Helium: The Coverage Stayed. The Demand Was Sold.
Helium built the largest crowdsourced wireless network in history — then dismantled the demand side. In 18 months: sub-DAO tokens unwound, CBRS 5G traded away, T-Mobile denying the offload deal, the free plan killed, Helium Mobile sold. What's left is coverage — and a $41.7M token.
Helium built the largest crowdsourced wireless network in history — and then, in eighteen months, dismantled almost everything it was built to serve. The CBRS 5G ambition was traded away. The sub-DAO tokens were unwound. The free plan was killed as « not sustainable ». The carrier-offload deal Helium touted with T-Mobile was publicly denied by T-Mobile. The buy-and-burn subsidy was suspended. And on June 2, 2026, Nova Labs sold Helium Mobile to Andrew Yang’s Noble Mobile. What survives is the coverage — and a $41M token for a network that once promised to replace the telcos.
Last update: July 11 2026 · Helium (HNT) / Ecosystem · By Cache256 Intelligence
In August 2025, this analysis described Helium as invisible infrastructure: 400,000+ hotspots, 1.16M daily users, IoT and 5G connectivity dissolving into enterprise operations. That framing is now obsolete, and the reason it is obsolete is the story.
Helium did not fail at building supply. It succeeded spectacularly. It is, by deployment, the largest decentralized wireless network ever assembled — a global LoRaWAN footprint plus 140,957 active Mobile hotspots. What Helium failed at is the other half of the equation: getting anyone to pay for it at scale. And between January 2025 and June 2026, rather than close that gap, the project systematically retired the ambitions that depended on closing it.
This is the DePIN supply-demand gap in its most instructive form — the same structure Cache256 documented in Filecoin, where storage power was never storage demand. Helium is the wireless edition, and it is further along the curve: the retreat is not hypothetical. It is documented, dated, and in the public record.
// THE EIGHT RETREATS (2025–2026)
// HISTORY 2019–2026
2019–2021 — The People’s Network
Helium launches LoRaWAN coverage rewarded by Proof-of-Coverage (PoC). Hotspot demand outstrips supply; hardware resells at multiples of MSRP. Deployment peaks near ~916,000 hotspots by August 2022 — a crowdsourced buildout with no precedent.
2022 — The sub-DAO architecture
HIP-51/52/53 create the Helium DAO plus IOT and MOBILE sub-networks, each with its own token. The thesis: let every network price its own economy.
April 2023 — Migration to Solana
The network migrates to Solana. Fees collapse; Pyth oracles price the HNT/DC conversion. The migration is executed cleanly and remains stable through 2026 — one of the few unambiguous successes.
2023 — Helium Mobile launches as an MVNO
Helium Mobile rides a wholesale agreement on T-Mobile’s macro network, supplemented by community CBRS/Wi-Fi hotspots. The famous $20/month unlimited plan — and later the free Zero Plan — drive signups.
Nov 2024 – Jan 2025 — HIP-138: Return to HNT
The sub-DAO experiment is reversed. MOBILE and IOT emissions end January 29 2025. Governance is unified under HNT/veHNT (HIP-141).
Feb 2025 — CBRS Trade-Up ends (HIP-139)
Deployments shift from CBRS 5G toward Wi-Fi hotspots.
Apr 2025 — AT&T Wi-Fi offload partnership
A real, named carrier partnership for Wi-Fi offload. Later joined by Movistar / Telefónica in Mexico and a Mambo WiFi partnership in Brazil.
Jun 2025 — The T-Mobile denial
T-Mobile publicly states it has no carrier offload agreement with Helium.
Jun 2026 — The divestiture
Noble Mobile acquires Helium Mobile (June 2). The Zero Plan is terminated (June 11). Nova Labs retreats to the data-offload business.
// TERMINAL
user@cache256:~$ helium status --detail
Two networks, one token
▸ IoT (LoRaWAN) — global coverage, unlicensed ISM bands. Demand: a few hundred USD/day in DC burn.
▸ Mobile (Wi-Fi offload) — 140,957 active hotspots; carries ~99.8% of Data Credit burn.
