DePIN Crypto Projects List 2026: A Category-by-Category Map of the Major Networks

Bottom line: the major DePIN networks sort into two families, and knowing which is which tells you what to check

Every DePIN worth mapping falls into one of two families. Physical Resource Networks (PRNs) supply things tied to a location that cannot be virtualised — wireless coverage, sensor readings, energy, maps. Digital Resource Networks (DRNs) supply fungible resources deliverable from anywhere — bandwidth, storage, compute. That split is the standard taxonomy (as of 5 June 2026), and it is the spine of the directory below.

What follows is a map, not a ranking and not a tip sheet. It lists only networks whose physical resource, chain, ticker and status could be confirmed against primary documentation on 16 July 2026. Where a widely-quoted number could not be traced to a primary source, it has been left out — and there are more of those than you would expect.

The name was picked by a poll, not a standards body

One piece of context first. "DePIN" was not defined by a standards organisation. It won a Messari poll on X (Twitter) in November 2022 with 31.6% of the vote, beating rival coinages including IoTeX's "MachineFi" (2022) and Lattice's "TIPIN" (July 2022). A category named by vote has no authoritative membership test — which is exactly why a directory needs to state its own.

The three tests this list uses

There is no canonical "three conditions of DePIN," so these are our own inclusion criteria. The academic taxonomy describes DePIN as a five-layer architecture — application, governance, data, blockchain, and infrastructure — and it is the infrastructure layer, the actual physical communication, compute and storage resources, that separates a DePIN from an ordinary crypto protocol.

Test 3 is where the field separates. The cleanest documented implementation is Filecoin's: Proof of Replication (PoRep) proves a provider made a unique copy of the client's data, and Proof of Spacetime (PoSt) proves continuous storage over time — as Winning PoSt (possession at a point in time) and Window PoSt (consistent storage across a defined period), as of 16 July 2026.

DePIN is often confused with RWA. The distinction worth holding: RWA centres on tokenising assets that already exist, while DePIN centres on getting distributed, independent participants to build and run the physical infrastructure itself.

Directory 1: Physical Resource Networks (PRNs)

Location-bound resources that cannot be virtualised. All confirmed as of 16 July 2026.

A few notes. Helium runs a global LoRaWAN network for IoT devices plus a cellular-offload/Wi-Fi "Mobile" network, both powered by community-operated hotspots. GEODNET bills itself as "the decentralized location layer for robotics," supplying centimetre-level positioning to drones, robots and autonomous vehicles — and it is a useful reminder that "which chain?" doesn't always have one answer: GEOD exists as an ERC-20 on Polygon, as an SPL token on Solana (approved via GIP3, using the Wormhole NTT framework), and on IoTeX. WeatherXM is deployed on Arbitrum One — an Ethereum layer 2, not Solana, despite what a surprising number of listicles imply.

DIMO deserves a flag. Its own governance proposal (DIP-10) lists token contracts on Ethereum, Base/Optimism and Polygon. It launched on Polygon and publicly announced a protocol migration to Base — but no primary source confirms the migration is complete. Calling DIMO "a Polygon project" full stop is wrong; so is calling it a Base project. The honest formulation is the multi-chain one.

Directory 2: Digital Resource Networks (DRNs)

Fungible resources that can be delivered from anywhere.

Filecoin runs its own chain — a chain of tipsets rather than individual blocks — secured by Expected Consensus, a probabilistic Byzantine-fault-tolerant algorithm in which the probability of being elected depends on how much provable storage a miner contributes. Arweave sells "pay once, store forever": most of a transaction fee is contributed to a storage endowment distributed to miners over time, rather than going wholesale to the miner of that block.

The compute networks — Render, Akash, io.net, Nosana — all sit close to AI demand (where AI and crypto meet). Akash's distinguishing mechanism is what Akash itself calls a "reverse auction": independent providers worldwide compete by submitting bids, and the user chooses on price, location and reputation, so providers bid the price down. The network upgraded to Mainnet 14 on 28 October 2025 at 13:23 UTC (block #23939793), jumping from Cosmos SDK v0.45 to v0.53. Render is no longer just 3D rendering — it now runs "Dispersed," a GPU network for AI workloads (Render Network Foundation monthly report, as of 31 March 2026).

Grass needs careful wording. Third-party articles routinely describe it as infrastructure for "ethical web scraping for AI training." Grass's own FAQ says something far tamer: "verified institutions" performing "tasks like checking local prices, viewing regional ads, or conducting academic research." It never says AI scraping. This article uses Grass's own wording, and notes the gap rather than resolving it. On the obvious privacy question, Grass states: "Your personal information is never accessible; we never access or use your browsing history, and we do not access or sell your personal information." That is the project's claim, not a verified fact.

Directory 3: DePIN-purpose chains (not resource networks)

Worth separating, because they get listed alongside the networks above as though they were the same kind of thing. They aren't — they're the substrate.

peaq uses PEAQ for gas and as the deposit locked when registering a machine, and issues machines a globally unique identity written did:peaq:0x… that isn't issued by any single company. Total supply is roughly 4.2 billion PEAQ (inferred from its CoinList allocation of 252,000,000 PEAQ representing 6% of supply, as of 9 May 2024). Note a positioning wrinkle: peaq's own homepage now leads with "peaqOS" and omnichain framing rather than its Polkadot parachain status — but the parachain tracker shows it ACTIVE as of 16 July 2026. IoTeX offers W3bstream, a real-time data attestation and compute layer connecting smart devices to smart contracts, and ioID, giving machines verifiable identities (as of 16 July 2026).

How to evaluate one honestly: real demand, or just emissions?

This is the point of the map. DePINs are good at recruiting the supply side — give out tokens and hardware appears. The hard part is whether outside customers actually pay. Four checks, in order.

