FOUNDATIONS · Last updated: 17 July 2026 · 12 min read · by Jordan, PassivePin Editorial

What is DePIN, really? A working definition

DePIN is the category of crypto that builds physical infrastructure: servers, cell towers, hard drives, sensors. Anyone with the right hardware can plug in and earn tokens. Here is how it actually works in 2026.

A working definition

DePIN stands for Decentralized Physical Infrastructure Networks. The acronym was coined by the analyst group Messari around 2022 to describe a category of crypto project that uses tokens to incentivize the build-out of real-world infrastructure by independent operators, rather than by a single company.

The pattern is consistent across the category. A project issues a token, distributes some of it to early operators, and uses the rest to fund ongoing payouts to operators who provide a measurable resource: bandwidth, storage, compute, location data, sensor readings, or wireless coverage. Customers (often enterprises) buy access to that infrastructure in stablecoins or project tokens. The protocol takes a cut, the rest goes to operators pro-rata.

The result is infrastructure that is cheaper to build, more distributed geographically, and censorship-resistant in ways traditional cloud providers cannot match. A bandwidth network with 100,000 residential nodes in 80 countries is a fundamentally different business from a cloud provider with 30 hyperscale data centers.

A short history: from BitTorrent to Helium

DePIN is not a 2024 invention. The pattern of using crypto incentives to coordinate physical resources dates back to the early 2010s, though the projects looked different then.

2003-2013: BitTorrent and the file-sharing era. BitTorrent proved that millions of individuals could coordinate to move enormous amounts of data without a central server. There was no token; coordination happened through gossip protocols. But the principle — peer-to-peer resource pooling — was the seed.

2014-2017: Storj, Sia, Filecoin, Maidsafe. The first wave of storage-focused projects. Storj and Sia launched in 2014-2015; Filecoin raised a then-record $257 million ICO in 2017. These projects aimed to use idle hard-drive space to compete with Amazon S3 and Google Cloud Storage. Most struggled with the cold-start problem (you need both supply and demand to launch, but neither side joins without the other), but they proved the model was technically viable.

2019: Helium. The first DePIN project to break into the mainstream. Helium paid people in HNT tokens for deploying small LoRaWAN wireless hotspots that connected Internet-of-Things devices. By 2022, Helium had over 1 million hotspots deployed across 180+ countries and a market cap in the billions. It also exposed the limitations: hardware requirements created geographic inequality, the subsidy model was unsustainable without enterprise demand, and the 2022 bear market cratered the token price.

2022: The Messari rebrand to "DePIN". The category name crystallized. Helium, Filecoin, Render, and Arweave were joined by dozens of new projects in storage, compute, bandwidth, and location. The 2022 bear market was brutal on token prices but turned out to be a build-out period: projects that survived the bear emerged with actual customers and real revenue.

2023-2025: AI compute + location data. The LLM boom created an insatiable demand for GPU compute. Render, Akash, and io.net all pivoted or expanded to capture it. Hivemapper and DIMO scaled location-data networks to thousands of dashcams. The category grew from ~$10B in 2022 to over $50B in combined market cap by mid-2025.

2026: The "real demand" year. This is the year DePIN stopped being a story about token emissions and started being a story about customer revenue. The mature projects — Helium, Filecoin, Render, Hivemapper — generate the majority of their operator payouts from enterprise B2B contracts, not from new token issuance. The flywheel is starting to turn on revenue, not subsidies. We are still early, but the speculative phase is ending.

The eight verticals

DePIN is not a single thing. Projects break into roughly eight verticals based on the resource they coordinate. The economics, hardware, and skill requirements differ across them.

Bandwidth — selling unused internet capacity. Grass, BlockMesh, EarnFM. Lowest barrier to entry; pays fractions of a cent per gigabyte routed. Covered in our bandwidth guide.
Storage — selling hard-drive space. Filecoin, Arweave, plus a long tail of smaller projects. Mid barrier to entry; needs reliable uptime and a meaningful stake. Covered in our storage guide.
Compute — selling GPU and CPU time. Render, Akash, io.net, Aethir, Aleph. Higher barrier to entry; needs a recent NVIDIA GPU and uptime. Covered in our GPU compute guide.
Location — selling GPS traces, dashcam footage, and street-level imagery. Hivemapper, DIMO, NATIX Network. Mid barrier; needs a dashcam or a running app on a phone.
Wireless — selling cellular and LoRaWAN coverage. Helium, World Mobile Token. High barrier; needs FCC- or equivalent-licensed hardware and varies by country.
Sensor — selling ambient data: air quality, soil moisture, weather. WeatherXM, dClimate. Low barrier but small payouts; mostly useful for projects that aggregate sensor data for B2B buyers.
Energy — peer-to-peer energy trading and grid optimization. Powerledger, Daylight. Niche but high-margin; regulatory exposure is the main risk.
Rewards — questing platforms that sit adjacent to DePIN. Layer3, Galxe, Zealy. Not infrastructure themselves, but the onramp for users into the rest.