▸ Token: HNT only since Jan 2025 (MOBILE/IOT emissions ended; legacy tokens still tradable, near-zero).
Mechanism
▸ Proof-of-Coverage rewards coverage, not billed traffic — the structural gap.
▸ Data Credits ($0.00001 fixed) minted by burning HNT → Burn-and-Mint Equilibrium.
▸ Discretionary buy-and-burn of HNT from subscriber revenue: suspended (early 2026).
What Nova Labs still owns
▸ The hotspot network + the carrier data-offload business (AT&T, Movistar/Telefónica Mexico, Mambo Brazil).
▸ Not: the consumer carrier (sold to Noble Mobile, June 2 2026).
system@cache256:~$ echo "Status: supply intact, demand divested — the offload layer is the remaining business"
// CORE MECHANISM — AND WHERE IT BREAKS
- Proof-of-Coverage (PoC) — Hotspots earn HNT for verifiable coverage: beacons, witnesses, location. This is the design decision that explains everything downstream. PoC pays for presence, not for traffic. A hotspot in an empty field with no paying user earns; a hotspot carrying real offload earns too. The incentive is coverage-first by construction, which is precisely why supply raced ahead of demand.
- Data Credits & Burn-and-Mint — Usage is paid in Data Credits, fixed at $0.00001 and minted by burning HNT. In theory this creates a clean flywheel: more usage → more burn → HNT scarcity. In practice, the burn is dominated by Mobile (~99.8%) and the IoT network — the larger deployment by far — contributes a few hundred dollars a day.
- The discretionary burn — Nova Labs ran an experiment routing subscriber revenue into HNT buys-and-burns. It made the burn chart look like organic demand. It was suspended in early 2026, which is the honest disclosure that it was a subsidy, not a flywheel.
- The MVNO underneath — Helium Mobile’s nationwide coverage was never the hotspots. It was a wholesale agreement on T-Mobile’s macro network. Community hotspots offloaded a portion of traffic, lowering wholesale cost where they existed. The decentralized network was the margin improvement, not the service.
// WHAT ACTUALLY REMAINS — THE OFFLOAD LAYER
Strip away the consumer carrier, the sub-DAO tokens, the CBRS ambition and the free plan, and one business is left standing — the one Nova Labs explicitly said it is refocusing on: selling data offload to carriers.
Named, verifiable partnerships
▸ AT&T — Wi-Fi offload partnership (April 2025). Real and named.
▸ Movistar / Telefónica — Helium Mobile Hotspots deployed in Mexico (Mexico City, Oaxaca) to extend coverage and offload mobile data.
▸ Mambo WiFi — Brazil, ~40k access points (announced late 2025).
The one that is not a partnership
▸ T-Mobile — the MVNO wholesale relationship is real. The carrier offload agreement Helium implied is not: T-Mobile denied it on the record. Helium’s explanation — an unnamed « third-party wireless company » — has not been substantiated.
This distinction matters more than any metric on this page. A DePIN network that sells offload to carriers has a business. A DePIN network that implies carrier relationships it does not have is doing marketing. Helium is currently doing both, and the reader is entitled to know which is which.
// METRICS
Reading the table honestly. Active-hotspot counts vary by source and by definition of « active » (deployed vs witnessing vs earning): helium.com promotes « over 358k »; independent tallies put total active nearer 184k (≈141k Mobile + ≈43k IoT); one May 2026 count reports ~376k. We anchor on the one hard, directly-observable figure — 140,957 Mobile hotspots — and flag the rest. What is not ambiguous is the demand side, and that is the number that decides the thesis.
// HIDDEN INFRASTRUCTURE
The LoRaWAN layer is real — and nearly unused. Helium operates the largest contiguous decentralized long-range IoT network on earth. It works. Sensors connect. It is genuinely useful infrastructure. It is also, by its own team’s admission, generating a few hundred dollars a day. Real ≠ demanded.
