Predict.fun's Airdrop Pays 35 Cents Today or a Dollar Next Year. The Other 65 Cents Has No Address.

At 09:00 UTC on Thursday, October 29, a form inside Binance Wallet closed its first deadline. Holders of Predict Points, the activity score that Predict.fun handed out to the traders who pushed its BNB Chain prediction market past $1.8 billion in cumulative volume, had to pick one of two boxes. Box A: claim 35% of your PREDICT allocation now, and the remaining 65% is "automatically forfeited and non-recoverable." Box B: stake for a year and receive all of it. One submission per person. No edits. Anyone who never fills in the form by November 26 gets Box B by default.

The pool at stake is 70 million PREDICT, 7% of a fixed one-billion supply, carved out of the ecosystem and community allocation. Predict.fun has framed the split as a choice. Most coverage has repeated the framing.

It isn't a choice. It's a price. Sixty-five percent, charged to the people who used the product, for the right to sell what they earned.

The airdrop menu is a liquidity fee, and the only people paying it are the users.

Start with who else is getting tokens around this launch, and on what terms. The same tokenomics post that set the airdrop rules put 26.05% of supply with early supporters and 10% with core contributors, all of it locked at launch behind a 24-month cliff and then 36 months of monthly vesting. That is a long schedule, and Predict.fun deserves credit for it. Insiders cannot sell anything for two years.

Then there is the money coming in the side door. Binance Wallet ran a Pre-TGE Prime Sale on October 29, capped at 6 BNB per person and open to holders of Binance Alpha points. Buyers claim a "PREDICT Key" after the sale, and the tokens are airdropped into their Binance Alpha accounts on TGE day. The announcement lists no price, no valuation, no number of tokens on offer and no lockup for those buyers.

So three groups touch PREDICT this week. Insiders wait two years and lose nothing. Sale buyers pay in BNB on terms that have not been published. And the traders whose $1.8 billion in volume is the entire reason the token has a story are the only group whose allocation shrinks by two-thirds if they want it liquid.

The 65% doesn't vanish. It goes somewhere, and the rules don't say where.

Here is the part that should bother anyone who reads token contracts for a living. Binance's version of the terms calls the forfeited share "non-recoverable." The TechFlow write-up republished by KuCoin says it is "automatically reclaimed or destroyed." TokenPost and CoinGabbar both describe the forfeit and neither says where the tokens land.

"Reclaimed" and "destroyed" are opposite outcomes. Destroyed means the supply shrinks and every remaining holder owns a slightly larger slice. Reclaimed means the tokens flow back into the ecosystem and community pool, which at 50.95% of supply is already the largest bucket on the chart and is controlled by the project. Under the first reading, impatient users subsidize patient ones. Under the second, they subsidize the treasury.

The arithmetic is not trivial. If every Season 1 recipient took Box A, 24.5 million PREDICT would go out and 45.5 million would be forfeited. That is 4.55% of total supply changing hands from users to an address nobody has named. For a sense of scale, the entire initial liquidity pool is 5% of supply.

A burn is a transaction. It has a hash. A return to a treasury is also a transaction, to a specific contract. Either one could be published before the first claim. Neither has been.

Even the basic numbers arrived with two versions.

The forfeiture is not the only loose thread. On launch day, a widely shared KuCoin flash item headlined a 10 billion total supply while the BlockBeats report underneath it said one billion. The 7% airdrop figure only works at one billion, so the smaller number is almost certainly right. But it took a reader doing division to settle it.

Box B has the same problem. Binance's terms describe it as receiving 100% after one year of staking, which reads like a single cliff at month twelve. The tokenomics write-up describes it as a linear monthly unlock over twelve months. Those are very different products. One pays you nothing for a year. The other pays you roughly 8% a month starting next month. CoinGabbar adds a third wrinkle, reporting that users who choose to vest can switch to the instant claim within one month of TGE, something Binance's own form rules, which forbid edits, do not mention.

A user deciding between Box A and Box B on Thursday morning was being asked to price an option without knowing its expiry, its payout curve or whether it could be exercised early.

Concede the design's point: forced patience is better than a dump.

The strongest defense of the menu is a good one. Airdrops have a long record of turning into sell orders on day one, because the people farming points are often the people least attached to the product. A plain unlocked drop of 70 million tokens into a market whose 16% launch float is mostly earmarked for market making and DEX depth could swamp that liquidity in hours. A 65% haircut on early exit is a blunt but effective way to make farmers think twice, and the default-to-staking rule protects users who simply never open the form.

It is also more disciplined than some of what launched alongside it. Cambria, a browser MMO going live on Base the same day, made its 750 million RSGP airdrop, 15% of supply, fully unlocked at TGE with no lockup. Predict.fun is at least trying to stop the day-one rush.

And there is a fair question about how much of the volume came from people who deserve protection at all. When a rival YZi-backed venue, Opinion, posted nearly $1.5 billion in weekly volume weeks after launch, one founder told Cointelegraph its growth looked engineered rather than organic. Points programs on BNB Chain prediction markets attract wash trading. A design that punishes the quick exit punishes the wash traders hardest.

All true. None of it explains why the penalty's destination is unpublished. Discouraging dumps is a reason to charge the fee. It is not a reason to keep the recipient vague. If the goal is supply discipline, a burn to a dead address, posted in advance, proves it. If the goal is to refill the community pool for future seasons, that's defensible too, but then users should know they are funding Season 2 with their Season 1 rewards. And if sale buyers face no comparable lockup, then the anti-dump logic applies to the users who earned tokens and not to the ones who bought them, which is backwards.

This is the gap that on-chain locks were built to close. A vesting contract or token lock with a public address lets anyone check, at any time, how much is locked, for whom and until when. Predict.fun's insider schedule is generous on paper. Putting it, the sale terms and the forfeiture destination on-chain where they can be verified would make it generous in fact.

The first batch of Box A claims goes out on launch day. Weekly batches follow on November 6, 13, 20 and 27. Every one of them will forfeit tokens to somewhere. The project can name the address before the next batch, or its users can find it themselves on a block explorer after the fact.