io.net’s New Tokenomics Links Burns to GPU Demand

DePIN networks, which aim to coordinate decentralized physical infrastructure like compute or storage, have increasingly faced the same tokenomics challenge. When incentives are delivered through fixed, time-based emissions, supply can rise even if real-world usage lags. That mismatch can pressure token prices, discourage operators from contributing resources, and ultimately weaken the network’s ability to deliver services.

According to a CoinDesk Research report commissioned by io.net, the company’s newest tokenomics approach, called the Incentive Dynamic Engine (IDE), is designed to break that cycle by tying parts of the token economy more directly to actual GPU demand. The report describes mechanics intended to keep supplier payouts in dollar terms and to direct surplus toward token buybacks and permanent destruction.

The core problem with fixed emissions in DePIN

Many DePIN token models follow a familiar pattern: tokens are minted on a predetermined schedule to reward contributors. The CoinDesk Research report argues that this structure creates a structural risk, particularly when demand drops. If emissions continue regardless of whether customers are actually using the network, the token supply can expand faster than usage. Over time, that can lead to a “negative spiral,” where operators reduce participation after token prices fall, further reducing throughput and usage.

While this issue is not unique to DePIN, it is especially salient for compute-focused networks, where the economics depend on sustained demand for workloads such as AI training and inference.

How io.net’s Incentive Dynamic Engine is meant to work

The IDE is described as a mechanism that replaces fixed token emissions with a demand-linked approach. Instead of issuing a constant number of tokens each epoch, the engine calculates the dollar payout owed to active GPU suppliers and releases only the IO tokens required at the live token price. In effect, when the token price is lower, fewer dollars translate into more tokens, and when the token price is higher, token payouts can be reduced.

The report also outlines a burn-and-retire element. After supplier payouts are covered, at least 50% of any remaining surplus is allocated to buy back IO and permanently destroy it. The stated objective is to retire a portion of the non-emitted reward pool, which the report characterizes as having 231 million IO remaining in an “un-emitted reward pool,” with a target of at least 115 million IO being destroyed.

Two reserves back the payout logic

CoinDesk Research describes the engine as drawing from two reserves. A Reward Vault, sourced from block rewards, is used first. A Fee Vault, sourced from client payouts, functions as a backup when the first reserve is insufficient. The report emphasizes that this reserve structure is central to the model’s ability to maintain supplier income under adverse conditions.

When does the system become net-deflationary?

The report introduces the concept of a “sustainability ratio,” framed as a condition above 1, where network earnings exceed supplier obligations. Under those circumstances, surplus is available for buyback and burn. The report describes that burning is already active at launch at a sustainability ratio of exactly 1, with the burn mechanism targeting at least 50% of block-reward emissions in that scenario.

It further states that the engine is expected to become net-deflationary once the margin above a sustainability ratio of 1 grows enough that burn activity on at least half of combined surplus exceeds the emission rate.

Stress tests and what they are trying to prove

A key part of the report focuses on simulations conducted by CryptoEcon Lab. It describes stress tests against a 55% demand drop and a 50% price crash. The scenario analysis is presented to show that supplier income in dollar terms can hold even when demand and token price move sharply, by drawing on reserves rather than forcing market-facing token dumps.

The report also discusses a prolonged low-activity scenario, in which obligations scale down as participation falls, since compensation is tied to active supplier participation rather than fixed commitments. The intention, as presented, is to prevent an accumulation of long-term liabilities when the network cannot sustain full usage.

Where io.net fits in the broader compute economy

io.net’s model aggregates idle GPU capacity from data centers, miners, and independent operators across multiple countries, then bundles that capacity for AI and machine-learning workloads. The company’s pitch, as reflected in the report, is that this structure can produce pricing below public cloud options.

CoinDesk Research also references network earnings trends, stating that daily network earnings have been trending upward since March 2026, averaging roughly $35,000 to $36,000 per day and moving broadly in line with the IO token price. The report frames this as context for how token and economic activity may be aligned under the IDE.

Enterprise contract and the metric to watch

In addition to the mechanics and simulations, the report cites an enterprise agreement as already active. It characterizes the contract as contributing about $650,000 per month in network earnings and projecting at least 12 million IO for burning in the first year, subject to the system’s performance metrics.

The report flags one signal it says matters for sustainability: the sustainability ratio holding at or above 1. In practical terms, this is the condition under which earnings remain sufficient to cover supplier payouts and still leave surplus available for buyback and burn.

Implications for DePIN token design

If the IDE performs as described, it highlights a broader shift in how DePIN projects may approach token economics. Rather than treating emissions as a fixed scheduling problem, the model frames token issuance as an outcome dependent on demand and network revenue. For investors and operators, the appeal is that incentives could be less vulnerable to price-driven reflexes, and that contributor compensation may be stabilized in dollar terms.

Still, the real-world test will depend on sustained demand, the stability of payout reserves, and how network earnings and token valuation interact over time. Tokenomics models are only as strong as their assumptions, and reserve-backed systems can still face constraints if revenue collapses for extended periods.

This is a research report, commissioned by io.net and produced by CoinDesk Research, and it does not constitute investment advice or a recommendation. As always with tokenomics changes, the market will likely focus on whether the sustainability conditions described in the model can hold across market cycles.