The PassivePin directory tracks 36 active projects across these eight verticals. The mix has shifted: in 2023, bandwidth and storage dominated; in 2026, compute and AI-adjacent projects are the fastest-growing segment, both in number and in dollars paid to operators.

How a DePIN project actually pays you

There are three payment models in use today, and most projects blend them.

1. Token emissions. The project mints new tokens and distributes them to operators pro-rata. This is the bootstrapping model: it works when the project is launching, but it is inflationary. If demand for the token does not catch up, the token price falls, and the real earnings of operators fall with it. Helium ran on this model for years; the 2022 token price crash exposed the risk.

2. Customer revenue. The project charges enterprise customers (AI startups, hedge funds, mapping companies) in fiat or stablecoins, and distributes a share of that revenue to operators. This is the sustainable model; it is the model that mature projects like Render and Hivemapper have moved toward. The risk: if the customer base shrinks (a competitor emerges, a customer's funding dries up), operator payouts shrink with it.

3. Hybrid. Most projects use a mix: emissions to bootstrap, customer revenue to scale. The ratio shifts over time. A project that is still 80%+ emissions-funded in year three is in trouble; a project that is 80%+ revenue-funded is healthy. We track this ratio for every project in the directory.

For an operator, the practical question is: what is the dollar value of my expected payout over the next 12 months, after the token can be sold. That is the number on each project card. It accounts for both the current rate and the implied sell pressure; it is not a guarantee, but it is the closest honest estimate we can produce.

Why 2026 is different

Three things have changed since the 2022 bear market that make 2026 a structurally different year for DePIN.

AI compute demand. The cost of training a frontier model is now in the hundreds of millions of dollars. AI startups are desperate for cheap GPU time, and the hyperscalers (AWS, GCP, Azure) cannot keep up. DePIN compute networks have a real cost advantage: they pay residential operators in tokens, often below the cost of equivalent AWS instances, because the operator's marginal cost (electricity, depreciation) is lower. Render Network reported $50M+ in customer revenue in 2025, up 5x year-on-year. io.net grew 10x. Akash 4x. The compute vertical is now the largest by operator earnings, and it is not close.

Geographic distribution. In 2022, most DePIN operators were in North America and Western Europe. By 2026, more than 60% of new node deployments are in Latin America, Southeast Asia, and Africa. This is not a vanity metric; it matters because the customers (B2B buyers) want coverage in those regions, and a centralized cloud provider cannot offer it. Helium's coverage maps went from "USA + Western Europe" to "180+ countries" almost overnight, and that is the unlock that made the enterprise contracts work.

Regulatory clarity. The 2022-2024 period was a regulatory fog. The 2025 US election and the subsequent SEC guidance clarified that most DePIN tokens, when structured properly, are not securities. The EU MiCA framework explicitly carved out DePIN from the strictest crypto rules. Australia's ATO published guidance on how DePIN earnings should be taxed. This is not full clarity, but it is enough that institutional investors (funds, family offices) are now deploying into the space.

None of this means DePIN is now "safe." Tokens can still drop 70% in a quarter. Projects can still rug. Hardware can still fail. The category is in a stronger position than it has ever been, and the operators who joined in 2022-2024 are now generating meaningful income, but a new operator in 2026 should still treat any capital deployed as money they can afford to lose.

Next steps

If you have read this far, you have a working definition of DePIN, the historical context, the eight verticals, and a sense of why 2026 is different. The next steps depend on what you want to do.

If you are evaluating whether to start. Read the node economics guide to understand how projects are funded and what risks are priced in. Then read the scam red flags guide before you commit capital.
If you are ready to start. Pick one bandwidth project (lowest barrier) or one GPU project (if you have a gaming GPU). The setup guides section has step-by-step instructions. Use the hardware buyer's guide if you are considering buying equipment.
If you are already running nodes. Compare what you are running against the alternatives in the comparisons page. Most operators we talk to are running one bandwidth and one compute project; if you are doing more, you have probably identified an opportunity we have not.
If you want the headline numbers. The project directory has a card per project with the current earn rate, the risk score, and a setup link. Filter by category to see what fits your hardware and time budget.

Continue learning: Node economics, explained →