.png&w=3840&q=75&dpl=dpl_AJqEjM5WAxjyRcSd6J56dfeSPSFt)
How to Start a Seed Fund
To start a seed fund you form the standard three-entity structure — a Delaware fund limited partnership, a GP LLC, and a management company LLC — then write portfolio construction that survives arithmetic: how many companies, at what check size, targeting what ownership, with what share of the fund held in reserve for follow-ons.
Seed is defined by stage rather than fund size, which is what separates it from a micro VC fund. A $75 million seed fund and a $15 million one both write first checks into pre-revenue companies; they differ in position count and reserve depth. LPs underwrite seed funds almost entirely on portfolio construction and sourcing, because at this stage there are no financials to diligence.
The Portfolio Math Decides the Fund Size
Seed returns follow a power law: a small number of positions produce nearly all the return, and the fund has to be constructed so that one outcome can return it. If your largest realistic exit cannot return the fund on its own, the portfolio is mis-sized regardless of how good the companies are.
Work it backwards. A $20 million fund writing $400,000 first checks into 30 companies deploys $12 million, leaving $8 million — 40% — for reserves. If a winning company exits at $500 million and you hold 4% after dilution, that single position returns $20 million. That arithmetic either works or it does not, and LPs will run it themselves.
- First check size and the ownership it buys at your target entry valuation — commonly 5% to 12% at seed.
- Position count: 25 to 40 is the common band, since concentration below 20 raises single-company risk and above 50 dilutes your best outcomes.
- Reserve ratio: 30% to 50% of the fund, with a written rule for which companies get follow-on capital.
- Expected dilution from seed to exit — typically 40% to 60% across subsequent rounds — applied to your entry ownership.
- Graduation rate: what share of the portfolio you expect to raise a Series A, which is the metric seed LPs track most closely.
Sourcing Is the Only Real Edge at Seed
Every seed fund claims access to great founders. The ones that raise describe a mechanism: an operator network in a specific domain, a community they built, a technical reputation that makes them the first call, or a geography institutional funds do not cover.
Support the claim with data from your own pipeline. Review the last 50 opportunities that reached you, mark which fit the thesis, and record how each arrived. If the answer is mostly inbound from a network you genuinely own, that is an edge. If it is mostly demo days and warm intros available to everyone, LPs will price it as no edge at all — and at seed, where the check is the commodity, edge is the entire pitch.
Reserves Are Where Seed Funds Quietly Fail
The decision that separates good seed managers from average ones is not which companies to fund first — it is which to follow on. A fund that reserves nothing loses its ownership in the winners through 3 or 4 rounds of dilution. A fund that reserves 50% and deploys it evenly across the portfolio has spent its best capital defending companies that will not return it.
Write the reserve rule before the first close: what triggers a follow-on review, the maximum any single company may receive, who approves it, and when unused reserves may be redeployed into new positions. LPs ask for this specifically because it is where discretion does the most damage.
How Fund Launch Builds It
The Fund Builder models the portfolio construction — fund size, check size, position count, ownership targets, reserve ratio, dilution assumptions, and the management fee and carry that sit on top — so the arithmetic an LP re-runs produces the numbers you showed them. Scroll Deck turns it into an LP-facing deck, and Legal Canvas prepares the formation and offering documents with your reserve policy and concentration limits carried as strategy disclosures for independent counsel review.
Because the deck, the model, and the partnership agreement come from one record, the ownership target on your slide is the one in the documents.
Typical Terms
Ranges typical of first-time seed funds. Venture terms differ from real assets — preferred returns are rare and fund lives are long.
| Term | Typical range | Notes |
|---|---|---|
| Management fee | 2.0% – 2.5% | Often stepping down after the 5-year investment period. |
| Carried interest | 20% | 25% to 30% for managers with a genuinely exceptional track record. |
| Preferred return | None | Unusual in venture — the return profile makes a hurdle largely meaningless. |
| Fund term | 10 years | Plus 2 one-year extensions. Seed outcomes need every bit of it. |
| Position count | 25 – 40 | Sized so one outcome can return the fund. |
| Reserve ratio | 30% – 50% | With a written rule for which companies receive follow-on capital. |
Typical ranges observed across private funds of this type. Actual terms depend on strategy, track record, and LP negotiation — treat these as orientation, not advice, and confirm your structure with counsel.
