DePIN Mining Explained: Not PoW, and the Honest Economics

Bottom line: DePIN mining pays you for a physical resource, not for a calculation

DePIN mining means earning tokens by contributing a real-world resource to a network — wireless coverage, storage capacity, GPU time, map data, weather readings — and having that contribution verified. It borrows the word "mining" from Bitcoin, but there is no hash race, no competition to find a solution first, and electricity is not the main cost. What you are running is closer to a piece of network infrastructure than to a mining rig.

Now the part most guides on this topic will not tell you. As of 2026-07-16, the number you would need to work out whether DePIN mining pays — a current per-device reward rate — is not public. Nearly every article promising a monthly figure is affiliate-grade guesswork. And on Helium — the flagship DePIN, the network that made "proof of coverage" famous — the reward mechanism sitting at the centre of all that coverage has been voted out.

How it differs from PoW and staking

Put simply: proof-of-work converts electricity into computation, staking locks up capital, and DePIN rents out something physical. It has something in common with running a blockchain node, but the basis for payment is different in kind — you are paid for being somewhere and doing work there.

What the "miner" actually is

At RAKwireless, an official Helium Maker, every purchasable RAK Hotspot V2 variant is $204 (as of 2026-07-16). The store advertises a $179–$324 range, but neither endpoint can actually be bought: all seven $324 4GB variants are sold out, and so is the $179 SKU (the KR920 2GB). What is in stock is six 2GB variants — EU868 (UK and Europe), US915, AU915 and AS923 (UK and US) — all at exactly $204. Budget from the range's floor and you are anchoring on a price no one can pay.

RAK describes the power draw as about 5 watts — "less than your standard internet router". Run continuously, that is roughly 3.6 kWh a month. In DePIN the dominant cost is not electricity; it is the hardware and the location. The proof-of-work mental model does not transfer.

Worth noting for the reader outside the US and Europe: the listed regional variants are EU868, US915, AU915, AS923 and KR920. There is no Japan-specific variant listed — which matters more than it sounds, as the regulatory section below explains.

How the work gets verified

If you could fake the work, none of this would function. Each project solves that differently — and the variety is the point. These are neutral examples, not recommendations.

Filecoin gives every 32 GiB or 64 GiB sector a 24-hour proving period, split into 30-minute deadlines — 48 a day — each answered with a zk-SNARK proof. WeatherXM computes rewards daily, bundles them into a Merkle tree and submits the root hash to a reward pool. Hivemapper directs 90.5% of newly minted HONEY to contributors (80.5% map coverage, 10% editing and QA), with 9.5% as an operational reward.

So "proof of coverage" is only one verification design among several — a detail that most explainers flatten. On Helium, everything that earns is represented as an NFT on Solana; oracles track each entity's lifetime rewards and a lazy-distributor program tracks what has been claimed (Helium docs, as of 2026-07-16). Because distribution is automated in code, smart contract risk is DePIN participant risk. Knowing how to read a block explorer lets you verify your own rewards rather than trust a dashboard.

The 2026 change that invalidates most guides

This is the heart of it.

Helium's HIP-149, "Helium Utility and Emissions Realignment" (start-date 2026-06-02), states as its third decision: "Retirement of Proof-of-Coverage on both Mobile and IoT. Mobile data deployers earn pro-rata of rewardable bytes; the Service Provider allocation stays unchanged at 24%. IoT data transfer continues at the existing $/DC peg." Its status as of 2026-07-16 is Approved — which Helium's own repo defines as "approved by consensus, and is pending development, testing or deployment". PoC emissions stop "when the program upgrade ships (at Council seating)". It has passed; it is not yet live.

The reasoning is stated bluntly in the proposal itself: "Proof-of-Coverage pays for existing rather than for serving subscribers. The work that grows the network now is offload utility, not coverage proofs." The protocol concluded that paying people merely to exist as coverage was not producing value.

