DePIN Unit Economics: The Missing Emissions Data

Introduction

The burn coverage ratio is the share of a network’s token burns funded by organic end-customer revenue rather than by treasury reserves, foundation purchases, or incentivized usage. As of late 2025, no outside analyst can compute it for GEODNET, Helium, or io.net, because none of the three publishes dollar-denominated emissions next to dollar-denominated revenue for the same period on the same surface.

What is available is the revenue leg alone. A three-month network fee comparison cited by SolanaFloor from Artemis data put Helium at roughly $4 million, GEODNET at $1.3 million, and Akash Network at $729,700. That ranks gross fee generation across three Solana-adjacent DePIN networks. It says nothing about what each paid in token issuance to produce those fees, which is the comparison required to establish whether the unit economics work.

This asymmetry characterizes DePIN disclosure. Revenue became legible across 2025: fee dashboards, daily burn counters, subscriber totals. The subsidy that produced that revenue did not. Every claim that a DePIN token has turned net deflationary currently rests on a chart selected and framed by the team making the claim, with the burn side denominated in dollars and the emission side denominated in tokens.

The three networks most often cited as evidence that the model works have responded to the same pressure in three structurally different ways. Helium routes subscriber revenue through a mechanical, non-discretionary burn. GEODNET reportedly commits a fixed share of data revenue to buybacks as a matter of policy. io.net is reported to have replaced its fixed emission schedule with an engine that keys issuance to measured network earnings. Those are a protocol rule, a management policy, and an algorithm, and they differ sharply in how much they would survive a bad quarter or a change of leadership. Two of the three cannot be verified from any source available to us.

What a burn proves and what it does not

Helium’s sink is the cleanest mechanism in the set, making it the appropriate place to establish what a burn demonstrates. Subscriber revenue is converted into Data Credits, non-transferable units priced at a fixed $0.00001 each and minted only by burning HNT. The fixed dollar price means the token quantity burned floats with the market price while the dollar value burned tracks usage. On top of that, the network runs a daily dollar-cost-averaged buyback executed through Jupiter Exchange, per the SolanaFloor account.

A burn of this kind is a sink tied to consumption rather than to a schedule, improving on discretionary buybacks announced at market peaks. It is also, strictly, a transfer. Revenue that would otherwise sit on a balance sheet is converted into supply reduction. That alters the token’s float. It does not establish that the underlying service earns more than it costs to deliver.

Burning gross revenue has a direct consequence. If 100% of subscriber revenue is routed into Data Credits, then the cost of delivering that service, the payments to hotspot operators, the hardware subsidy, and the wholesale capacity purchased from a carrier partner, is being paid in some other currency. In DePIN that other currency is token emissions. Burns and emissions are two legs of the same transaction, and a dashboard showing one leg tells an observer little. A network can burn $50,000 a day while issuing considerably more than $50,000 a day in supply-side rewards, and both facts can be true at once without contradiction.

The emissions-to-revenue ratio, rather than the burn figure, is the metric that resolves the question. The ratio asks how many dollars of new token supply a network must issue to produce one dollar of end-customer revenue, and whether that number is falling. A network whose ratio drops from five to two over a year is converting subsidy into business. A network whose ratio is stable is paying a permanent tax to maintain supply-side participation, and its burn program recycles the proceeds of that tax back into the token rather than funding anything new.

Three measurements, one of which is currently public

Of the three quantities needed to score DePIN unit economics, only gross fee generation is publicly observable today, and even that is reported through third-party aggregators rather than audited disclosure.

Hardware customer acquisition cost is the fully loaded cost of bringing one revenue-generating node online and keeping it there: device subsidy, installation and referral incentives, logistics, and the emissions paid to that node through its first year before organic demand reaches it. None of the three networks discloses this per node, and none publishes cohort data showing whether nodes deployed in one period are still online and still earning in the next. Churn matters here. A node that collects a subsidy for eight months and then goes dark has consumed acquisition cost without ever contributing to the revenue line.

