Crypto AI Agents Face Reality Check as Virtuals Protocol Revenue Collapses
Crypto AI agents were once celebrated as a groundbreaking platform at the intersection of artificial intelligence and blockchain. However, Virtuals Protocol is now facing a significant decline in user engagement and revenue. Daily earnings, which previously exceeded half a million dollars, have plummeted to just $500, raising serious concerns about the future of the platform.
The Rise and Fall of Crypto AI Agents
The dramatic collapse in activity on Virtuals Protocol is part of a broader cooling of investor enthusiasm toward crypto AI agents.
According to Dune Analytics data highlighted by Blockworks researcher Sharples on April 8, the number of new AI agents launched on Virtuals has slowed to a crawl—there hasn’t been a single new agent in over a week.
Also read: How To Create an AI Agent
Virtuals Protocol daily trading revenue (Source: Dune Analytics)
This is a far cry from the platform’s heyday in late November, when it saw over 1,000 new agents created daily.
The decline appears to have set in shortly after the Virtuals token (VIRTUAL) peaked at $4.61 on Jan. 2, coinciding with a spike in daily revenue that exceeded $500,000.
While the token’s high was a milestone, it also seems to have marked the top of the hype cycle. Revenue and token value have since plummeted, with VIRTUAL trading as low as $0.42 this week.
Also read: Crypto AI Agents: A Booming but Speculative Trend, Says Sygnum Bank
Virtuals’ attempt to reinvigorate interest by expanding to the Solana ecosystem in late January has done little to stem the decline.
On April 7, Sharples reported that the platform’s daily revenue had dropped below $500, despite the project’s lifetime revenue still standing at an impressive $39.1 million.
Several factors are contributing to the platform’s sharp reversal of fortune.
DeGen Capital core contributor Mardo pointed out that current macroeconomic headwinds, including the broader crypto market downturn and geopolitical uncertainties like US tariff escalations, are placing pressure on speculative digital assets.
But Virtuals Protocol’s internal decisions may also be partly to blame. Mardo highlighted policies such as withholding token taxes from builders—funds that other platforms might redistribute—as potentially damaging to the developer ecosystem the platform relies on.
The broader market for crypto AI agents is not faring much better. Dune Analytics puts the total market cap of AI agent-related tokens at $153.81 million, though nearly half of that is tied up in AIXBT, a sentiment-analysis tool built on social media trends.
AIXBT has seen its value plunge by 92%, falling from a January peak of $0.90 to just $0.07 as of this week.
Adding fuel to the fire is a growing chorus of criticism from within the crypto and AI communities. Many argue that crypto AI agents, as they exist today, are little more than “ChatGPT wrappers”—a term used dismissively by AI commentator BitDuke, who suggested their novelty has worn off.
Fellow AI analyst “DHH” echoed the sentiment, stating, “You’re delusional if you think any AI agent is full-on replacing a great programmer today.” Their skepticism is reflected in the dwindling interest from developers and users alike.
Infinex founder Kain Warwick was even more blunt, labeling the current generation of AI agents as “slop”, while holding out hope that future iterations may offer more utility and appeal.
Is There a Second Act for AI Agents?
The rapid fall of Virtuals Protocol may well be a cautionary tale about the fragility of hype-driven markets.
While the platform once symbolized the promise of decentralized AI integration, its declining revenue, user engagement, and token value paint a starkly different picture in 2025.
Also read: DeepSeek AI: Exploring the Capabilities
Still, not all is lost. As Warwick and others have hinted, this may merely be the end of the first wave of crypto AI agents, rather than the end of the category altogether.
To regain traction, platforms like Virtuals will need to focus less on tokenomics and more on real-world functionality, usability, and sustainable incentive structures.
Until then, it appears the AI agent bubble has burst—at least for now.