The consequence for operators is equally blunt: "Hotspots earning predominantly through PoC today will see their rewards drop to zero from the PoC bucket and depend on data utility." In its own Drawbacks section the HIP concedes: "Some deployments that brought network value primarily through PoC may not be sustained on data utility alone."

But reporting only the retirement misreads the HIP

The same HIP's Decision 1 creates something in PoC's place for Mobile data deployers: a band tied to the per-GB payer rate, with a target minimum at half that rate (currently $0.05/GB) and a cap at three times it ($0.30/GB). The proposal calls this "downside protection from the target minimum (binding at current HNT prices)", and adds: "HNT trades below the floor threshold today (≈$0.27), so the target minimum is already binding and is the protection that matters now." The accurate picture is therefore two-sided: PoC goes, but Mobile data deployers gain a revenue-linked floor.

Three caveats keep that from being a promise:

1. The HIP itself says the target minimum "is what the protocol aims to deliver, not a hard guarantee every epoch. After a sharp drop in the HNT price, the amount delivered can fall short for 1 to 2 weeks, until the burn average catches up."
2. The band is Mobile-only. HIP-149 states that "IoT data transfer is unaffected by Decision 1 and continues on the existing $/DC peg." A LoRaWAN unit like the RAK Hotspot V2 above sits on the IoT side — it loses PoC and gains no floor.
3. It still does not yield a per-device number. $0.05/GB is a floor per gigabyte; how many gigabytes one device carries in a month is not published.

There is corroborating structural evidence. Helium's live documentation contains no page at all for proof of coverage, mining, or earnings — 69 non-blog pages in its sitemap, organised around LoRaWAN operation, Wi-Fi conversion, tokens and network data (measured 2026-07-16). The flagship DePIN's own documentation has stopped marketing a "mining" framing.

The honest economics

Why aren't early-Helium-style returns repeatable? Not as an opinion — the protocol's own documents give the mechanism, with numbers.

1. Dilution. From HIP-149: "Over the past year, network rewardable bytes grew ≈4× (from ≈24K GB/day in June 2025 to ≈97K GB/day in April 2026) while HNT issuance stayed fixed under the HIP 20 schedule and HNT price more than halved." More contributors, fixed issuance, falling price. Each participant's share falls structurally.

2. Density scaling — though this one is transitional. HIP-17 (status: Deployed) reduces rewards for hotspots in dense areas by multiplying scaling factors across H3 hex resolutions — because where hotspots sit close together, "very little additional coverage is demonstrated". Note, though, that HIP-149 lists HIP-17 itself among the "HIPs retired by IoT PoC removal" — this scaling dies with PoC, so unlike dilution it is not a durable structural reason. The conclusion survives either way: today HIP-17 scales your second unit down, and after PoC retires the PoC reward is zero regardless. Adding a second unit next to your first does not double anything.

3. The token. HNT traded at roughly $0.21 as of 2026-07-15, against an all-time high of $54.88 set on 2021-11-12 — about −99.6% (CoinGecko). Market cap is around $38M. The conditions that produced the famous returns no longer exist.

4. The rate has been cut. The widely quoted "$0.50/GB" is stale. HIP-149 records that, separately from the vote, Nova reduced the payer rate from $0.50/GB to ≈$0.10/GB under existing HIP-143 authority, "reflecting current commercial offload rates". The $0.05 floor and $0.30 cap above are derived from that $0.10. Any article quoting $0.50/GB as current is out of date on its face.

5. More issuance is coming. HIP-149's second decision authorises ≈141M HNT over 36 months — about 77% of current on-chain supply (≈182.5M) — front-loaded with ≈50% in the first 12 months, minted to a Squads multisig vault administered by Nova Labs. Effective max supply rises from ≈206M to ≈347M, and the HIP lists "Dilution" in its own Drawbacks. HNT's long-advertised max-supply property is not preserved. Note that public trackers lag this: CoinGecko still displayed a 223,000,000 max supply as of 2026-07-15 — a useful reminder that a tracker's number is not automatically current.

