Play-to-earn crypto games after the bubble

Play-to-earn crypto games promised something genuinely new: a game that paid you instead of charging you. For a period in 2024 the tap-to-earn version of that promise was the largest consumer crypto phenomenon in the world by user count, and it required no download, no wallet and no purchase. Then, within one quarter, most of the audience left. This page is a retrospective rather than a list, because the evidence supports a retrospective and does not support a list.

That is a deliberate editorial choice and worth stating plainly. We could not verify current earnings, active player counts or token economics for any individual game from primary sources. The projects named below appear only as retention statistics from published reporting, and none of them is a recommendation, under the sourcing rules in ourreview methodology. What we can do is explain the mechanism that produced the outcome, because the mechanism has not changed and will produce the same outcome again.

The model, stated honestly

A play-to-earn game pays rewards in its own token. Those tokens come from a pre-allocated supply, an emissions schedule, or both, and the value of the payout depends on somebody buying the token from the players who sell it. In a growing market that works: new players arrive, demand exceeds the emissions, and early participants are paid in an appreciating asset. The reward feels like a wage.

It is not a wage. A wage is funded by revenue from customers who receive something they value. These rewards are funded by a token budget and by the inflow of the next cohort of players. That is a subsidy with an end date, and the end date is whenever growth stops. The structure is not unique to games — the Anchor protocol advertised 19.45% on UST with an income base nowhere near the payout obligation, and the shortfall was covered until it could not be, asthe 2022 record shows. Games differ only in that the subsidy buys attention rather than deposits.

There is one reliable test. Ask what the game sells to anyone who is not trying to earn from it. If the answer is nothing, then every unit of reward paid out must ultimately come from another participant's capital or from a treasury that is being depleted. That is not a moral judgement; it is an accounting identity, and it predicts the shape of what follows.

What the retention data shows

The clearest sector-level series is daily active wallets on TON, because the largest tap-to-earn games settled there. That figure exceeded one million in September 2024 and fell below 500,000 by December 2024. Over broadly the same window, Google Trends interest in the phrase "tap-to-earn" fell around 80% from its June 2024 level. Attention and usage declined together, which rules out the comforting explanation that players simply stopped talking about a format they were still using.

The individual project numbers point the same way. Hamster Kombat lost more than three million users in a single month, Notcoin 428,000 and Catizen 407,000 over the same period. These are cited here as measurements rather than as verdicts on the teams involved; several of them executed distributions at a scale almost nobody in consumer software achieves.

The behavioural data explains the timing. Per Chainplay, 17.8% of players quit after their first airdrop and 50.6% reduced their activity. The reward was not a retention device; it was a completion signal. Animoca figures put retention at roughly 60% of peak daily active users for the strongest projects and about 25% for weak ones, which means that even the successes lost a substantial minority of their audience and the typical project lost most of it.

What the format did achieve

It is worth being fair about this, because a retrospective that only records the decline misses why serious people built in the category. Tap-to-earn solved a distribution problem that had defeated crypto for a decade. Onboarding normally requires a download, a wallet, a seed phrase, a funding step and a tolerance for irreversible mistakes, and each of those steps loses most of the people who reach it. A Telegram mini-app removed all five at once, and tens of millions of people who had never held a token ended up holding one.

The mistake was treating that reach as retention. An audience assembled by the promise of a distribution is optimised for receiving the distribution, and the survey data showing that almost a fifth of players left immediately after their first airdrop is the measurement of exactly that. The lesson generalises beyond games: any product that acquires users with a token is acquiring users who value the token, and when the token arrives the reason to stay arrives with it and then leaves.

The hourly rate nobody calculates

Earnings in this format are the product of three variables: how many tokens you accumulate, what the token is worth when you can finally sell it, and how many hours you spent. The first is the only one the game controls and the only one it advertises. The second is determined by the same flow of new players that sets how many people are competing for the same emissions, which makes token count and token price negatively correlated at exactly the wrong moment — the same emissions-versus-organic-revenue problem set out onDeFi yield farming. The third is the one players consistently underestimate, because the daily actions are short and the total is not.

Set that against the broader outcome data for token distributions — 88% of airdropped tokens lost value within three months and 64% of recipients sold at the token generation event — and the reason honest hourly figures are so rare becomes clear. Calculating them requires an exit price, and the exit price is usually well below the one displayed while the tokens were being accumulated. Anyone publishing a confident earnings-per-hour figure for a live game is quoting a number that has not happened yet.

Clones, drainers and the fake-game problem

Every successful distribution creates a market for imitations, and this format is unusually easy to imitate. A clone needs a name, an icon and a link, and it reaches players through the same channels the real game uses. The most common pattern places the malicious step at the end of a perfectly ordinary experience, so that hours of legitimate-feeling engagement precede a single hostile request that the player has been conditioned to approve quickly.

The defensive rules are unglamorous and effective. Reach a game only from a link published by the project itself. Treat any message announcing an early claim, a bonus window or a snapshot deadline as hostile, because urgency is the mechanism. Keep the wallet used for any claim separate from the one holding assets. And accept that a reward small enough to be a bonus is not worth the risk of an approval you did not read. The wider catalogue of loss mechanisms is on crypto earn risks.

The tax position

Tokens received as rewards are generally income at market value on the day of receipt in most jurisdictions, and that value becomes the cost basis for a later sale. This produces the category's most unpleasant surprise: a distribution valued highly on the day it lands and sold months later for a fraction creates an income tax liability based on the first number and a capital loss based on the difference, with the two frequently taxed under different rules and offsettable only in limited ways. Where custodial broker reporting exists it does not cover this activity, so the records are yours to keep. The mechanics are on tax on crypto earnings, and none of it is advice.

What genuinely survives

One thing: games people would play anyway, with rewards as a bonus rather than the reason. When the token is incidental, a falling price costs the player nothing they were counting on, the developer is funded by people who want the product rather than by the next cohort of earners, and the economics stop depending on growth. That is an ordinary games business with an unusual payment layer, and it is the only version of the idea the data does not contradict.

Everything else in the category is a distribution mechanism wearing a game's clothing. That is not worthless — a distribution mechanism that reaches tens of millions of people is a genuine achievement, and several projects executed it well. But it should be evaluated as a one-off token distribution with the outcome characteristics set out on earn free crypto, not as an income. Time spent this way is time, and time has a price even when nobody invoices for it.

The verdict

Play-to-earn as an income strategy is not supported by any data we could verify. The participation figures fell by half in a quarter, the behavioural data shows the reward itself triggering departure, retention analysis treats heavy decline as normal, and the token distributions that pay everyone out have historically lost value within months. Nothing in that chain is a scandal, and none of it required bad intent from anyone — it is simply what happens when the payment to participants is funded by participation rather than by revenue.

Play-to-earn as entertainment with an upside is a different proposition and a defensible one, provided the budget you allocate to it is a budget for entertainment. If you want a return on capital, the methods that produce one are surveyed on how to earn crypto, the venue question is handled on where to earn crypto, and the realistic scale of the whole opportunity is set out on passive income with crypto. If you want a game, pick one you would play for nothing, and treat anything it pays you as a surprise rather than a plan. For the mobile products where these two worlds increasingly overlap, see crypto earn apps.