A blockchain grants program produces builders when it funds a written list of gaps, pays against milestones, and measures what is still running after the money stops. Open rounds without that discipline draw farmers: Optimism's third retro funding round took 1,594 applications, and its own review named spam and airdrop-farming entries as a problem.
Who decides this
The head of ecosystem owns the program: what it funds, how much, and whether the chain is better off a year later. The grants committee owns individual decisions and answers for each one in writing. The foundation council sets the budget and should see a retention report, not a list of grants awarded.
Should we run open applications or requests for proposals?
Run most of the budget through requests for proposals and a published wishlist, and keep open applications small. Open applications reward whoever writes the best form, not whoever fills the gap your chain actually has.
Two of the best-known programs have moved this way. The Ethereum Foundation's Ecosystem Support Program paused open applications in 2025 and relaunched in November with two routes: a wishlist of broad themes and RFPs with defined scope, deliverables and timelines. Its stated reason was that rising application volume strained a lean team and crowded out strategic work.
The Solana Foundation lists three funding types: milestone-based grants for public goods, milestone-based convertible grants for public goods with a commercial side, and RFPs for ideas it wants built. Projects that are mainly commercial generally do not qualify for a standard grant.
What we would do: publish ten to twenty named gaps, each with an owner on your team who can judge the work. Leave a small open track for ideas you did not think of, capped per grant.
Which grant instrument fits which goal?
Match the instrument to what you are buying: shared infrastructure, a specific product, or usage. Each instrument draws a different applicant, and the wrong one attracts farmers by design.
| Instrument | What it buys | Grant-farming risk | Committee workload | Use it for |
|---|---|---|---|---|
| Open application grant | Ideas you did not request | High: rewards application-writing | High: every form needs review | A small, capped exploratory track |
| RFP or wishlist grant | A named gap with defined deliverables | Lower: scope is set by you | Medium: fewer, better-fit proposals | Tooling, infrastructure, missing primitives |
| Convertible grant | A commercial product that may raise later | Lower: converts only if the team succeeds | Medium: needs a legal agreement | Teams likely to build a business on your chain |
| Usage or liquidity incentive | TVL, volume, users for a set period | High: capital leaves when rewards end | High: needs data and monitoring | Short, measured pushes in new markets only |
| Retroactive funding | Work already delivered | Medium: lobbying and self-reported impact | Very high at scale | Rewarding public goods after the fact |
The convertible grant deserves more use than it gets. Solana's foundation introduced it in 2023 so it could back closed-source or for-profit projects; the grant becomes an investment only if the project hits agreed milestones, such as a priced funding round. A farmer gains little from it. A real company gets aligned capital.
How big should the program be, and how fast should it spend?
Size the program from the gap list, then release money against milestones over six to nine months, not in one allocation. Budget set from the treasury balance tends to be spent on whoever is in the room.
For scale, the Ethereum Foundation reported $32.6 million across 101 grants in the first quarter of 2025, as covered by one crypto publication citing the foundation's own update. Few chains need that. In our view, most need fewer, larger, scoped grants.
Incentive programs show why the payout schedule matters. Arbitrum's STIP round allocated 50 million ARB to about 30 protocols from November 2023 to March 2024. One efficacy analysis found the chain's TVL rose by more than $1.31 billion over that window. Grantees grew TVL by 118.2%, against 35.6% for protocols outside the program. It did not measure what stayed.
A later review for the Arbitrum DAO by two research teams did. It found TVL, users and volume typically rose during incentives and fell back toward pre-program levels afterwards, sometimes within weeks. It recommended stating each program's purpose up front, longer timelines where product-market fit exists, and milestone-gated payouts instead of one-time allocations.
Who should sit on the grants committee, and how should it decide?
Use a small committee of reviewers who know the domains you fund, pay them for their time, and require a written rationale for every decision. Large volunteer panels drift toward popularity.
Optimism's retro round 3 shows the cost of scale. Its own review found the median badgeholder spent about 16 hours on the round, which it called unsustainable. The 17-vote quorum led applicants to lobby voters, which some participants saw as a popularity contest. Impact metrics were self-reported and often vanity numbers.
What we would do: five to seven reviewers, each assigned to two or three wishlist areas. A conflicts register, updated before every round. Decisions and reasons published. Any reviewer with a stake in an applicant steps out of that decision, in writing.
How do we tell builders from grant farmers?
Fund teams with a track record you can check, pay against deliverables you can verify, and judge each grant on usage measured after payments end. Farmers are good at applications and bad at shipping.
Sybil behaviour is the extreme case. Gitcoin's grant rounds were targeted by coordinated fake donors gaming quadratic matching. Gitcoin answered with identity checks, a machine-learning model to flag accounts, and human review before sanctions. Its stewards also stopped paying a "fraud tax" that had made grantees whole, removing the attackers' payoff.
Reporting is the quieter failure. One independent newsletter on Arbitrum's STIP, citing a DAO forum analysis, reported that around 35% of recipients had not posted a final report. The DAO review above also found that over 37% of one later incentive round went to fee rebates and points programs, which it linked to weak retention.
Signals we would require before a second tranche:
- Shipped code or product that predates the application, in a public repository or live on another chain.
- An accountable entity that signs the grant agreement and can be held to it.
- Verifiable milestones, preferably on-chain: contracts deployed, integrations live, users transacting.
- A 90-day post-grant report on usage, with the final tranche held until it arrives.
What should the council see each quarter?
The council should see what grantees have shipped and what is still in use, not how much was awarded. A grants report that lists commitments tells the council how fast money left, not what it bought.
The Arbitrum review made the same point from the other side. It recommended one standardised, centrally run reporting process instead of leaving data collection to each grantee. It also recommended a dedicated unit with authority to rebalance or pause funds when results fall short, and measuring success after incentives end, not only during the funded window.
What we would put in front of the council each quarter:
- Gap coverage: which wishlist items are funded, which are delivered, which have no credible applicant.
- Delivery: milestones met, missed and renegotiated, by grant.
- Survival: grantees still shipping on the chain 6 and 12 months after their last payment.
- Cost per surviving project: total paid divided by grantees still active, by instrument.
- Pauses: grants stopped or clawed back, and why.
If a whole category of grants keeps failing the survival test, close it and move the budget. That is a council decision, and the head of ecosystem should bring it with the data.
What goes wrong: distribution left to chance
The common failure is a grants program with a budget and a form, but no plan for which builders the chain needs. Capital then goes to whoever applies best and lobbies hardest, not to the gaps that limit the ecosystem.
It looks like this. The foundation announces a large ecosystem grants pool. Applications flood in, many from teams who apply to every chain's program each season. The committee, short of time, funds the most polished proposals and the loudest community favourites. A year later, most grantees have moved on to the next program, the dashboards that justified the grants have fallen back, and the chain still lacks the oracle, the wallet integration or the lending market it needed.
Nobody decided where the money should go, so the applicants decided. The fix is upstream: write the gap list, assign an owner to each gap, and make every payment conditional on something that stays.
How Alcyone approaches it
Alcyone starts a chain engagement from the gap list and the owner for each gap, then plans the launches, partnerships and onboarding that fill it. Alcyone's ecosystem advisory case work includes a major Layer-1 whose first token launch on the chain brought 5,000+ new holders. Engagements begin with a 30-minute call with a senior partner and a fixed-scope, fixed-fee assessment credited against further work.