📈 Get daily crypto insights that make you smarter about your money

Sky Mavis Closes Axie Vibeathon With 90 Community Games Built by AI and Only 11 That Need a Wallet

Sky Mavis has closed the books on the Axie Vibeathon, its AI-powered game-building hackathon, and the final numbers tell a striking story: 402 registered builders produced 90 playable games, nearly all of them built with AI coding tools rather than traditional development pipelines — and only 11 of them actually needed a wallet.

From idea to playable in two weeks

The Vibeathon was structured as a bet that clear ideas plus AI tooling were enough to ship something playable, aimed at non-programmers as much as experienced developers. Registration opened Sept. 1 and closed Sept. 7, with the build round running Sept. 8 to Sept. 21. Entrants competed solo, each submitting a single game hosted on a server under their own control — the platform links out to entries rather than running them itself.

Every submission had to be playable in a browser and ship with a title, a pitch, a thumbnail, control instructions, a code repository link, a fallback demo video, documentation of the AI tools used, and a statement explaining how the game fits Axie Core. Sky Mavis left the engine choice open, and the use of Axie art, NFTs, onchain data or the API was optional at the prototype stage.

The results went live in the Axie Games Hub on Sept. 22 at 11 a.m. Philippine time, with most entries running directly in the browser. All 90 surviving titles can be browsed and played on the official Vibeathon gallery site.

Judging: likes don’t count

Community enthusiasm is visible but carries no official weight. Logged-in users can like entries and leave comments for builders, but judging follows five weighted criteria: Axie Core concept at 35 percent, gameplay at 25 percent, product vision at 20 percent, feasibility at 10 percent, and prototype and documentation quality at 10 percent. Every entry must pass an eligibility and safety screen before receiving three independent judge scores.

Early community feedback has already sorted the field. According to the company, the five leading games have drawn 40 percent of all positive responses, with the 3D auto-battler Lunacia Command collecting the most likes. Nearly half of all submissions leaned into strategic combat — a testament to how deeply the original Axie Infinity formula shaped what builders imagine when they think “Axie game.” Notable projects include Axie Infinity: Tactics, a tactical RPG inspired by classic games, and Lunacia Synergy, a 3D auto-battler that builds Axie DNA directly into its mechanics.

The wallet number is the headline

Perhaps the most revealing statistic: builders were allowed to add blockchain features, but only 11 of the 90 games required a Ronin wallet or used real Axie NFTs. For a chain-native studio like Sky Mavis — the company behind Ronin and one of the pioneers of play-to-earn — that ratio is not a failure. It is evidence of where the ecosystem’s own community instinctively draws the line: games first, crypto hooks only where they earn their place.

Sky Mavis co-founder Jeffrey “Jiho” Zirlin, posting as Jihoz, made the point explicitly when the first batch went live, tagging Anthropic’s Claude and OpenAI in a post celebrating “page after page of new axie experiences, built by YOU.” His read: this is something unique that he has not seen anywhere else in gaming.

What happens to the winners

Up to eight finalists were scheduled to be announced Sept. 29, each receiving 250 USDC toward AI tooling. Round 2 runs from Oct. 4 to Oct. 31, where finalists must build a full game loop — with any Axie asset or blockchain hooks cleared by Sky Mavis. Winners are expected Nov. 5, sharing a community prize pool of 20,000 bAXS, the locked form of AXS roughly worth 18,000 USD when the event was announced. A separate 500 bAXS Promising Prototype bonus is reserved for a strong Round 1 entry that misses the final cut.

The company has been careful about expectations: no winner, finalist or recognized prototype automatically earns a production pathway. That said, Sky Mavis has said winning concepts could become ecosystem integrations or standalone games, and the studio framed this run as a first iteration — settling the tools, format and route from an idea to a published Axie game.

Why it matters

The Vibeathon arrives at a moment when Axie’s ecosystem is deliberately broadening beyond its origins. Origins Season 19 is currently running with an 80,000 bAXS prize pool, and the franchise’s surrounding catalog has expanded to more than 140 titles. A hackathon that converts 400 community members into 90 shipped prototypes — most browser-native, wallet-optional — is a cheap way to scout both talent and demand, and the judging weights make the priorities clear: capture what makes Axie feel like Axie first, then decide if a chain is needed.

For Web3 gaming at large, the 11-of-90 wallet figure may be the most quoted stat of the bunch. It suggests the next wave of onchain games may look exactly like the current wave of browser games — until, and unless, ownership mechanics prove they add something a leaderboard can’t.

13 thoughts on “Sky Mavis Closes Axie Vibeathon With 90 Community Games Built by AI and Only 11 That Need a Wallet”

    1. And only 11 of 90 needing a wallet is the interesting number. Even Sky Mavis quietly admitting crypto login is friction.

      1. Or wallet-free entries pull in actual game devs who would never touch a crypto hackathon. Trojan horse strategy if it works

    2. @spawnpoint_ right, and every entry hosting on its own server means Sky Mavis shipped zero moderation overhead too. they basically crowdsourced the QA and the hosting bill

  1. axie core weighted at 35 percent means the lore nerds beat the engineers. explains why lunacia command is sweeping the likes

  2. 402 registered builders and only 90 shipped something playable. even with ai doing the coding most people just dont finish. feels accurate honestly

    1. 22 percent completion is actually solid for a two week sprint, most game jams sit under 15. ai mostly raised the floor on finishing at all

    2. The wilder stat is only 11 of 90 games needed a wallet. Axie running a hackathon where 87% of entries skip onchain entirely says a lot about where the ecosystem actually is.

  3. Likes carrying zero weight in judging is the right call. Otherwise it becomes a popularity contest and the bots win.

  4. vibe coding a game in two weeks with open engine choice… most will be junk but a few of these 90 will be genuinely good. also nice that every entry needs a fallback demo video

  5. requiring documentation of the ai tools used is actually smart. makes entries auditable instead of trust me bro i wrote it

Leave a Comment

Your email address will not be published. Required fields are marked *

BTC$83,581.00-2.2%ETH$2,689.18-1.6%SOL$116.96-4.4%BNB$767.73-1.2%XRP$1.48-3.5%ADA$0.2459-3.3%DOGE$0.0941-3.5%DOT$1.19-5.0%AVAX$10.92-3.4%LINK$14.25-3.3%UNI$8.94-2.4%ATOM$1.72-2.2%LTC$66.75-2.0%ARB$0.1985-4.9%NEAR$4.85-9.4%FIL$1.00-7.6%SUI$1.15-4.5%BTC$83,581.00-2.2%ETH$2,689.18-1.6%SOL$116.96-4.4%BNB$767.73-1.2%XRP$1.48-3.5%ADA$0.2459-3.3%DOGE$0.0941-3.5%DOT$1.19-5.0%AVAX$10.92-3.4%LINK$14.25-3.3%UNI$8.94-2.4%ATOM$1.72-2.2%LTC$66.75-2.0%ARB$0.1985-4.9%NEAR$4.85-9.4%FIL$1.00-7.6%SUI$1.15-4.5%
Scroll to Top