Raising Without Romance: SAFEs, Token Warrants & VC Term Sheets in 2026

Overview

We've written about VC before — an ode and epitaph to it back in 2021, and a review of the founder/investor relationship in 2024. This guide takes the less subjective route: what the standard instruments actually say, what the current market data shows about how they're used, and where the ground has shifted since we first covered this in early 2025.

A. How to Safely Use a SAFE

SAFEs (Simple Agreements for Future Equity) are now the default fundraising instrument at pre-seed and seed, and the numbers back that up hard. According to Carta's State of Pre-Seed: Q1 2026 report, SAFEs accounted for roughly 92–93% of pre-seed rounds, while convertible notes fell to a record low of 7–8% of instruments and dollars. By Q2 2026, Carta's follow-up report clocked the average pre-seed instrument size at $276,000 — a 27% year-over-year jump and the highest figure in over four years, concentrated in fewer, larger checks going mostly to AI companies (49% of all pre-seed dollars in H1 2026).

If you're raising pre-seed or seed capital today, you will most likely be using a SAFE — possibly combined with a Token Warrant if your project plans to issue a native token down the line (see Section B).

In essence, a SAFE gives an investor the right to convert into shares once your company's valuation is actually set, at a future "priced round." Investors get visibility on roughly how many shares they'll end up with, without anyone having to agree a valuation on day one.

Y Combinator's four SAFE flavors

Y Combinator, which created the SAFE and still runs the most widely copied template, offers four standard variants:

Variant How it works
Valuation Cap only Investor gets a hard ceiling on the conversion valuation, no discount. The most common variant.
Discount only Investor gets a fixed discount (typically 15–25%) off the price of the next priced round, no cap.
Cap + Discount Both provisions combined — investor converts at whichever is more favorable.
MFN only The most founder-friendly variant: no cap, no discount, just the right to inherit better terms if you issue a more investor-favorable SAFE later.

YC's own standard deal for accepted companies — $500,000 total, split as $125,000 for a fixed 7% (post-money SAFE) plus $375,000 on an uncapped MFN SAFE — has been unchanged since 2022 and remains identical across every batch through 2026. It isn't negotiable, and it's a useful anchor for what "market" looks like even if you're not going through YC.

The visibility of post-money SAFEs comes at the price of more founder dilution

The easiest way to give SAFE investors clarity on their eventual ownership % is a "post-money" SAFE — the dominant structure in the market today. A post-money price cap is a predetermined maximum valuation at which the investment converts, calculated after accounting for all outstanding SAFEs. That means:

Fixed ownership percentage — investors lock in a specific % relative to other shareholders, including other SAFE holders.
Clarity for both sides — everyone can calculate ownership percentages before the priced round that triggers conversion.
Founder dilution risk — because SAFE holders' percentages are fixed, founders absorb more of the dilution in future priced rounds than they would under a pre-money structure.

Example: an investor puts in $1M at a $10M post-money cap. They lock in 10% ownership on conversion — even if the company eventually prices at $12M — because SAFE holders convert at whichever is lower, the cap or the actual round valuation.

By contrast, a pre-money SAFE defers the valuation question to a later round, which creates more uncertainty about final ownership — a problem that compounds with every additional pre-money SAFE stacked on top.

Other SAFE terms worth reading closely
Conversion triggers — the specific event(s) that flip the SAFE into equity, usually a priced round or a liquidity event. Ask what happens if neither ever occurs.
Refund clauses — be wary of terms requiring you to return invested funds if no Financing Event happens; that money is usually long spent by then.
Financing Event deadline — aim for a realistic runway (6–18 months) so you're not back in market too soon.
Participation rights — may let investors join future rounds to maintain their %.
Lock-up periods — may restrict founders from selling shares for a period.
Representations and warranties — the company must confirm the SAFE is properly authorized and doesn't conflict with existing obligations.
Governing law / dispute resolution — specifies how conflicts get resolved.
Assignment restrictions — limits on transferring SAFE rights to third parties.

B. A Warrant for Tokens

When a Web3 project raises through a SAFE, investors typically expect a token allocation on top of it, if the project intends to launch a token later. The standard structure for this — now the preferred deal shape for pre-seed/seed crypto VC deals — pairs a SAFE with a separate Token Warrant.

The Token Warrant is a distinct legal instrument from the SAFE (and separate from a SAFT — more on that below). It gives the holder the right, but not the obligation, to receive or purchase a set number of tokens at a predetermined price or condition, usually within a set window. Mechanically:

Issuance — the warrant specifies the token quantity, exercise price (if any), expiration date, and conditions for conversion.
Triggering event — conversion activates when the holder exercises their right, or when a predefined condition is met (manual exercise or automatic conversion).
Token amount — set by the warrant's formula: a fixed ratio, a market-value-based calculation, or — most commonly — a "1-for-1" allocation, where the % equity a SAFE holder receives is mirrored with an equal % of the token supply for a nominal payment.

Because most SAFEs today are post-money, investors can calculate their expected token % fairly easily — the open question is usually what pool of tokens that % applies to (total supply vs. tokens actually in "free float" at conversion, which is the more founder-favorable definition to negotiate for).

If the warrant isn't exercised by expiration, or the conditions aren't met, it typically expires worthless — think of it as a stock warrant adapted for digital assets.

