How to Design Airdrops Without Attracting Only Speculators

Airdrops should buy durable ownership, not temporary order flow

An airdrop that mainly attracts speculators is usually mispriced treasury spend. It transfers ownership to wallets whose highest-return move is to sell, loop activity across addresses, or farm the next campaign instead of helping the network compound. The strongest recent designs did not rely on one-shot wallet rewards. They used staged budgets, time-weighted participation, explicit anti-Sybil filters, and post-claim alignment mechanics.

Budget the airdrop like treasury capital, not like growth marketing

Airdrop budgets work better when they are treated as capital allocation with hurdle rates, not as a launch-day headline. Arbitrum said outright that its March 2023 design had to balance broad inclusion, meaningful per-user allocations, and reserves for future DAO grants. That is the right frame. If every token is spent in the first event, there is no room to correct mistakes, reward later contributors, or defend the treasury against low-signal demand.

Optimism followed the same discipline at the portfolio level. In its November 30, 2023 budget update, the Foundation listed 816,043,786 OP allocated to user airdrops as a category, but only 245,901,955 OP had been allocated at that point. That gap matters. It means the treasury preserved dry powder instead of assuming the first distribution logic would be optimal.

Jupiter made the treasury logic even more explicit in January 2025. Instead of exhausting the annual pool in one distribution, Jupiter carved 200 million JUP out of a 700 million JUP program for later “Carrots,” including claim-and-stake incentives, appeals, and future growth programs. That is a better fit for uncertain recipient quality. It also reduces the odds that the entire budget gets captured by users who were only present for a single scoring window.

Eigen’s Season 1 structure points in the same direction. Only 5% of initial supply went to Season 1 restakers, with the broader stakedrop category capped at 15% of total supply across multiple seasons. Unclaimed Season 1 tokens were set to be reallocated to future stakedrop seasons instead of leaking into permanent deadweight. From a treasury-risk perspective, this is what disciplined optionality looks like.

The practical rule is simple. Manage the airdrop as treasury capital: budget should be released in tranches that correspond to observable retention, governance, or usage targets. Wallet count is not a treasury KPI. Six-month activity, delegation participation, staking persistence, fee contribution, and product breadth are closer to one.

Reward behavior that is expensive to fake and repeated over time

Airdrops should reward behavior that cannot be rented for a weekend. Optimism Airdrop #2 did this by paying governance rewards based on OP delegated × days and by rebating 80% of gas fees, up to $500 in fees rebated per address, only for users above a minimum usage threshold. Airdrop #3 kept the delegation-days structure and added a bonus for delegating to an address that actually voted onchain. Both choices favor recipients who behaved like future governors, not just future sellers.

Jupiter reached a similar conclusion from a different starting point. The 2025 design added a swap consistency bonus for users who traded in at least 8 months of the prior year. That matters because repeated usage is a much stronger indicator of product fit than a burst of last-minute volume. The same logic appeared on the staking side, where Jupiter used time-weighted stake instead of a single snapshot.

Eigen’s published methodology is unusually clear on this point. Season 1 allocations considered amount, duration, earliness, and loyalty. Native restakers received extra weight, and the Foundation also imposed a 33% concentration cap so that no single LST, LRT, user, or entity could dominate the distribution. That is a more defensible path than flat rewards for every address that touches the protocol once.

Optimism Airdrop #5 shows how far this principle can be pushed on the product-usage side. To qualify as a “Superchain Power User,” an address needed at least 20 unique contracts and a contracts-to-transactions ratio of at least 10% during the eligibility window. That is a strong signal that the user explored the ecosystem rather than cycling one cheap action.

Do not subsidize testnet tourism if the strategic goal is mainnet retention. Eigen explicitly said future testnets, including Holesky, would not carry token allocations and that the goal of testnets is testing and preparation for mainnet. Many teams would save themselves a future farmer problem by adopting that rule earlier.

Score behavior, not just volume, and assume users will optimize against you

Raw volume is a poor proxy for aligned demand. Jupiter published one of the clearest reasons why. In the 2025 design, the top 10% of swap users accounted for 99% of total swap volume, and the top 1% accounted for 97%. A purely linear per-dollar airdrop would have let a very small group absorb most of the distribution. Jupiter responded with tiers, volume normalization, and behavior discounts.

Jupiter then filtered the data the way a treasury team should expect to filter it. It removed wallets that traded for less than 3 weeks, wallets with 50%+ failed transaction rates, and wallets that only ran rapid circular swaps. It ignored swaps under $5 and heavily discounted stablecoin-to-stablecoin and SOL-to-SOL flows because those patterns were common among farmers. The broader lesson is not “copy Jupiter exactly.” It is “assume the scoring model will be gamed and decide in advance which behavior you are willing to pay for.”

