Io.net unveils dynamic token model for DePIN AI GPUs

Io.net has unveiled a redesigned token model for its decentralised AI compute network, in a move that aims to link token rewards more closely with real usage and demand.

The company announced the Incentive Dynamic Engine, or IDE, as a new framework for how it issues and allocates its native tokens across hardware providers, users and investors.

Io.net runs a distributed network of graphics processing units, or GPUs, that supports AI training and inference workloads. The company said its network has handled more than $20 million in compute leases since June, across tens of thousands of GPUs.

The new model is part of a wider effort in the decentralised physical infrastructure network, or DePIN, sector. Many DePIN projects use token incentives to attract hardware providers but face questions over inflation, market volatility and limited real-world usage.

Dynamic token system

IDE uses a dynamic control system that adjusts token release and payout levels based on real-time network conditions. The company said the design links token emissions with factors such as demand for compute and utilisation of available GPUs.

Io.net said the approach differs from earlier models that focused on bootstrapping supply through high token rewards. Those models often produced persistent inflation and did not always align with end-user demand.

The company said IDE aims to stabilise income for GPU providers in more than 130 countries. It also said it aims to support more predictable pricing for users of the network.

Gaurav Sharma, CEO, io.net, said, "We're at a crucial juncture for AI: either continue with centralised hyperscalers underpinned by obscure, circular finance, or build decentralised, open markets for compute."

"DePINs in their current form are not fit for purpose. IDE is a pioneering new model, and is the foundation for the next phase of growth of io.net. GPU providers, users and investors will benefit from the first reliable and open compute network, accurately aligning incentives. Following its implementation next year, IDE will enable startups, researchers and enterprises to develop and deploy AI systems on io.net for the long term."

Enterprise trust gap

Decentralised compute networks have so far seen limited adoption among large enterprises. Corporate users often cite concerns about reliability, transparency and the relationship between resource supply and application demand.

Io.net said many blockchain-based networks obscure the link between GPU availability and usage. It said this lack of transparency has eroded trust in the stability of decentralised infrastructure.

The company said IDE works alongside a commitment to greater visibility into network data. It said the combination is designed to create a more open market for compute that is easier for enterprises and investors to understand.

Io.net said an open market that directly matches any hardware provider with any user can support long-term capital deployment by investors and hardware manufacturers. It said this structure can reduce concerns about circular financing arrangements in AI infrastructure.

Benefits for stakeholders

The IDE framework includes mechanisms that the company describes as deflationary in certain conditions. These mechanisms link token flows with real-world usage rather than speculative trading.

Io.net said the model reduces income volatility for GPU providers. It also said the design rewards investors who support actual network utilisation.

The company said users of the network could see more resilient access to compute as a result of incentive realignment. It said this is important for AI developers that depend on consistent access to GPU capacity for training and running models.

Io.net's network supports AI projects at startups, scale-ups and larger organisations. The company said customers span media, entertainment, blockchain infrastructure and research institutions.

Io.net said it operates one of the largest pools of distributed GPUs in the market. Its platform supports model training, deployment and large language model workloads.

Community review

The IDE proposal has been released in a litepaper for review by the io.net community. The company has set a feedback period that runs until late February.

It said it plans to incorporate community input into a final tokenomics design. The updated document is scheduled for publication at the end of March.

The company expects to roll out the redesigned token model on its network in the second quarter of 2026. It said the launch will place its global GPU base into a market structure that aligns all stakeholders around maintaining an open and resilient decentralised AI compute network.

Io.net said it views IDE as the basis for the next phase of its growth, as AI developers seek alternatives to centralised cloud providers.