Token subsidy per unit of real demand requires both legs and is therefore unavailable. The distinction between organic and synthetic demand matters here. Revenue counted at the protocol level can include ecosystem spending: foundation purchases of the network’s own service, grant-funded pilots, and usage that exists because someone was paid to generate it. Nothing in the available data separates end-customer demand from internally funded demand for any of the three networks, and in compute markets specifically, usage that exists to earn rewards is a documented failure mode rather than a hypothetical one.

Burn coverage ratio inherits both problems. A ratio built on a revenue figure that includes ecosystem spending will overstate coverage, and a ratio that ignores emissions describes supply mechanics rather than business quality.

There is also a comparability problem that better disclosure will not fully resolve. Helium’s fees come from consumer wireless subscriptions and reported carrier payments, GEODNET’s from enterprise correction-data subscriptions, and Akash’s from compute rental. These have different gross margins, different contract durations, different churn profiles, and different sensitivity to a single counterparty. Ranking them by gross fees is roughly as informative as ranking a mobile carrier, a data licensor, and a hosting provider by revenue without examining anything below the top line.

Helium Mobile’s fee stack and its single point of dependency

Helium’s reported figures contain an internal tension that points directly at where the revenue originates. The SolanaFloor article cites 505,505 sign-ups, roughly double the May level, alongside 113,891 hotspots and more than 1.2 million daily users. Daily users exceeding total sign-ups by a factor of more than two is not a contradiction if “users” counts devices served through carrier offload rather than Helium Mobile subscribers. That reading matches the reported shift toward carrier partners paying for traffic carried on Helium radios, though we could not verify the carrier arrangements, their onset, or their pricing terms from any source available to us.

The daily figures give a sense of scale. On October 26, per Artemis data cited in that article, Helium generated $57,900 in fees against $34,500 in buybacks, with HNT near $2.10 at the time, implying roughly 16,400 tokens purchased that day. Data Credit burns were reported running consistently near $50,000 per day, exceeding $52,000 on October 27. The source does not specify the year for those dates, and surrounding references point to 2025.

One line in that breakdown warrants attention. Of the October 27 burn, approximately $124 came from IoT usage. The LoRaWAN network that defined Helium for its first several years, that justified its original hardware build, and that absorbed the bulk of its early emissions now contributes on the order of a quarter of one percent of daily network fees. That is a concrete illustration of what fixed-emission bootstrapping can produce: a large deployed base, a functioning network, and minimal revenue. Whatever the mobile business becomes, the IoT segment is the control case, showing that density achieved through subsidy does not automatically convert into demand.

The reflexivity problem in net deflation

A Helium core team member cited a Dune chart reportedly showing revenue-funded burns outpacing emissions since late September 2025, described as net deflation for HNT. The mechanical problem with that framing is that the burn side is fixed in dollars and the emission side is fixed in tokens.

Because Data Credits are priced at a fixed $0.00001, a given dollar of subscriber revenue burns a quantity of HNT that is inversely proportional to the market price. If the token price halves with the business unchanged, the same daily revenue retires twice as many tokens, emissions stay on schedule, and net deflation deepens. If the price doubles, net deflation can flip to net inflation without a single subscriber leaving. A deflation claim measured in token units is therefore partly a statement about price, and reading it as a statement about business performance inverts the causality. The verifiable version of the claim is dollar-denominated: revenue burned per day versus the dollar value of emissions per day. That number has not been published.

Offload revenue as a concentration exposure

If a material share of Helium’s fee line comes from carrier offload, the network has traded one dependency for another. Consumer subscription revenue is slow to build and hard to defend against incumbent pricing, but it is diversified across hundreds of thousands of small accounts. Offload revenue is a wholesale capacity sale to a small number of large buyers who set the price, control traffic steering, and can renegotiate at contract renewal. The decentralized supply side changes who owns the radios. It does not change the fact that wholesale capacity is a commodity sold into a buyer’s market.

