Token warrants and side letters: how equity investors get token exposure.

The problem the instrument exists to solve

Picture a typical structure. An operating company employs the engineers, owns the code and the trademarks, and raises a priced round or a SAFE from venture investors. A separate entity — commonly a foundation or association established offshore — issues the token, holds the treasury, and stewards the network. Both are part of the same project in every ordinary sense of the word.

Now ask what a preferred stockholder in the operating company owns. They own a residual claim on that company: its revenue, its assets, its eventual sale. They do not own the token, because the token was issued by a different legal person that they hold no interest in. If protocol value accrues to the token and the treasury rather than to the company’s income statement, an investor can be right about the project and still capture very little of it.

The token warrant and the token side letter are the market’s answer. They give the investor a contractual route to token exposure, negotiated at the same time as the equity, sized off the equity position, and documented separately because the thing being promised is issued by someone else. That separation is deliberate rather than sloppy — the reasons are set out in /equity/dual-cap-table.

Warrant or side letter

The two terms are used loosely and sometimes interchangeably, which is unhelpful because they behave differently at the moment tokens are issued. The distinction worth holding on to is whether the investor has to do something — pay, elect, exercise — or whether delivery simply happens.

The terms that get negotiated

The following are the clauses that consume the negotiation. Each is a genuine choice with consequences on the token register, and none has a settled market answer.

The covenant chain, and why it is the real risk

Here is the structural weakness that gets least attention in template discussions. The document is usually signed by the operating company. The tokens will be issued by a different entity. A company cannot grant rights over property that belongs to someone else, so what the investor actually receives is generally a covenant: a promise by the company to procure that the issuer delivers, or to use defined efforts to bring it about.

Whether that covenant is worth anything depends on facts outside the four corners of the warrant. Does the company control the issuing entity? Is the issuer a party to the agreement, or bound by a back-to-back arrangement? What happens if the foundation’s council, exercising independent duties, declines? And note the awkward corner this sits in: the more tightly the company can compel the issuer, the less independent the issuer is — which cuts against the very separation the two-entity structure was built to achieve.

That is not a drafting nuisance; it is the same tension the SEC and CFTC described in their joint interpretation of 17 March 2026, which frames a crypto asset as becoming subject to an investment contract where an issuer’s representations or promises to undertake essential managerial efforts lead purchasers reasonably to expect profits, and as ceasing to be subject to one when purchasers can no longer reasonably view those promises as attached to the asset. A structure engineered to guarantee investors token upside from company effort is pulling in one direction; the analysis of the token’s own status pulls in the other.

How it lands on the two registers

A token warrant is easy to lose, because for most of its life it is not on either cap table in a visible way. It does not issue shares, so cap table software has nowhere to record it. It does not consume tokens until it is exercised, so the token register often carries it as a footnote or not at all. Then a token generation event happens and a set of contingent claims arrives at once.

• On the company cap table: normally nothing. The warrant is over tokens, not shares, so it does not appear in the fully diluted share count and does not dilute other stockholders directly.
• On the token cap table: a contingent claim against supply, which should be modelled in the investor bucket from the day it is signed rather than the day it is exercised.
• In the company’s obligations: a covenant that may need disclosure in diligence, and that a later acquirer will certainly ask about.
• In the reconciliation: a list of which investors hold one and which do not, maintained continuously — because the moment a new round opens, the ones who do not will want to know why.

The practical discipline is to model outstanding token warrants as if they were exercised, alongside the contributor and treasury buckets, so the token fully diluted figure is honest. That is the same reflex a good equity operator applies to an unexercised option pool, and it is described in the context of the whole register at /equity/token-cap-table.