First, the filter, because "actually shipping" needs teeth. A project qualifies if it passes all four: (1) playable today by anyone without an allowlist; (2) a patch, event or balance change shipped in the last 30 days; (3) observable player activity — concurrent players or market volume — from independent trackers, not the team's own dashboard; (4) economy design where sinks exceed faucets at current usage. Roughly 300 projects announced, ~40 launched, 5 still standing. That's the survival rate. Keep it in mind whenever someone pitches you "the first AAA web3 MMO".

We're deliberately not naming the five in the headline list-style you'd expect, for a reason you'll see. Instead, the five archetypes they occupy — because the archetypes are the transferable knowledge.

Archetype 1: The strategy layer game

Turn-based, server-authoritative, on-chain only where the rules matter (ownership, trades, tournament results). The lesson: put the chain where cheating hurts. Nothing else is on-chain, and players don't miss it. Survived because its economy is a tournament meta, not a wage.

Archetype 2: The collect-and-trade world

A persistent world whose core loop is collecting, breeding/crafting and trading — the one genre where the marketplace is the game. Wallet-optional storefront, gasless trades via L2, items settle on-chain when it matters. The lesson: if trading is the game, web3 is a feature; if trading is the exit, web3 is a scam.

Archetype 3: The esports-anchored title

A competitive game with on-chain tournaments, prize escrow and transparent brackets. This one understands something most studios missed: the chain's killer app in gaming isn't items, it's trustless prize pools. No more "the organizer vanished with the pot" stories.

Archetype 4: The UGC platform

Player-created content with creator royalties enforced at the asset layer. The chain's role is plain: pay creators programmatically, forever, without trusting the platform's accounting department. Closest thing to a real "ownership economy" that actually works, because the earners are creators, not clickers.

Archetype 5: The compute-hungry oddball

Our favorite, for obvious reasons: a persistent world with AI-driven NPCs and simulation-heavy gameplay, whose team made headlines by cutting infrastructure costs ~40% by running non-critical simulation workloads on decentralized GPU markets (see our piece on who's buying crypto compute). The lesson: in 2026, the web3 part of a web3 game might quietly be the infrastructure, not the tokens.

The pattern behind the pattern

  • Every survivor treats blockchain as infrastructure for trust — ownership, escrow, royalties, verification — never as a marketing channel or a payroll.
  • Every survivor has an economy designed by someone who has read one book about monetary history and one about game design, in that order.
  • Three of five run meaningful workloads on commodity/decentralized compute. Games are brutally exposed to infra costs — servers scale with players, not with funding rounds.
  • None of them says "play-to-earn" anywhere in their store listing. Not one word. The market has voted.

The names of the five matter less than people think — two of them were near-death last year and one will probably fumble it eventually. The archetypes are the durable part. If you're evaluating a web3 game (or building one), find which archetype it honestly belongs to. If it belongs to none — if the chain is there "for community" — you already know how that movie ends.