The scenario that would test this is clear: a carrier partner reduces offload payout rates or redirects traffic. Daily Data Credit burns fall with it, the buyback shrinks proportionally, and net deflation reverses without any failure of the network itself. We have no basis to estimate the probability of that scenario, because contract terms are not public. A revenue stream whose durability cannot be assessed from outside should not be underwritten as though it were recurring consumer revenue.

Reports that Helium suspended a discretionary burn program in order to fund growth could not be confirmed against any source available to us. If accurate, the structural reading is plain: a burn that can be suspended is a capital allocation decision, not monetary policy, and the distinction matters to anyone treating supply reduction as a durable property of the token.

GEODNET has the strongest claimed economics and the thinnest public record

GEODNET is the case where the gap between asserted and verifiable figures is widest. The claims circulating about it, roughly $7.3 million in annual recurring revenue, growth of about 154% year to date, and a policy of directing 80% of data revenue into buying and burning GEOD toward a net-deflationary target, are not supported by any source we could access. The Blockworks research report commonly cited for them returned a 404 at the time of writing. We flag them as unconfirmed rather than repeating them as findings.

The one externally sourced data point is GEODNET’s $1.3 million in fees over the cited three-month window. Annualized, that implies something in the range of $5 million, which is directionally compatible with a $7.3 million run rate given growth within the period. Compatibility is not corroboration. It narrows the range of plausible figures without confirming any of them.

The underlying business is the most differentiated of the three. Real-time kinematic positioning, or RTK, uses a network of fixed base stations to broadcast correction data that reduces GNSS error from meters to centimeters for nearby receivers. The value is a function of proximity, so coverage density is the product. Buyers are agricultural equipment operators, surveyors, robotics firms, and automated machinery operators, and they pay on recurring per-rover subscriptions rather than per-transaction.

That demand profile has two distinct properties. Recurring enterprise contracts churn more slowly than consumer subscriptions and do not reprice daily, which makes revenue more predictable than Helium’s. But the incumbent set is formidable: established positioning vendors, publicly funded continuously operating reference station networks, and satellite-delivered correction services that require no terrestrial base station. The token-subsidized thesis holds that emissions can fund base station density in regions where incumbents will not deploy, and that density is sticky once achieved. The counterargument is that correction data becomes a commodity priced per rover per month once density crosses the operational threshold, at which point an incumbent with existing enterprise relationships can match coverage and undercut on price without funding its own supply side.

A definitional question also hangs over the burn policy as described. A commitment to burn 80% of data revenue is not the same as 80% of total revenue, and DePIN networks that sell hardware typically book a meaningful one-time line alongside the recurring one. Without a revenue breakdown, an outside reader cannot tell whether the burn is funded by the subscription business or partly by device sales that recur only as long as the network expands. Nor can the growth rate be interpreted without a base: 154% on a small base is a common early-stage trajectory and carries no information about what happens at ten times the size.

io.net and the failure modes of earnings-linked emissions

io.net’s reported Incentive Dynamic Engine, said to have launched in June 2026 and to tie both emissions and burns to measured network earnings, could not be verified. The io.net blog post describing it returned no readable content when fetched, and no secondary source in our research covered its mechanics. What follows treats earnings-linked emissions as a design, not as a deployed system with observable results.

The design responds to an actual defect. A fixed emission schedule prices supply growth without reference to demand, which means a network pays the same subsidy whether its capacity is fully utilized or sitting idle. Replacing that with issuance keyed to measured earnings removes the mispricing. It introduces three specific problems.

The first is procyclicality. Fixed emissions are countercyclical: when demand is weak, the network continues paying operators to stay online, preserving capacity that becomes valuable when demand returns. Earnings-linked emissions cut rewards precisely when earnings fall, accelerating operator exit in a downturn. Capacity leaves when it is cheapest to retain and must be re-acquired at full cost later. Whether that trade is worth making depends on supply elasticity, and for GPU operators with alternative buyers, supply is elastic.

