The numbers hit my terminal at 3:47 AM Dublin time. Veridian’s preliminary Q2 revenue: $11.5 billion. That’s not a typo. The same protocol that scraped by with $787 million in Q2 last year is now printing cash like a central bank. Annualized revenue hit $47 billion in May, and the team is claiming an adjusted operating profit for the quarter. The chat rooms are exploding. “Bullish,” “L2 king,” “AI cash cow.” But I’ve been staring at the on-chain ledger for the past six hours, and something is off. Red candles don’t lie, and neither does the data. Let me walk you through why this revenue spike might be the most dangerous signal in crypto right now.
—
Context: The Rise of Veridian
Veridian is a Layer-2 blockchain optimized for AI model inference. It launched in 2023, positioning itself as the settlement layer for machine learning workloads. The pitch was simple: execute AI computations on-chain, pay fees in VER tokens, and let developers build decentralized AI agents. For the first year, it was a ghost town. Then, in early 2025, a single large AI firm—let’s call it “NexMind”—started processing millions of inference requests daily. Veridian’s transaction count exploded. Its fee revenue went from a trickle to a flood. By Q1 2025, it reported $4.73 billion. Now, Q2 is $11.5 billion. That’s a 14x year-over-year increase. For context, the next largest L2, Optimism, does about $500 million in annualized fee revenue. Veridian is in a league of its own.
But here’s the thing—I’ve been in this game since 2017. I’ve seen ICOs with zero code commits, DeFi farms blowing up overnight, and NFT floors cratering 40% in a day. The pattern is always the same: when a single entity controls the majority of a protocol’s economic activity, the rug is not a question of if, but when. Veridian’s success is not decentralized. It’s a single-client dependency wearing a decentralized mask.
—
Core: The Data Doesn’t Lie
Let me show you what I found. I pulled the on-chain data from Veridian’s block explorer. I wrote a quick Python script to analyze the top 100 transaction fee payers for the last 90 days. The result? One address—0x9f4e…3b2a—accounts for 82% of all fees paid. That address belongs to NexMind. I verified this by cross-referencing the smart contract interactions: 99% of the calls are to a single inference engine contract deployed by NexMind. The terminal output is brutal: 82% concentration. That’s not a network; that’s a private server.
Now, let’s talk about the revenue. Veridian’s native token, VER, is required to pay fees. The protocol collects these fees, burns a portion, and distributes the rest to stakers and validators. The $11.5 billion figure is the gross fee revenue. But here’s the catch: NexMind is not paying these fees in VER they bought on the market. They are receiving massive token grants from the Veridian Foundation. In Q2 alone, the foundation unlocked 400 million VER tokens, worth roughly $8 billion at average prices. NexMind sold a portion of these tokens to cover its operational costs, including the fees it pays back to Veridian. It’s a circular shell game. The revenue is artificially inflated by the very token issuance that is supposed to be the protocol’s lifeblood.
I cross-checked this with the tokenomics data. The adjusted operating profit that Veridian boasts excludes the cost of token dilution. If you account for the foundation’s token grants, the net profit is negative. The protocol is bleeding capital to keep the revenue numbers high. This is classic wash trading, but at a protocol level. Wash trading: The digital casino’s oldest trick.
Let me give you a concrete example. In May, Veridian announced that its annualized revenue surpassed $47 billion. That same month, the foundation sold 150 million VER tokens to NexMind at a discount. NexMind then used about 10% of those tokens to pay fees, creating the illusion of demand. The rest went to the market, depressing the token price. VER is down 60% from its all-time high. The market is pricing in the dilution, but the narrative is still bullish because of the revenue headline. This is a classic valuation trap.
—
Contrarian: The Unreported Angle
Everyone is comparing Veridian to OpenAI. OpenAI’s annualized revenue is $40 billion, Veridian’s is $47 billion. But the comparison is apples to oranges. OpenAI’s revenue comes from real customers paying for subscriptions and API access. Veridian’s “revenue” is largely internal recycling. The real metric to watch is the net revenue after token dilution. Based on my calculations, Veridian’s net revenue is closer to $300 million—a far cry from the headline.
But there’s an even darker angle. The positive adjusted operating profit is only possible because of the token price. VER’s price is still above $20, allowing the foundation to sell tokens at a premium. If the price drops below $5—which is likely given the inflation—the entire economic model collapses. The protocol will have to either slash rewards or increase fees, both of which will drive away the few remaining organic users. Exit liquidity is someone else’s problem, but not for long.
I also found that the majority of Veridian’s validators are controlled by NexMind and a few venture capital firms. The network is not truly decentralized. The sequencer is a single entity. This is the same problem I’ve been calling out for Layered protocols for years. Layer2 sequencers are basically single centralized nodes. Veridian is just a more extreme example. The “decentralized sequencing” PowerPoint is still just a PowerPoint.
—
Takeaway: What to Watch Next
The next three months will be critical. Veridian’s token unlock schedule is set to release another 1.2 billion tokens by year-end. If NexMind reduces its usage, or if the foundation fails to find new customers, the revenue will collapse. The token price will follow. I’m watching the on-chain activity of address 0x9f4e…3b2a. If it starts moving tokens to exchanges, get out.
Veridian’s Q2 numbers are a masterpiece of financial engineering. But in crypto, engineering always has a shelf life. The question is not whether the revenue is real, but whether the market will realize it before the tokens run out. Red candles don’t lie. And when they start, they don’t stop.
—