The Mobile layer is real — and it is a cost-reduction product, not a network. Where Helium hotspots exist, carriers can offload traffic more cheaply than building or roaming. That is a legitimate, sellable service — it is the business Nova Labs kept. But it is a margin play inside someone else’s network, not a replacement for one.
The token layer carried the rest. PoC emissions built the coverage. The buy-and-burn made the demand chart look better than it was, until it was suspended. The sub-DAO tokens were supposed to price each economy and instead lost ~99% before being retired. The flywheel had real parts; it was spun by subsidy.
What Helium proves, and it is not nothing. Token incentives can bootstrap physical infrastructure at a speed and capital efficiency no telco can match. ~916,000 radios deployed by strangers, at their own expense, in under three years. That is a genuine discovery about how infrastructure can be financed. The unresolved half is whether anything makes people pay for it afterward.
// WHAT FAILS — HONEST AUDIT
- PoC pays for coverage, not usage — the original sin. A reward function indexed to presence produces presence. It produced ~916k radios and a few hundred dollars a day of IoT demand. Later HIPs (usage correlation, active-data reward pools) try to bolt demand onto the incentive; they are patches on a design decision.
- The sub-DAO experiment failed outright — MOBILE ~−98.8%, IOT ~−99%, emissions terminated after ~2 years. HIP-138 is framed as simplification. It is also a reversal.
- Structural MVNO dependency — nationwide Mobile service ran on T-Mobile wholesale. The hotspots were the discount, not the network. And the carrier-offload relationship Helium implied with T-Mobile was denied by T-Mobile on the record.
- The free tier was the growth engine — and it was unsustainable — ~596k cumulative signups were inflated by a $0 plan discontinued as « not sustainable long term ». Post-conversion paying base and churn are the numbers that matter, and they are not public.
- The demand-side business was sold — the clearest verdict available on Helium Mobile’s standalone economics is that Nova Labs divested it and called the remaining offload business « more economical ».
- Spectrum and regulatory exposure — the CBRS 5G route was abandoned (HIP-139). The FCC’s ongoing Part 96 review could raise power limits in ways that favor incumbents over low-power shared-spectrum deployments.
- Governance concentration — veHNT voting on the economics-critical HIPs shows Nova Labs proxies and aligned large holders controlling a majority of participating votes. The DAO is real; the distribution is not diffuse.
// COMPETITIVE LANDSCAPE
// VERDICT
Scalability — Proven, on the wrong axis. Token incentives scaled physical deployment faster and cheaper than any telco could. They did not scale a single paying customer.
Adoption — Supply-side: historic. Demand-side: the IoT network burns a few hundred dollars a day, and the consumer carrier was sold. What remains is a real but modest carrier-offload business (~$3.5M annualized on-chain revenue).
Token economics — Unwound and repriced. Sub-DAO tokens terminated after ~99% drawdowns; the buy-and-burn subsidy suspended; HNT at $41.7M for a network of hundreds of thousands of radios. The Burn-and-Mint design is sound; the burn is too small.
Decentralization — Genuine at the physical layer (community-owned hardware, Solana execution). Concentrated at the governance layer (Nova Labs proxies + aligned whales carry the economics-critical votes).
Regulatory — The CBRS route was abandoned rather than defended. IoT rides unlicensed ISM bands with low regulatory friction and low demand. The FCC Part 96 review is a live variable for anyone still betting on shared-spectrum density.
// TRAJECTORY 2026–2027
Helium’s next two years are no longer a story about growth. They are a story about whether the remaining business — carrier data offload — is large enough to justify the token, the emissions and the hotspot base that were built for something else.
The variables that decide it
▸ Does offload revenue scale materially beyond ~$3.5M annualized, now that the discretionary burn no longer flatters the number?
▸ Does the Noble Mobile relationship convert into a durable, arm’s-length offload contract — or does Helium lose its consumer distribution entirely?
▸ Do named carrier deals (AT&T, Telefónica) expand, and does Helium ever substantiate the T-Mobile claim?
▸ Does the IoT network find any demand, or does it remain a functioning network nobody pays for?