The second is measurement integrity. Any rule that mints tokens as a function of reported earnings creates an incentive to manufacture earnings. In a compute marketplace where the same party can rent and supply, wash renting to trigger issuance is an obvious attack vector, and defending against it requires revenue attestation independent of the rewarded parties. Nothing in the available material describes how io.net addresses this.

The third is lag and reflexivity. If the engine reads a trailing earnings window, emissions respond to demand shocks weeks late, which is tolerable for noise and damaging during regime changes. If it reads closer to real time, and earnings are measured in a volatile token, issuance responds to price as well as demand, amplifying feedback in both directions. We do not know which mechanism io.net selected.

Underneath the tokenomics sits the central demand problem of compute markets. GPU rental is close to fungible: buyers compare price per hour for a given accelerator, switching costs are low, and competing supply includes hyperscalers and centralized GPU clouds with higher reliability guarantees and conventional procurement paths. Decentralized compute competes mainly on price, which caps gross margin structurally. The single fee figure available for the category is Akash at $729,700 over three months, less than a fifth of Helium’s over the same window. If that reflects category scale, the primary issue for compute DePIN is not emission design. Demand has simply not materialized at the scale the supply build assumed.

Scoring on what can actually be checked

The three networks cannot be ranked on unit economics because the emissions leg is missing for all of them. They can be ranked on two observable factors: gross fee scale and the structural permanence of the burn commitment.

The second axis is the more diagnostic one. A burn enforced at the protocol level, as Helium’s Data Credit conversion is, executes irrespective of management preferences, survives changes of leadership, and can be verified on chain. A burn set by policy, as GEODNET’s is reported to be, is a discretionary allocation of operating cash that can be reduced in any quarter where capital is needed elsewhere. That is often a defensible commercial decision, which is precisely why a policy burn should not be underwritten as an immutable monetary property. An algorithmic burn tied to earnings sits between the two: structural in form, but only as credible as the earnings measurement feeding it and the governance process that can modify the formula.

Ranked on structural credibility, Helium’s mechanism is the strongest, GEODNET’s is the weakest, and io.net’s is unassessable. That ordering runs inverse to the ordering by claimed economics, where GEODNET’s reported margins and io.net’s reported engine appear more compelling than Helium’s daily fee volume. The networks with the most attractive narratives possess the least verifiable infrastructure.

What disclosure would settle, and what to watch instead

The data required to make DePIN unit economics auditable is straightforward. Four disclosures would address the gap: the dollar value of emissions published beside the dollar value of revenue for the same period on the same dashboard; revenue split between end-customer demand and ecosystem or grant-funded usage; node cohort data showing how many devices deployed in a given quarter remain online and earning twelve months later; and, where a material share of revenue comes from a small group of counterparties, customer concentration disclosure. None of this requires audited financial statements. It requires publishing the cost leg with the same prominence as the revenue leg.

Until that exists, three observable conditions offer practical signal. The first is whether Helium’s daily fee run rate holds across a carrier contract renewal cycle rather than across a single quarter of user growth, since the primary downside is a renegotiation that reduces burns without an operational outage. The second is whether GEODNET’s reported 80% burn commitment survives a flat quarter, which is the condition that separates a policy from an immutable rule. The third is whether io.net publishes the IDE’s earnings measurement and attestation specification, because an issuance formula keyed to unverified revenue functions as an unconstrained subsidy.

The Helium IoT line remains the central case study. A network bootstrapped by fixed emissions to over a hundred thousand deployed devices now generates roughly $124 a day from that original segment. The hardware exists, coverage exists, and demand never arrived at the scale the subsidy assumed. Revenue-linked tokenomics alters how a network compensates supply. It does not produce buyers.