning: how low Haystack will go
Semil shares that Haystack generally won’t go below ~5% ownership at entry, aiming for 5–10% and “earning the right” to maintain it. They contrast US vs. Europe dynamics, where some large firms expect 15%+ and leave less room for smaller investors.
- •Haystack generally stops around ~5% ownership at entry
- •Target posture: 5–10% to start; maintain through rounds if deserved
- •Positioning message: never own more than the founder; be a non-threatening partner
- •US model: invest a round earlier than Accel/Lightspeed/Greylock-style leads
- •Debate on speed: large multi-stage firms can be fast on small checks, but not always
- 22:09 – 26:31
Deployment cadence and investing mistakes: speed, selection errors, and course correction
Semil revisits his earlier “three-year deployment” comment and updates it to 24–30 months for seed. He highlights Fund 4 as a period of concentrated mistakes driven by pushing for ownership—leading to selection errors—and explains how that learning improved subsequent funds.
- •Haystack promises ~24–30 month deployments rather than 36 months
- •Fast deployment makes it harder to return to LPs without deep trust/returns
- •Biggest mistake cluster: Fund 4 (2017–2019) when pushing for ownership
- •Diagnosis: ~80% poor selection vs. ~20% adverse selection
- •LP heuristic: over 5 funds, expect a mix (so-so, a turd, and a spectacular one)
- 26:31 – 36:16
Fundraising lifecycle: finding LPs, building proof, and assembling the right collateral
Semil advises emerging managers to create a track record (often via AngelList) before approaching institutional LPs, and to rely on GP references earned through real work. He outlines practical fundraising materials—one-pagers, decks, data rooms—while arguing reputation and “chatter” matter more than slide polish.
- •AngelList and early investing can create a concrete proof-of-work narrative
- •Institutional LPs are predisposed to pass on Fund I and wait for Fund II
- •GP-to-GP vouching is crucial and must be earned via shared work contexts
- •Collateral stack: one-pager, short deck, data room; reduce friction via templates
- •Key priority: get into LP “chatter zone” through helpfulness, insight, and outcomes
- 36:16 – 41:57
Creating LP urgency and negotiating terms: walking away, preferential deals, and GP stake asks
Semil’s approach to urgency is to set clear timelines and be willing to accept a “no” in various forms rather than waiting indefinitely. They discuss preferential fee/carry terms, selling GP economics, and why any concessions should be tightly scoped to getting off the runway—never perpetual.
- •Urgency tactic: explicit deadlines + assume “not a fit” if silent
- •Preferential terms: Semil avoids them to prevent long-run precedent
- •If needed, concessions can be a one-time “get off the runway” tool
- •GP stake/carry share: can be rational for a high-quality anchor with real leverage
- •Kickers/hurdles: Haystack hasn’t used them; leans toward simple, standard terms
- 41:57 – 47:57
Fundraising mistakes in hindsight: LP psychology, red flags, and managing difficult partners
Semil describes a “minor omen” moment—an LP who simply didn’t engage despite a top-tier intro—and a deeper mistake: assuming LPs who commit will always act aligned with the GP’s success. He shares examples of blocking LPs in future funds due to behavior and unrealistic demands.
- •Early warning: even great intros and track can be met with indifference
- •Major lesson: LPs can have divergent motives even after committing
- •Semil blocked two LPs from future funds due to behavior and term demands
- •Behavioral fit matters; red flags can be a gift if revealed early
- •Recalibrating expectations: don’t assume support is automatic or permanent
- 47:57 – 53:08
LP churn is coming: liquidity mismatches, stranded capital, and shifting demand toward smaller vehicles
Semil predicts increasing LP churn driven by uneven liquidity and limited visibility into portfolio values. He argues some capital in oversized funds becomes “stranded,” pushing LPs to reassess and potentially favor smaller funds where losing $10M hurts less than losing $100M.
- •Churn drivers: idiosyncratic liquidity profiles across LP types
- •Visibility problem: LPs unsure what they own is worth and when it returns cash
- •Stranded capital: stuck in weak companies or funds too large to clear hurdles
- •Institutional pullback from conservatism vs. need to rethink large-fund exposure
- •Downshifting: LPs may prefer smaller vehicles as risk becomes clearer
- 53:08 – 58:46
Who wins in venture next: firepower + early access, barbell outcomes, and the “Chanel vs. Walmart” framing
They debate which fund archetypes thrive: boutique brands with a clear product and giant platforms with lifecycle capital, while “middle” firms risk identity and decision-making bloat. Semil adds that a single breakout deal can rewrite a firm’s trajectory, making long-range predictions fragile.
- •Winners: funds combining early access/ownership with real follow-on firepower
- •Growth of $100–$200M seed funds bridging the seed-to-A gap
- •Harry’s archetypes: “Chanel” (boutique) vs. “Walmart” (walls of cash)
- •Middle-tier risk: unclear brand, many partners, slow decisions
- •One deal can redefine a franchise; outcomes often overpower narratives
- 58:46 – 1:06:45
LP mistakes and better endowment construction: fund-size escalation, bucket traps, and how to allocate smarter
Semil argues LPs’ biggest sin was enabling fund-size escalation and overly flexible blind pools instead of using separate vehicles as governance. They discuss bucket-driven adverse selection and outline how an endowment could build conviction: spend time in market, name-check networks, do directs with experienced help, and join a handful of franchises.
- •LP sin: going along with fund-size escalation and overly broad allocation mandates
- •Separate vehicles for separate strategies act as a governance “governor”
- •Bucket allocation can cause adverse selection (buying what’s available, not best)
- •Endowment playbook: spend real time in SF/NYC, meet networks, triangulate names
- •Portfolio structure: mix early funds, selective franchises, and directs via experienced VC talent
- 1:06:45 – 1:11:50
The ecosystem’s “sins”: VCs lowering checkpoints, founders not respecting capital, and the importance of reporting
Semil lists systemic mistakes: VCs funded too speculatively and reduced true milestone gating; founders accepted money without discipline and failed to keep investors meaningfully informed. He frames the reset as Darwinian: capital efficiency plus integrity, transparency, and respect for the dollar will determine survival.
- •VC sin: lowering financing checkpoints; capital deployed like speculative real estate
- •Result: companies with huge cash balances but no product-market fit
- •Founder sin: lack of self-imposed milestones and weak investor communication
- •Reporting cadence enables advocacy and future fundraising; absence reduces support
- •New era: disciplined capital use + transparent partnerships become prerequisites
- 1:11:50 – 1:24:19
Quickfire: macro worries, underrated angels, hardest moments, biggest wins/misses, and recent investment
In quickfire, Semil shares what worries him most (macro risk + low money supply), highlights underrated angels, and names fundraising as the hardest part of Haystack’s journey. He details his biggest return (HashiCorp from a tiny Fund I check), biggest miss (OpenSea), his favorite advice (“do good deals”), and a recent investment (Impart Security).
- •Macro concern: prolonged “administration of pain” + low M2 money supply
- •Underrated angels: Charlie Songhurst; also praises Scott Belsky
- •Hardest moment: first four fundraises—continuous, pre-AngelList ease
- •Biggest return: HashiCorp (25K into ~$30–35M); Fund I had multiple mega-wins
- •Biggest miss: passing on OpenSea multiple times; lesson on overthinking vs. usage
- •Best advice: Paul Martino—“do good deals”; Recent investment: Impart Security