▸ Do HNT emissions compress toward the size of the actual burn, or does dilution continue?
Closing assessment
Helium is not a fraud and it is not a failure of engineering. It is a failure of the demand assumption — and, to its credit, a project that has responded by cutting the ambitions it could not fund rather than pretending otherwise. Killing the free plan, unwinding the tokens, suspending the subsidy and selling the carrier are not scandals. They are disclosures. Read them as such.
The open question for the whole DePIN category is the one Helium is now answering in public: when the token stops paying for the coverage, does anyone pay for the coverage?
// FAQ
Is Helium dead?
No. The network runs, the hotspots earn, the offload business is real and has named carrier customers (AT&T, Telefónica Mexico). What ended is the maximalist story: decentralized 5G, sub-DAO economies, a consumer carrier. Nova Labs now operates a data-offload business on top of a crowdsourced radio network.
Who owns Helium Mobile now?
Noble Mobile, Andrew Yang’s low-cost carrier, which acquired it on June 2 2026. Noble takes the subscriber base and will run on the Helium Network. Nova Labs keeps the network and the offload business.
What happened to MOBILE and IOT tokens?
Emissions ended January 29 2025 under HIP-138 (« Return to HNT »). Rewards reverted to a single HNT. The legacy tokens remain tradable at near-zero value; holders can swap to HNT in the Helium app.
Does Helium have a T-Mobile offload deal?
T-Mobile says no. On the record, to Light Reading: « T-Mobile does not have a carrier offload agreement with Helium Network. » Helium says T-Mobile subscribers reach its network via an unnamed third-party partner. The MVNO wholesale relationship — which is how Helium Mobile got nationwide coverage — is a different thing, and it is real.
Why did the free Zero Plan end?
Helium Mobile stated it was « not sustainable long term ». It ended June 11 2026; users default to the $15/month Air Plan. The plan was the primary driver of the headline signup number.
Is the IoT network used?
Barely, in economic terms. Helium’s own blog notes the LoRaWAN network « has yet to generate significant network demand », with daily Data Credit burns limited to a few hundred dollars. The coverage is real; the demand is not there.
Should I still run a hotspot?
Cache256 does not give financial advice. The relevant facts: PoC still rewards coverage; HNT is at $0.22 with a $41.7M cap; emissions continue while measured demand is small; and returns depend heavily on placement. Model your own electricity, hardware and backhaul costs against those numbers.
// REGULATORY & SPECTRUM
CBRS (US, 3.5 GHz shared) — Helium Mobile’s 5G ambition ran on CBRS’s three-tier shared framework. The CBRS Trade-Up Program (HIP-139) closed in February 2025, migrating deployments toward Wi-Fi. The FCC’s ongoing Part 96 review (power limits, out-of-band emissions) is a live variable: higher power limits would favor incumbent carriers over dense low-power community deployments.
LoRaWAN / ISM (unlicensed) — The IoT network operates in unlicensed ISM bands (e.g. 915 MHz US). Low regulatory friction, duty-cycle and interference constraints. Easy to deploy; that is precisely why the supply glut was possible.
EU / international — Unlicensed ISM is broadly permissive under CEPT/ECC frameworks; licensed or shared-spectrum cellular models are far more nationally fragmented. Helium’s carrier traction to date is US + Mexico + Brazil.
« Spectrum as a service » — Helium demonstrated that crowdsourced coverage on shared and unlicensed bands can be sold to carriers. It did so within existing regulatory categories; it did not create a new one. The precedent is real and under-appreciated — and it survives the retreat.
