It's worth remembering how insane this was. An entire industry raised billions on the thesis that the fun part of a game could be replaced with a payroll. Axie Infinity at its peak had more daily players than Belgium has people, and most of them were not playing. They were shift-working — clicking through a game they didn't like, in an economy managed by guilds that functioned as payday lenders, earning tokens that were always one halving away from ruining them.

You know how it ended. But the ending is less interesting than the diagnosis, because the diagnosis applies to half the crypto industry.

The three-body failure

1. Paying people selects the wrong people

The moment a game pays, it stops competing with other games and starts competing with other jobs. Your player base is now applicants. They optimize for yield, not fun, and they leave the moment the APR does. Retention curves of P2E games didn't look like games; they looked like emission schedules — because that's what they were.

2. The economy was a check-kiting scheme

Every P2E economy had the same accounting: value flowing out to earners had to come in from newcomers buying NFTs. That's not an economy, it's a sequence. When the sequence ended — and sequences always end — withdrawals and token price entered the death spiral together. Games with "sinks" (ways to burn tokens) fared marginally better, but a sink can't save you when 90% of participants have zero interest in what the sink produces.

3. Ownership was the point, and nobody owned anything

The cruelest irony: the one genuinely good idea — you keep your items when the game ends — was violated first. When Axie's economy collapsed, the "player-owned" assets became worthless. Ownership of an asset class that depends entirely on one company's server is not ownership; it's a loyalty card with extra steps.

What the survivors kept

The games still standing in 2026 kept one web3 property and discarded the rest, and it's instructive which one: portable, tradeable assets. Not earnings. Not governance tokens. Items.

  • Fun first, wallet optional. The healthy 2026 web3 games are playable start-to-finish without a crypto wallet. Blockchain is the item layer, not the front door. Conversion to on-chain assets happens when a player cares enough to want it.
  • Sinks before faucets. New economies are designed sink-first: every token entering a player's wallet was spent on something they wanted. Emissions are treated like radioactive material — handled carefully, minimized, scheduled to decay.
  • No promises of income. Marketing that mentions earnings is now, correctly, read by players as a scam warning. The genre vocabulary has shifted from "play-to-earn" to "play-and-own", which sounds like a rebrand but is actually a business-model statement.

The general lesson

Here's why this autopsy belongs on a compute forum and not just a gaming one. P2E, DePIN emissions, liquidity mining, airdrop farming — they're all the same machine with different skins: pay people to do the thing, discover that the people who show up only do the thing for pay, and watch the whole structure evaporate when the pay does.

The fix is always the same and always unglamorous: make the thing worth doing without pay, treat incentives as scaffolding with a demolition date, and let the asset — the item, the verified compute hour, the shipped feature — be the product. Gaming ran the experiment at extreme scale and paid full tuition. Everyone else should read the transcript instead of re-enrolling.