Early teams building in web3 or on AI agents usually meet two kinds of first money: a grant that pays for a piece of work, and a pre-seed round that buys part of the company. The useful question is not which one is better but which one fits what you need to prove next, because the two cost different things and leave you in different positions for the money that follows.
For AI-agent teams in 2026 the question comes early: a prototype is quick to build, while showing that it is safe, useful and worth a company still takes real work. We also have an interest here, since we run a grant program and the fund behind it invests in rounds. So here is an honest comparison, including when a grant is the wrong choice.
What each one costs the founder
- Time. A grant costs an application, then reporting against milestones. A pre-seed costs a fundraising process: introductions, calls, follow-ups and diligence. That is often weeks of founder attention that does not go into the product.
- Dilution. A real grant takes no equity and no tokens. A pre-seed does, by definition, whether as shares, a convertible instrument or a claim on future tokens. It also sets a reference point: the next round is priced against it, for better or worse.
- Obligations. A grant obliges you to deliver what you promised and report on it. Some programs add conditions, such as an open-source licence, a deployment on their chain or public reporting. A pre-seed obliges you to the company: investor updates, consent rights written into the documents, and an expectation that you are building something that can grow to venture scale, usually full-time.
The short version: a grant costs you a commitment to a piece of work; a pre-seed costs you a share of the company and a commitment to a trajectory.
When a grant is the better first money
- The idea is unproven. If you cannot yet show that the thing works, a valuation is mostly guesswork, and guesses at this stage rarely favour the founder. A grant lets you prove the core claim first and discuss valuation later, with evidence.
- The work has an infrastructure or public-good flavour. A prover library, a benchmark suite, an agent framework, an open standard. Work like this can matter a great deal without an obvious business model yet, and grants exist largely to pay for it. Forcing it into a pitch deck too early can distort what you build.
- You need one milestone to show. Many teams are a single working demo away from a much better conversation with investors. A grant sized to that demo is the cheapest way to get there.
- The team is not full-time yet. If you are still deciding whether this is the thing, investor money commits you before you know. A grant commits you to a milestone, which is a decision you can actually make today.
When a pre-seed is the better first money
- The team is ready to go full-time. Grants pay for milestones, not for a year of salaries. If the plan needs several people working full-time for many months, grant money will run out between milestones and you will spend your time writing applications.
- The market is clear. You know who pays, why, and roughly how you reach them, and the main risk is execution rather than whether the idea works. That is the risk investors are set up to fund.
- You need runway more than a milestone. Hiring, sustained go-to-market work, a security budget for mainnet: none of these is one deliverable on one date. They are what a round is for.
- Timing matters. If waiting for a grant decision means someone else ships the same product first, dilution may cost less than being second.
Some teams fit both lists. Then the question is order, not choice.
Sequencing: grant first, then the round
For many early teams, the strongest path uses both. The grant pays for the first milestone, the milestone becomes the evidence, and the round is priced on something that exists rather than on a deck. Investors are then pricing a working product and a delivery record, not a plan, and the conversation moves from “will this work?” to how big it can get and what it needs next.
To make this work, choose the grant milestone with the round in mind. Pick the deliverable that removes the biggest doubt an investor would have, not the one that is easiest to build. And read the grant’s terms before you sign: a grant should not quietly commit you to taking the grantor’s investment later, or stop you from raising elsewhere.
Red flags on both sides
On the grant side:
- A grant that asks for tokens or equity. If a “grant” comes with a token allocation, a warrant or shares, it is an investment. That is not automatically bad, but evaluate it as one: what is the price, and what else are you giving up?
- Hidden rights. A right of first refusal on your next round, exclusivity that stops you applying elsewhere, or ownership of your code passing to the grantor. Any of these can cost more than the grant is worth.
- Fees. A real grant program does not charge you to apply, and does not deduct “services” from the amount you were awarded.
On the pre-seed side:
- “Investments” that are really services. If much of the money comes back to the investor as marketing, listing or advisory fees, it is not runway.
- Token terms before there is a token design. Committing a large token allocation before you know what the token does locks in decisions you are not ready to make.
- Pressure to launch early. An investor who wants a token out before the product works is optimising for their exit, not for your project.
- Terms you cannot explain. If you cannot explain each right in the term sheet to your co-founder in one sentence, get advice before signing.
Our FAQ covers the questions founders ask at this point, including whether grants take equity or tokens and whether taking our grant obliges you to raise from the fund afterwards.
Where we sit
ZK Prime Capital is built around the grant-then-round sequence. Our grants are $5K–$75K and non-dilutive, with no equity taken, for idea- and prototype-stage teams working on on-chain AI agents, DeFAI and zero-knowledge infrastructure. Every application gets a reply within 5 business days. Teams that ship their milestone can continue into a $50K–$250K round and a 3–6 month incubation through to token launch; the round is where investment terms come in. So far 60+ teams have been funded with $3.1M. If our focus is not your fit, the Web3 Grants Tracker lists the other programs we follow.
Not sure which one you need yet? Start with one milestone, then make the bigger decision with evidence in hand. Apply with your project ▸