Artificial intelligence agents will transform Web3 in 2025, with cryptocurrency staking and onchain trading emerging as early use cases, industry executives told Cointelegraph.
Agentic AIs — machines pursuing complex goals autonomously — are already reshaping the digital economy, building Web3 applications, launching tokens and interacting with humans autonomously.
In 2025, “AI agents are expected to take on a more prominent role within decentralized communities,” J.D. Seraphine, a Web3 AI developer at Raiinmaker, told Cointelegraph.
They will also face headwinds, including technical challenges, regulatory hurdles and centralization, Michael Casey, co-founder of the Decentralized AI Society, told Cointelegraph.
“Without decentralization, centralized, misaligned systems will drive us off a cliff, especially with AI,” Casey said.
Related: AI memecoins will become utility tokens
One million AI agents
As of December, Web3 hosts approximately 10,000 AI agents, collectively earning millions of dollars each week from onchain activities, according to a report by VanEck.
The asset manager expects upward of 1 million AI agents to populate blockchain networks by the end of 2025.
“The potential universe of AI agents interacting with crypto is infinite,” Matt Hougan, asset manager Bitwise’s head of research, told Cointelegraph.
In 2024, tokens tied to agentic AI bootstrapped upward of $10 billion in market capitalization, mostly in Q4, according to CoinGecko.
Agentic AI projects include ai16z, which aims to use AI to direct onchain investments, and Virtuals, a platform for launching AI agents on Coinbase’s Base network.
Early use cases
Cryptocurrency staking on behalf of human tokenholders is a viable early use case for agentic AIs, Hougan told Cointelegraph.
“We’ll see a lot of experiments, but only a few will stick. AI agents participating in staking, for example, seems like a logical first step,” Hougan said.
Staking involves securing a blockchain network by locking up tokens with validators in exchange for a cut of transaction fees.
Meanwhile, ai16z’s agent, Eliza, autonomously manages an onchain liquidity pool, purportedly clocking annualized returns in excess of 60%, according to data from daos.fun.
Still, onchain AI models lag centralized counterparts, such as OpenAI’s ChatGPT, on technical dimensions such as speed and computational power, Casey said.
Creating viable decentralized AI agents depends on finding “decentralized solutions to ensure high-quality training data while safeguarding user privacy,” Seraphine said.
Meanwhile, “AI will face increasing pressure to be regulated, and big players like OpenAI are lobbying for rules that align with their own models,” potentially disadvantaging decentralized AIs, Casey added.
For investors, “it’s okay not to know exactly what will happen as long as you recognize the significance and position yourself for exposure,” Hougan explained.
Related: Why tech giants like Amazon may hesitate to adopt Bitcoin