What LPs Will Ask
Can one company in this portfolio return the whole fund?
If your ownership after dilution multiplied by a realistic best-case exit does not clear the fund size, the construction is wrong. This is the first calculation a venture LP runs and it takes them 30 seconds.
Show me your last 50 opportunities and how each one reached you.
Sourcing is the only durable edge at seed. Pipeline data distinguishes a real network from a manager who attends the same demo days as everyone else.
What is your reserve rule, and who approves a follow-on?
Reserve discretion is where seed funds destroy value quietly. A written trigger, a per-company cap, and a named approver show the decision is governed rather than emotional.
What ownership do you target, and what do you model for dilution?
Entry ownership of 8% means little if you model no dilution. LPs expect 40% to 60% erosion by exit and will check whether your returns assume otherwise.
What is your graduation rate to Series A?
It is the earliest observable signal that your picking works, long before exits. Managers with prior angel or SPV activity can usually produce it.
How are you funding the management company for 10 years?
2.5% on a $20M fund is $500,000 a year before the step-down, against a decade of operations. LPs want to know the firm survives long enough to manage the portfolio.
Frequently Asked Questions
What is the difference between a seed fund and a micro VC fund?
Stage versus size. A seed fund is defined by investing at the seed stage and can be $100 million; a micro VC fund is defined by being small, commonly under $25 million, and may invest across stages. They overlap often, which is why both terms get used loosely — but LPs underwrite the portfolio construction, not the label.
How small can a first seed fund be?
Funds of $5 million to $15 million get raised regularly, usually from individuals and family offices. Below roughly $10 million the management fee cannot support a full-time team — 2.5% of $10 million is $250,000 a year — so the manager is typically part-time or subsidising the firm personally. That is workable for Fund I and not sustainable beyond it.
Do I need prior investing experience to raise a seed fund?
You need something checkable. Angel investments with your own capital, SPVs you syndicated, or a deal record from a prior fund role all work. Operating success alone is weaker than founders assume — being a good founder and picking good founders are related but distinct skills, and sophisticated LPs know the difference.
Should I take board seats?
At seed, usually only where you are leading and the company wants it. Board seats consume time that scales badly across 30 positions, and they create fiduciary obligations to the company that can conflict with your fund. Many seed managers take observer rights instead and reserve formal seats for the largest positions.
What returns do seed funds target?
Seed managers commonly underwrite 3x to 5x gross on the fund, driven by a small number of outsized outcomes rather than portfolio-wide performance. These are underwriting targets, not promises, and the offering documents must present them as assumptions with the drivers visible.
Related Guides
How to Start a Venture Capital Fund
What it takes to start a VC fund: the three-entity structure, the documents, why VC terms skip the preferred return, the ERA path, and real costs.
How to Start a Micro VC Fund
Starting a micro VC fund under $25M: portfolio construction math, running on a small fee, the 250-investor qualifying VC fund variant, and typical terms.
How to Start a Rolling Fund
How rolling funds work: quarterly subscriptions, the 506(c) advantage, why reserves are harder, and when a traditional closed-end fund is the better structure.
How to Start a Solo GP Fund
The solo GP fund model: decision speed as the edge, key-person risk as the LP question, succession provisions, outsourced operations, and 506(c) raising.
GP Commitment: How Much Skin in the Game LPs Expect
What LPs expect a GP to put into their own fund, why cash beats a fee waiver, how the commitment is funded over the call schedule, and what first-timers get wrong.
The First-Time Fund Manager's Guide to Launching
How first-time fund managers actually launch: proving the strategy, sizing the fund to your LP base, terms that close, the document step, and honest raise math.
How to Start a Venture Secondaries Fund
Structuring a venture secondaries fund: LP interests versus direct positions, pricing against stale marks, transfer restrictions, and the shortened J-curve LPs buy.
Build Your Seed Fund
Set the check size, position count, and reserve ratio once — Fund Launch carries the arithmetic into the model, the deck, and the offering documents.
Start building your fundThis guide is educational material, not legal, tax, or investment advice. Fund Launch is not a law firm and does not provide legal advice; fund formation documents prepared on the platform are reviewed by independent counsel before use. Consult your own advisors about your specific situation.