Arbitrum made the same move from a different angle. Its initial airdrop used a point system across multiple forms of network usage and deducted points for Sybil-linked usage patterns. It also published the criteria and underlying data, which matters because hidden scoring models without later transparency tend to become credibility problems.

Identity layers help when the pool is large enough to justify the friction. Human Passport was built originally to defend Gitcoin Grants from Sybil attacks. Its model aggregates multiple identity attestations, allows different weights by community, and deduplicates reused credentials in the default scorer. That is useful because no single credential is reliable enough on its own, and one-size-fits-all identity scoring rarely maps well to the actual threat model.

False positives are the unavoidable cost of stronger filters. Jupiter said it identified more than 750,000 wallets that matched its Sybil or bot filters and gave users 3 months to appeal misclassification. That appeal path is not a courtesy. It is part of the design. If a team cannot fund the analytics, appeals, and support burden, the team should narrow the airdrop scope rather than pretend the filter can be both aggressive and effortless.

Underinvesting in pre-distribution filtering pushes the cost downstream. Optimism said it reclaimed and redistributed recovered OP after identifying more than 17,000 Sybil addresses that slipped through Airdrop #1. That is better than ignoring the problem. It is still a reminder that weak front-end screening turns airdrop design into post-launch incident response.

Use claim mechanics to favor aligned holders, but do not pretend the trade-off disappears

Claim mechanics can materially change who stays. Arbitrum delayed claiming for one week after announcement so users could delegate during the claim flow and so a broader set of delegates could nominate themselves first. That is a subtle but important design choice. It moves the token from pure receipt toward governance activation.

Eigen used a much stronger version of the same idea. Season 1 launched with an initial non-transferable period, and the Foundation said the reason was to build social consensus, decentralization, and operational readiness around the token’s novel staking and forking model before enabling transferability. The claim flow itself pushed users to claim, stake, and delegate, and by June 19, 2024, Eigen said more than 88% of claimed EIGEN had been staked.

This is evidence that transfer restrictions and staking-centric claims can reduce immediate speculative exit. It is not evidence that every project should copy them. The trade-off is real. Non-transferability weakens price discovery and can push activity into side agreements or unofficial derivatives. Eigen itself warned that trading derivatives based on non-transferable EIGEN would be detrimental and could affect future-season eligibility.

Jupiter’s 2025 framework shows a softer approach that many teams will find more practical. It allocated 60 million JUP to stakers, weighted rewards by time-weighted stake, gave extra bonuses to the top 10% most active voters, and rewarded “Super Stakers” who had never attempted to unstake as of January 11, 2025. That design does not ban selling. It simply pays more to people who already proved they value governance and persistence.

The treasury answer is usually not to eliminate speculation. A token also needs liquidity, float, and price discovery. The goal is narrower. Make sure speculation is not the only behavior with a positive expected value.

Govern the reserve, publish the rules, and measure the right outcomes

Large discretionary airdrop reserves without governance constraints usually become lobbying pools. Arbitrum handled this correctly by publishing its full recipient list, criteria, and underlying dataset. Jupiter forced a second community vote, raised the passing threshold to 70%, and said the final vote passed with 87% support after incorporating verified feedback. Governance friction is not always bad. Sometimes it is the control that stops a treasury program from drifting into arbitrary favoritism.

The metrics should also change. Airdrop design is usually evaluated on claim count and day-one social reach because those are easy to report. They are weak indicators of whether treasury capital was deployed well. Better metrics are 30-, 90-, and 180-day retention, delegation persistence, stake duration, fee contribution after receipt, breadth of product use, and the share of recipients who remain active without a second subsidy. The examples above do not prove one universal formula. They do show what serious teams measure when they stop optimizing for vanity. Our methodology page covers the broader framework behind those choices.

At FinDaS Tokenomics, this is the recurring error we see in tokenomics design work. Teams treat the airdrop as a user-acquisition line item when it is really a transfer of scarce treasury assets and future governance power. That framing change usually improves the entire design. It forces budget caps, staging, explicit success metrics, and tighter thinking about who should own the token in the first place.

• Reserve part of the pool for later seasons, appeals, and corrective action.
• Reward time, breadth, and repeat behavior instead of isolated transactions.
• Discount or exclude behaviors with obvious farming signatures.
• Use identity or Sybil-resistance tooling when the pool is large enough to justify the friction.
• Favor delegation, staking, or other alignment mechanics, but keep enough liquidity for a real market.
• Publish the rules early, publish the outcomes later, and let governance constrain discretionary reallocations.

If the only profitable action after receiving the token is to sell it, the airdrop was never a community strategy. It was an unsecured distribution of treasury assets.