Io.net has introduced a new token burning strategy that is directly linked to its network earnings. The initiative could eliminate as many as 12 million IO tokens from circulation within the next year, according to the decentralized GPU provider, which also reports surging enterprise demand and record-breaking AI inference activity.
The first token burn was carried out on June 11, coinciding with the network’s third anniversary, and unlike typical burn programs, it is funded by actual revenue from customer usage rather than minting new tokens. Under the Incentive Dynamic Engine (IDE), at least half of the post-payout network revenue received in IO tokens will be permanently destroyed. Based on current earnings and the commercial pipeline, io.net anticipates burning up to 12 million tokens over the first year.
This news comes as io.net experiences its strongest commercial period ever. The company has secured an $8 million enterprise agreement—its largest contract to date—which contributes roughly $650,000 in monthly on-chain network earnings. Additional enterprise deals are in advanced negotiation stages. Beyond enterprise adoption, io.net is now the largest decentralized physical infrastructure network (DePIN) inference provider on OpenRouter, a platform that enables developers to access various AI models. The network processes over 4 billion inference tokens daily, competing alongside centralized cloud computing providers.
Demand for AI computing resources continues to climb, with major tech companies committing over $500 billion to AI infrastructure projects in 2025 and 2026. Io.net argues that high-performance GPU access remains constrained by hyperscaler capacity limits and pricing structures, creating opportunities for decentralized alternatives.
The IDE also aims to stabilize supplier earnings, addressing retention challenges common in token-based infrastructure networks. Supplier payouts are now pegged to a stable US dollar value rather than volatile token prices. Reserve mechanisms absorb market volatility, enabling providers to maintain predictable earnings even during token price downturns. Independent tokenomics research firm CryptoEcon Lab tested the model under stress scenarios, including a 55% drop in demand and a 50% decline in token price, and found that supplier returns remained stable in both cases.
Gaurav Sharma, CEO of io.net, emphasized that most token economies rely on price appreciation hopes, but io.net’s model is based on real network usage payments. Looking ahead, io.net is developing capabilities for AI agents to autonomously source and manage computing resources through its Agent Cloud platform, aiming to build a self-sustaining on-chain compute economy supported by decentralized infrastructure providers worldwide.