1. Is there a pay-or-burn loop for actual usage? Helium is the clearest example: hotspot operators earn HNT for maintaining coverage, while enterprises and developers burn HNT to create Data Credits, which pay the fees for wireless data transmission. Usage consumes supply. WeatherXM's customers must acquire WXM to use the network's data or services. On io.net, all payments are ultimately converted to $IO behind the scenes.

2. Is the focus moving from supply to demand? One network has said publicly that it is. The official 2026 Filecoin Network Strategy, published 19 February 2026, states: "This year marks a shift in ecosystem focus, away from growing supply and towards scaling demand." Its objectives are driving paid onchain deals, strengthening network profitability and cryptoeconomics, and scaling paid flagship client adoption. Note the framing: the strategy presents this as the community "shifting focus to demand at just the right moment" — not as a concession that supply ran ahead, which it never says. Still, when a network that calls itself "the world's largest decentralized storage network" with "exbibytes of capacity" (its own words, and stated only qualitatively) redirects its roadmap toward paying customers, that is worth noting.

*3. Is that scale number registered or active? This is the single most common way DePIN scale is overstated. Querying Helium's own entity API on 16 July 2026 returns 1,035,520 IoT hotspot entities and 56,701 Mobile hotspot entities. But those are registered entities, and the same API carries an is_active flag showing many are not active. A registered hotspot count is not a measure of live coverage.* This article publishes no active-rate figure, because none could be established reliably — and declining to print a number you can't stand behind is part of the map's accuracy, not a gap in it.

4. Can anyone name an outside customer? Rare in DePIN, but it exists. Telefónica and Nova Labs announced Helium Mobile Hotspots in Mexico on 24 January 2024, letting Movistar SIM customers offload mobile data onto the Helium Mobile Network while Telefónica retained full control of the customer experience — a pilot with select customers in Mexico City and Oaxaca.

XYO co-founder Markus Levin, quoted in the 30 January 2026 report on Messari's data, put the test bluntly: "When token prices are flat, the only thing that matters is whether someone is actually paying for the service, and whether the network can sustain itself without subsidies. That shift is healthy." Messari analyst Dylan Bane adds that DePINs should not abandon supply-side growth but must prioritise finding product-market fit on the demand side. This is really a question about centralisation versus decentralisation: decentralising is not the goal in itself — there has to be a reason the decentralised version is cheaper, faster or more durable.

Risks

Regulatory: the Helium case did NOT settle whether DePIN tokens are securities. This is widely misreported, so here is the record. On 23 April 2025 the SEC obtained a final judgment by consent against Nova Labs, creator of the Helium Network. It was limited to allegations that Nova Labs violated Section 17(a)(2) of the Securities Act of 1933 by misrepresenting, when offering and selling preferred stock in a private placement, that Lime, Nestlé and Salesforce were using the Helium Network when they were not. Nova Labs paid a $200,000 civil penalty without admitting or denying the allegations.

The SEC dismissed the other claims with prejudice — but stated in its own litigation release that the decision "rests on its judgment that the dismissal will facilitate the Commission's ongoing efforts to reform and renew its regulatory approach to the crypto industry, not any assessment of the merits of the claims alleged in the action," and "does not reflect the Commission's position on any other case." The securities question remains open. Any article claiming this settled it for DePIN tokens contradicts the SEC's own document.

Rewards are not guaranteed. Helium's Carrier Offload Program FAQ states plainly that "MOBILE rewards are subject to change" and that Nova Labs "does not issue, guarantee, refund, and/or warrant HNT rewards." The same FAQ notes some carriers are unlisted for confidentiality. Hardware is a fixed cost; rewards are a variable expectation.

Businesses get sold. On 2 June 2026, Andrew Yang's Noble Mobile acquired Helium Mobile — the direct-to-consumer cellular business, roughly 600,000 cumulative sign-ups including unsubscribed users — from Nova Labs, which retained the data-offloading program aimed at large carriers. Crucially, the Helium Network and the HNT token did not change hands: tokens remain under the purview of the Helium DAO, and Noble Mobile becomes a user of the network, not its owner. Describing Helium Mobile as a Nova Labs product has been wrong since June 2026.

Chain migrations leave stragglers. Render's community voted via RNP-002 to move to Solana, converting RNDR to RENDER 1:1 from both the Ethereum and Polygon versions. Render's own knowledge base says the upgrade assistant "will stay open indefinitely" — but legacy RNDR on Ethereum and Polygon is no longer maintained by the Foundation and cannot be used on the Render Network.

Assumptions can break. Arweave's permanence rests on storage costs continuing to fall. Its yellow paper's own figure: "Over the past 50 years, the average annual rate of decline of GB-hour cost has been 30.57%." Secondary articles frequently cite 38% — the primary source says 30.57%. If that decline stalls, the endowment maths weakens. That is the honest risk in the model.

Read self-reported numbers as self-reported. peaq advertises 6M+ robots, machines and humans onboarded, 48M+ machine transactions and 60+ apps across 22 industries; Grass says it is "trusted by over 8.5M users worldwide" (both taken from the projects' own homepages, as of 16 July 2026). These are the projects' own marketing figures, not independently verified.

Finally, you can check some of this yourself. Helium's own docs publish the HNT Solana mint address, hntyVP6YFm1Hg25TN9WGLqM12b8TQmcknKrdu1oxWux. Querying Solana mainnet on 16 July 2026 returns an initialized, live SPL mint with 8 decimals and a supply of roughly 181,841,487 HNT — consistent with, and below, the ~223,000,000 documented cap. Reading a block explorer walks through how to do that. Pulling one mint yourself is often faster than trusting a list — including this one.