6. Hardware can be voted obsolete. Under HIP-139 (Deployed), an entire hardware class was retired: "All CBRS radios will go offline and stop receiving rewards by March 1st, 2025." Owners were offered replacement Wi-Fi hotspots through foundation grants and could re-flash their equipment to stock firmware and keep it — "CBRS owners may keep their stock CBRS radios", which "can be sold or applied for a variety of use cases" such as rural fixed wireless and private LTE/5G, with prepaid SAS fees refunded pro-rata. But there was no cash reimbursement for the stranded investment. This is not a hypothetical risk; it has already happened once.

So what does a device actually earn? I could not find a primary source that publishes a current per-device rate. The only protocol-sourced per-device number available is a modelled table in HIP-147 (Deployed): under legacy HIP-53 rules, with 30,000 hotspots, 40TB of daily data transfer and an assumed HNT price of $1, the average PoC reward per Helium Mobile hotspot was $0.12 per day. That is a modelled scenario, not a promise and not a current figure — HNT is actually around $0.21, and that PoC bucket is the one being retired.

Without the current per-device reward rate, a payback period cannot be calculated — so this article will not calculate one. Any article that shows you one has manufactured an input somewhere.

The checks most guides skip: radio law and tax

Two questions get almost no coverage in English-language DePIN guides. Japan makes a good worked example, and the underlying lesson transfers to any jurisdiction.

Your DePIN hotspot is a regulated radio transmitter. In Japan, using a radio device without the 技適 (technical conformity) mark can breach the Radio Act. The Ministry of Internal Affairs and Communications states the penalty directly: imprisonment of up to 1 year or a fine of up to ¥1,000,000 — and where a device interferes with a radio station of high public importance, up to 5 years or ¥2,500,000. Two specifics land squarely on common DePIN practice:

• Modification voids certification. MIC (answer 10): modifying a certified device nullifies the technical conformity certification, and the mark must be removed; using such a device risks breaching the Act. Swapping in a higher-gain antenna to chase rewards — routine advice in DePIN communities — can fall here.
• The mark, not the passport, is the test. MIC states that "most foreign-made radios without the 技適 mark do not follow the rules for use in Japan" (answer 27) — but also, in the same answers, that "even among foreign products, most models that can be used in Japan carry the mark" (answers 21 and 27), and that "some transceivers obtained through personal imports cannot be used in Japan, so please check whether the mark is present" (answer 21). The question is not whether the gear is foreign; it is whether it carries the mark. Most DePIN hardware is bought from overseas stores — a personal import — so that check is yours to make.

This article does not assert whether any specific model is certified. MIC runs an official searchable database of certified equipment — check the maker, model and certification number yourself before buying. That is the actual first step, and it takes minutes.

On tax, the timing is what hurts. Japan's National Tax Agency is explicit: where crypto assets are acquired through mining, staking or lending, the market value at the moment of acquisition is included in gross revenue, with associated costs deductible as expenses. You are taxed on what the tokens were worth when you received them — so if the token then falls (recall HNT at −99.6% from its high), the liability does not fall with it. You can owe tax on value you no longer hold. Profits are in principle miscellaneous income; where crypto revenue for the year exceeds ¥3,000,000, it becomes business income if you keep books and records, or business-related miscellaneous income if you do not. Hardware lasting over a year above a certain value generally cannot be expensed in year one — it must be depreciated (NTA FAQ, December 2025 edition). Your own jurisdiction will differ, but taxed on receipt is a common enough rule to check before you plug anything in.

Scams and markups

DePIN is easy to pitch as "buy the box, collect money" — which makes it excellent bait. Here are checkable anchors rather than vibes.