Token Warrant vs. SAFT — these are not the same thing

A SAFT (Simple Agreement for Future Tokens) is a different, older instrument that promises tokens directly, without an equity component. SAFTs sit on considerably shakier legal ground in the US: the SEC hasn't issued formal guidance specifically on SAFTs, but its enforcement history signals real skepticism toward the "future tokens = not a security today" theory, and access is generally restricted to accredited investors. A Token Warrant attached to a SAFE avoids some of that exposure by tying the token right to an equity instrument rather than presenting it as a standalone token sale — but it doesn't eliminate securities-law risk entirely, and structuring this correctly is exactly the kind of thing to get counsel on before you send a term sheet.

The split personality of projects with both equity and tokens

In Web3 projects with both stockholders and token holders, revenue flow is a hybrid: it blends centralized corporate structures with decentralized tokenomics, and the exact path depends heavily on the project's design.

Revenue generation — typically transaction fees, service fees, subscriptions, or partnerships, flowing into the project's treasury or operating entity.
Stockholders' share — tied to the centralized legal entity: dividends, buybacks, reinvestment, equity value, and voting control.
Token holders' share — tied to the decentralized side: staking rewards, profit-sharing, buy-and-burn mechanisms, ecosystem incentives, and governance power.
The split — how revenue divides between the two groups isn't always clear-cut, and depends on legal structure, tokenomics design, and regulatory pressure.
In practice — Uniswap runs fully DAO-driven with no stockholders; Sky Mavis (Axie Infinity) runs both stockholders and token holders side by side; Layer 1s like Ethereum have no stockholders at all, just early backers holding tokens.
C. Ready for a Priced Round — the VC Term Sheet

At a priced round, every SAFE investor converts into stock using the SAFE's conversion formula (usually post-money terms). If the company's actual valuation lands below the SAFE cap, the cap — not the round price — sets the conversion math, with Employee Share Option Plans also entering the picture at this stage.

A useful way to read any term sheet is to split its clauses into two buckets: economics (who gets paid what, when) and control (who gets to decide what). Most of the outcomes founders regret — no upside left, or getting pushed out mid-flight — trace back to how these two categories interact once they're written into the long-form Stockholder Agreement.

Clause What it decides Market standard in 2026
Valuation (pre-/post-money) What your company is worth before/after the investment Negotiated per deal; higher isn't automatically better if it sets up a down round later
Liquidation preference Who gets paid first, and how much, in a sale or wind-down 1x, non-participating is the market standard under the NVCA model documents; "participating" preferred lets investors double-dip and should be pushed back on
Board composition Who sits on, and controls, the board Investors often want a seat; keep the board advisory, not a body that can override you on operations
Vesting schedule How founders "earn" their own equity over time 4-year vesting with a 1-year cliff is the near-universal standard: 25% vests at the 1-year mark, then the remaining 75% vests monthly over the following 36 months
Anti-dilution protection How investors are shielded from a future down round Broad-based weighted average is the common, founder-friendlier standard; full ratchet is the harshest version and worth resisting
Option pool Equity set aside for future hires Typically 10–20% of fully diluted shares; push for it to be sized post-money or kept smaller, since a pre-money pool dilutes founders more than investors
Control rights (protective provisions) Investor veto powers over major company decisions Keep the list short enough that it doesn't functionally hand over operating control
Drag-along rights Whether investors can force all shareholders to sell A high approval threshold (around 75%) with founder input is the more balanced version
Pro-rata rights Investor's right to invest in future rounds to maintain %age Common; consider capping it or making it optional rather than automatic
No-shop clause How long you're locked into exclusive talks with one investor 30 days is tight and standard; 60 days starts to work against you

These clauses aren't just legalese — they set your leverage, your payout, and how much autonomy you keep. The investor's goal is to de-risk their bet; yours is to keep enough skin in the game that the outcome still motivates you. Get a lawyer who has actually negotiated these before — ideally one not primarily retained by VCs — and don't be afraid to push back, especially on valuation, liquidation preference, and control rights.

How to Actually Negotiate This

If you're choosing between a SAFE and going straight to a priced round — a SAFE is almost always faster and cheaper for pre-seed/seed; save the priced round for when you have enough traction to actually negotiate valuation from a position of strength.

If you're adding a Token Warrant to a SAFE for a Web3 raise — get the "free float" definition and the conversion ratio nailed down in writing before you get to signature; this is where founders lose the most ground without realizing it.

If someone hands you a SAFT instead of a SAFE + Token Warrant — treat that as a flag to get counsel involved before proceeding; the legal footing is meaningfully less settled.

If a term sheet shows up with full-ratchet anti-dilution — push for broad-based weighted average instead; it's the market standard for a reason and full ratchet can wipe out founder ownership in a down round.

If the option pool ask feels large — check whether it's being sized pre-money or post-money; a pre-money pool effectively comes out of the founders' side of the cap table, not the new investors'.

If the no-shop clause runs past 30 days — negotiate it down; a long exclusivity window mostly benefits the investor's optionality, not yours.

None of this replaces a lawyer who's actually read your specific documents. If you want a second set of eyes before you sign anything, that's exactly what the call above is for.

FAQs