io.net operates what it describes as the world's largest decentralised GPU network. Instead of renting computing power from Amazon, Google or Microsoft, customers rent it from a distributed network of independent GPU owners, coordinated through blockchain infrastructure.
The company is now introducing a system that permanently destroys its own tokens based on how much money customers spend on the network. The first burn takes place today.
What a token burn is, in plain terms
io.net has its own cryptocurrency, called $IO. Like any currency, its value depends partly on supply and demand. If there are fewer tokens in circulation and the same or more demand to hold them, each remaining token becomes more valuable.
A token burn removes tokens from existence permanently. They are not sold, transferred or held in reserve. They are destroyed.
io.net's new system, called the Incentive Dynamic Engine, takes at least 50% of the network's revenue, denominated in $IO tokens, and burns it. Based on current earnings and pipeline projects, the company expects to destroy at least 12 million tokens over the next year.
The logic is straightforward: every dollar a customer pays to use the network makes the remaining tokens slightly scarcer and, in theory, slightly more valuable.
No connection to economic reality
Most cryptocurrency tokens have no connection to real economic activity. Their price rises and falls on speculation, social media sentiment and broader crypto market cycles.
io.net is attempting something different: tying the burn rate to actual customer revenue. CEO Gaurav Sharma framed it directly: "Ours is built around the certainty that people are paying to use the network."
The company has closed an $8m enterprise contract generating roughly $650,000 in monthly on-chain earnings and says it processes up to 4 billion AI tokens per day through routing platforms including OpenRouter.
How suppliers get paid
GPU owners who contribute computing power to the network are paid in dollar-pegged values, not in fluctuating crypto prices. Built-in reserves absorb price swings in the $IO token, so a supplier's income does not collapse if the token price drops.
CryptoEcon Lab stress-tested the model across scenarios including a 55% demand collapse and a 50% token price crash. The system is designed to keep paying suppliers even when market conditions deteriorate.
The question is scale
The model is more sophisticated than most crypto tokenomics. Tying burns to revenue rather than arbitrary schedules, and pegging supplier payments to dollars rather than volatile tokens, addresses two of the most common criticisms of crypto infrastructure projects.
The question is scale. An $8m contract and $650,000 in monthly revenue are meaningful for a startup. They are a rounding error relative to the hyperscaler GPU market io.net is positioning itself against. The burn mechanism works if revenue grows. If it plateaus, 12 million fewer tokens will not change the fundamental economics.
The first burn is today. The real test is whether there is enough demand to keep burning next year.