• Legitimate manufacturers are published. Hivemapper's docs state that "Bee Maps is the only certified manufacturer of devices for the Hivemapper Network" and point to a list of authorized resellers. Helium runs its approved Maker programme in the open: under the HIP-19 approval process, applicants fork a public template, submit a proposal, pass a hardware audit ($3,000–$8,000) and Helium Foundation KYC ($2,000), and are reviewed by the Manufacturer Compliance Committee (MCC), with application status published — see dewi-alliance/hotspot-manufacturers (as of 2026-07-16). An "official-looking" seller can be checked, not trusted.
• The projects themselves promise nothing. Hivemapper's documentation publishes no earnings figures at all. A reseller quoting guaranteed monthly income is asserting something the project does not.
• A live example of the markup dynamic. As of 2026-07-16 the official Hivemapper/Bee Maps store (shop.beemaps.com) is password-gated and reads "Opening soon. Be the first to know when we launch." The official shop is not selling — while third-party listings market the device with "HONEY bonus" and pre-order pitches. Buying official hardware from unofficial sellers at a markup is exactly the pattern to distrust. (The only citable Bee price is the 2024-02-21 launch pricing of $549 LTE / $449 Wi-Fi — not a current price.)
• Vendor marketing lags the protocol. As of 2026-07-16, RAKwireless's RAK Hotspot V2 page is still titled "Mining for Helium HNT rewards" and advertises that the device "Earns Helium HNT ($HNT) tokens through Proof-of-Coverage and data transfer" — while PoC retirement was approved on 2026-06-02. A product page is not a current source of truth.

In Japan, the Consumer Affairs Agency publishes a standing warning about crypto-related trouble jointly with the FSA and National Police Agency, and complaint volumes recorded by the National Consumer Affairs Center have stayed high: 5,625 in FY2022, 8,496 in FY2023 and 7,227 in FY2024. (A figure of 751 appears for FY2025, but it counts only up to 2025-05-31 — a partial year, not comparable to the full-year figures above.) Reported entry points include "easy side income" pitches on social media. General defensive habits are collected in our scam defense guide.

So — is it worth it?

The honest answer is that the inputs required to answer that are not publicly knowable right now.

That is not hedging. Computing a payback period requires a current per-device reward rate; no primary source publishes one, and on the flagship network the reward category itself is queued for removal. You cannot divide by a denominator that is mid-flight.

What can be said, as decision material:

• The hardware cost is a certain, up-front outlay ($204 for a RAK Hotspot V2, as of 2026-07-16). The rewards are not certain.
• Electricity is not the constraint (~5W). Do not import proof-of-work instincts here.
• Downside is not always floored at zero. Filecoin's storage-proving documentation states that failing to submit a WindowPoSt proof within the 30-minute deadline — or at all — results in slashing, with a portion of the collateral forfeited to the f099 burn address and the provider's storage power reduced. But the docs also publish exceptions: an honest provider missing a single WindowPoSt incurs no penalty in the current proving period, and there are no fees if the sector is recovered in a later one (FIP-0002, "Free Faults on Newly Faulted Sectors of a Missed WindowPoSt"); penalties apply only to sectors already faulty from previous proving periods. So one missed deadline is not an instant confiscation. A sustained fault is different: fault fees accrue for each day a sector is offline, equal to 3.51 days of expected block rewards, and continue "until the associated wallet is empty and the storage provider is removed from the network". Some DePIN can cost you money, not merely fail to make it.
• Space is finite and cities are the worst of both worlds. WeatherXM caps how many stations can earn in a cell: exceed the cell's capacity and only the top-ranked stations are paid, and relocating resets a station's PoL score to zero along with its seniority. Meanwhile RAK notes that flat rural areas can reach 10 miles or more while "high-density urban areas typically see coverage ranges up to a mile". Dense cities are simultaneously the most saturated and the shortest-range environments.

The most interesting signal may be that DePIN's own commercial story is moving. Bee Maps — Hivemapper's certified manufacturer — now positions itself around "Data for Physical AI", real-world video and imagery for training. The centre of gravity is drifting from a machine that earns toward a network that collects training data (see AI and crypto). If you engage with DePIN at all, the more truthful question is probably not "how much this month" but "is there a buyer for this data?" — which is, almost word for word, what HIP-149 concluded: what grows a network is utility, not proofs of existing.

Sector-wide activity can be tracked on cross-project explorers such as DePINscan. This article does not quote aggregate figures it could not measure directly.