The story of Axie’s 2021 boom showed what happens when a funny idea meets hyper financialised tokenomics. The game had a huge upside very early, followed by inflation, an SLP collapse and finally a player exodus. This is reminiscent of a whole scholarship industry that is built around extracting yield and not playing the game. By mid 2025, it had become apparently clear that the old play to earn formula of game was exhausted and dead. In Q2 2025, blockchain gaming activity fell 17% quarter on quarter to 4.8M wallets and funding collapsed 93% year on year to $73 million which is the lowest in 2 years. Over 300 gaming dapps went inactive in Q2 2025 which is about 8% of all games on DappRadar. Condesk also noted that funding is pivoting to infrastructure such as engines, asset layers and tooling instead of new Axie clones.
However, besides all this, gaming still remains the top use case on-chain. In May 2025 there were 4.9m daily active gaming wallets with gaming the largest dapp category but it was down to 19.4% market share from around 28% at the beginning of the year. In Q3 2025 gaming accounts for about 25% of all active wallets with 4.66M daily wallets and web 3 gaming funding was at $293M year to date. This makes up about 25% of the 2024 levels signalling a 75% fall in funding. This fall acts as a reminder that the Axie-style hype is gone but the blockchain gaming space is quietly rebuilding. Let’sdig in!
The new meta in 2025 is AI agents catching up to gaming
As gaming cools, the AI agent industry has become 2025’s top funded Web3 trend. By June 2025, AI-agent projects raised $1.3B against $380 million which was raised in web3 gaming. Over 17000 AI agents were launched on Virtuals Protocol since late 2024 which averaged 85+ agents per day. In addition, AI linked dapps account for 19% of on-chain activity which is nearly on level with gaming at 20%.
Chains like Matchain, opBNB and Nebula are among the most active AI agent ecosystems. DappRada’s Q2 and Q3 reports show that opBNB is the leading gaming chain by active wallets while other high throughput chains like Aptos, Sei and SKALE gain ground are ideal environments for always on AI agents. In traditional gaming, AI agents are becoming normal and it seems like the experimental phase is over. A Google Cloud Survey in August 2025 found 87% of video game developers already use AI agents to automate tasks. It was shown that 44% use them to process code, text, audio and video while 94% of them expect AI to cut long term costs of development. AI agents are not a niche web 3 toy but they are becoming the default background process of all game development. And in crypto they can be players in the economy, not just any other tools to use and forget.
Onchain agents and dynamic economies
On-chain agents are small AI bots with their own wallets and rules and they are able to click buttons for you on-chain. With the button clicking they can gather resources, trade items, join battles and vote in DAOs, without you micromanaging every transaction. Dynamic economies in gaming on the other hand use automatically adjusting prices, emissions and quests based on real world demand and supply instead of fixed rewards and drop rates.
Both on chain agents and dynamic economies matter to the Axie ecosystem due to a few reasons. Axie’s SLP economy died because rewards didn't react fast enough to changes in the economy. Axie onboarded more players, this resulted in more SLP printing which resulted in the price collapse. This cannot happen in an agentic economy because AI agents and smart contracts can adjust token rewards when too many people farm one activity. They can also automatically sink demand for crafting, upgrades and taxes when supply is too high. And finally, AI agents and smart contracts can simulate and stress test economic changes in advance.
In 2025, there are a few real examples that we can name drop. World of Dypians (opBNB) rolled out AI powered NPC guides and social hubs like Museums and Academies alongside major gameplay updates. This development showed how AI characters can be on-chain participants in web 3 MMO. AWE Network on Base is running “Autonomous Worlds” where thousands of AI agents and humans share one persistent on-chain world. And as of August 31, 2025, there were 9198 daily users, 5515 total agents and 1.7M transactions across multiple launched worlds.
In short, I can say that agent driven, self balancing economies are the logical upgrade from the play to earn that came with earlier games like Axie.