// REFERENCES — EXTERNAL
▸ Helium — helium.com · world.helium.com (network explorer) · docs (accessed 2026-07-11)
▸ HIP-138 — « Return to HNT » (GitHub) · Helium — Path to Simplicity (accessed 2026-07-11)
▸ HIP-139 — CBRS Trade-Up Program (closed Feb 2025, accessed 2026-07-11)
▸ Helium Blog — Ecosystem Health: IoT, LoRaWAN, DePIN & 5G usage (« yet to generate significant network demand », accessed 2026-07-11)
▸ Light Reading — Helium teases offload from T-Mobile, but there’s no formal deal (accessed 2026-07-11)
▸ Fortune — Andrew Yang’s business acquires Helium Mobile · Fierce Network · Helium Mobile — Noble (June 2 2026, accessed 2026-07-11)
▸ Android Authority — Helium Mobile ends the free Zero Plan (« not sustainable long term », June 11 2026, accessed 2026-07-11)
▸ Fierce Network — AT&T partners with Helium for Wi-Fi offload (April 2025) · Telefónica & Nova Labs — Mexico (accessed 2026-07-11)
▸ Messari — State of Helium Q4 2025 · CoinMarketCap — HNT (accessed 2026-07-11)
// READING — RELATED CACHE256
The category map. Helium is its most advanced test case — and its most legible warning.
The same gap, in storage. Supply rewarded, demand absent. Read the two together.
The IoT DePIN peer. Same unlicensed bands, same demand question.
The comparator Helium’s own 2025 fiche used. Token-incentivized supply, again.
Where DePIN devices are supposed to transact. Helium is the hardware precedent.
Helium’s execution layer since April 2023 — the one migration that worked cleanly.
DePIN compute. Supply-side incentives, demand-side reality check.
The marketplace model: match supply to paying demand, or don’t build it.
// CRITICAL BALANCE
What we credit. Helium proved that token incentives can finance physical infrastructure at a speed and capital efficiency no incumbent can match — ~916,000 radios deployed by strangers, at their own cost. The Solana migration was executed cleanly. The AT&T and Telefónica offload partnerships are real, named and verifiable. Killing the free plan and selling the carrier were honest decisions, not evasions.
What we will not soften. The T-Mobile offload relationship Helium implied was denied by T-Mobile on the record, and the substitute explanation — an unnamed third party — remains unsubstantiated. The IoT network, the largest deployment of its kind on earth, generates a few hundred dollars a day. The sub-DAO tokens lost ~99% before being retired. The buy-and-burn that made the demand chart look organic was suspended.
What we do not claim. We do not claim Helium is worthless, fraudulent, or finished. We do not have its unit economics — nobody outside Nova Labs does. We do not know the paying-subscriber base post-Zero-Plan. We flag those gaps rather than fill them with estimates dressed as facts.
Position. Helium is subsidized coverage with a real but small offload business inside it. It is not, at current economics, invisible load-bearing infrastructure. It could become one. It is not one today.
// DATA UNCERTAINTY NOTES
▸ Active hotspot totals conflict by source and definition — helium.com promotes « over 358k »; a July 2026 independent tally gives ~184k active (≈141k Mobile + ≈43k IoT); a May 2026 count reports ~376k. « Active » may mean deployed, witnessing or earning. We anchor on 140,957 Mobile hotspots (world.helium.com) and treat the rest as a range.
▸ Paying subscribers are not public. ~596,000 is cumulative signups to end-2025 and includes the free Zero Plan. Post-June-2026 paying base, conversion and churn are unpublished. Any figure claiming otherwise is an estimate.
▸ Revenue figures vary by scope. On-chain explorer revenue ≈$9.5k/day; Messari’s Q4 2025 average daily DC burn was materially higher (≈$56k) and included discretionary burns that have since been suspended. The two are not measuring the same thing.
▸ Helium Mobile unit economics, MVNO wholesale rates, and the share of traffic actually offloaded onto Helium hotspots are not disclosed. They are commercially sensitive. We do not model them as facts.
▸ Noble Mobile deal terms were not disclosed. The strategic reading (Nova Labs refocusing on offload) is taken from the reporting, not from a filing.
// CACHE256 · ECOSYSTEM · Helium (HNT) · © 2026 Cache256 — Not financial advice · You are sovereign · Read what proves what. The rest is theater.
Helium built the coverage. Then it killed the free plan, unwound the tokens, suspended the subsidy and sold the carrier. The question it is now answering in public is the one every DePIN network will face: when the token stops paying for the coverage, does anyone pay for the coverage?