How infrastructure and chains are quietly setting up Axie 2.0
Funds in the blockchain gaming are still moving but under the hood. In Q2 2025, only $73M went into Web3 gaming, but almost 75% of it went to infrastructure, not game studios. In Q3 2025, Web3 gaming raised $129M, the strongest quarter of a weak year, however, nearly half of that money still went to infrastructure such as engines, app-chains and tooling.
The new Axie chains are the appchains and L2s. DappRadar highlights GUNZ for Off The Grid, SANDchain for The Sandbox and gaming focused L2s as examples of chains built around a single flagship game or vertical. At the same time, Ronin has evolved from Axie only to a multi game hub. In Q3 2025, Ronin’s audience grew 55% to 419K daily wallets, with titles like Axie Infinity, Lumiterra and Pixels all active. Lumiterra is the new hit and it attracted 18.6M unique wallets in Q3 2025, it has one of the fastest growth caves in Web3 gaming.
The Business Research Company projects Web3 gaming growing from $39.65B in 2025 to around $88.57B by 2029. This would give the industry a 22% annual growth which is driven by tokenized economies, L2 scaling and better dev tools. Infrastructure heavy funding, AI tooling and gaming first chains like Ronin, Base, Aptos, Sei and opBNB are the soil where next Axie scale hit will sprout
So, what does Axie 2.0 probably look like?
The first group of games that would bring a blockchain gaming revolution like Axie will have AI native companions as NFTs. In such a game, players don’t just own static characters but they own agents that learn and improve. They can become battle partners that adapt to your play style, farmer bots that run your in game production when offline and social NPCs that remember your choices and relationships. These agents live on-chain, have their own wallet and can be traded and rented out.
The next group of games is those which allow you to play to train your own agent, not just play to earn. Rewards will flow both to the human and the agent. As you play, you will be training an agent and the agent works and optimizes in the background doing things like farming, trading and crafting. The economic loop rewards skills like good strategies and good training, not pure time spent or capital deployed.
Games with dynamic and AI stabilized economies will also be the next big thing. In such games token emissions and item drops react quickly to player and agent behaviour. If bots over-farm a dungeon, drop rates and yields get nerfed automatically. IF certain regions or items are under used, the system boosts quests of rewards there. This economy will be like that of Axie, but with an upgrade of self healing to avoid the death spiral.
Final thoughts and conclusion
The real alpha is not grinding non stop on the latest farm token but its learning how to command and co-own AI agents inside these worlds. Builders need to know that winners will not be the noisiest play to earn forks, they will be teams that treat AI agents as first class citizens in their world, who put the most funding into infrastructure and economy design, not marketing, and teams that are willing to let their economy evolve on chain instead of hardcoding everything offchain. Investors on the other hand should ask questions on chains that are becoming agent hubs, AI native worlds that are shipping to mainnet and who has both the sticky game and a robust infrastructure.
References
DappRadar – “State of Blockchain Gaming in Q2 2025” (July 10, 2025) https://dappradar.com/blog/state-of-blockchain-gaming-in-q2-2025
CoinDesk – “Web3 Gaming Faces Ongoing Turmoil, Market Metrics Reveal Persistent Decline” (July 11, 2025) https://www.coindesk.com/web3/2025/07/11/web3-gaming-faces-ongoing-turmoil-market-metrics-reveal-persistent-decline
DappRadar – “State of Blockchain Gaming Q3 2025” (Q4 2025) https://dappradar.com/blog/state-of-blockchain-gaming-q3-2025
Juice Gaming News – “AI Agents Outpace Crypto Gaming as 2025’s Top‑Funded Web3 Trend” (July 7, 2025) https://juicenews.io/article/ai-agents-outpace-crypto-gaming
The Business Research Company – “Web3 Gaming Industry Forecast to Hit $88.57 Billion by 2029 With a Strong 22.2% CAGR” (November 27, 2025) https://blog.tbrc.info/2025/11/web3-gaming-market-3/
