DePIN: The Smarter Way to Build Physical Infrastructure

A Nairobi driver, Austin homeowner, and Seoul designer earn passive income from everyday devices. That’s the promise of DePIN infrastructure.

Building a mobile network used to require tens of billions of dollars, thousands of employees, and a government license. As a result, a small number of telecom companies controlled pricing, coverage, and infrastructure deployment across entire countries. Users had no say and no alternative. If you wanted wireless coverage, you paid the carrier’s price. If your neighborhood wasn’t profitable to serve, you didn’t get served.

A decentralized wireless network works differently. Thousands of people run hotspots from their homes and offices and earn tokens for the coverage they provide. This model costs a fraction of the traditional approach. It also grows where people actually live and work instead of where corporations expect the highest returns.

From Corporate Ownership to Community Ownership

That is the idea behind Decentralized Physical Infrastructure Networks, or DePIN. Blockchain token incentives coordinate thousands or even millions of people to build and run physical infrastructure that has traditionally depended on centralized ownership and massive capital investment. The people who use the infrastructure can also own it, operate it, and earn from it.

The idea is not new. Cooperatives have existed for centuries. However, blockchain adds something cooperatives never had: a programmable, transparent, and tamper-resistant incentive layer that coordinates millions of participants across geographies without requiring them to trust one another or a central organization. The token is the contract, the protocol is the management, and the contributors are the company.

By mid-2026, DePIN tokens had a combined market capitalization of more than $15 billion, and the network included more than 2 million active nodes worldwide. It includes wireless connectivity, GPU compute, data storage, mapping, energy, and vehicle telemetry. Meanwhile, several leading projects generate tens of millions of dollars in annual revenue from enterprise customers, not token trading alone. This article tells the full story of where DePIN came from, what it is, who is building it, what the data shows, and which questions remain unanswered.

Before DePIN: The Infrastructure Problem Blockchain Was Built to Solve

The infrastructure problem DePIN addresses existed long before blockchain. It is the same problem that led to the internet’s decentralized design in the 1960s, the open-source software movement in the 1980s and 1990s, and peer-to-peer file-sharing networks in the early 2000s.

The problem is that building physical infrastructure at scale requires enormous amounts of capital. That concentration of capital leads to concentration of ownership. You cannot build a cellular network without spending billions. You cannot build a data center without spending hundreds of millions. Those costs create barriers to entry, limit competition, and produce the monopoly and duopoly structures found across telecommunications, cloud computing, mapping, and energy distribution.

Those concentrated structures are not bad on their own. They provide services people depend on every day. However, they also create problems that decentralized alternatives can address.

Coverage decisions are based on profitability instead of need. As a result, rural areas, developing markets, and low-density communities receive less investment because serving them does not generate enough return for centralized operators.

Pricing reflects market power instead of cost. Incumbents charge what the market will pay instead of what it costs to provide the service.

Data generated by the infrastructure belongs to the operator instead of the contributor. Your location data, driving patterns, and energy usage belong to the company whose infrastructure you use, not to you.

Every DePIN project tries to solve one or more of these problems for a specific type of infrastructure. The solution changes from one category to another. The approach stays the same. Token incentives coordinate distributed hardware providers to build infrastructure that one company would otherwise own and operate.

The Prehistory: Projects That Existed Before the Name (2017–2022)

The term “DePIN” first appeared in late 2022. However, the projects that came to define the category had been building for years before anyone gave them a shared name.

Filecoin launched its mainnet in October 2020, but its story began much earlier. Protocol Labs was founded in 2014, and Filecoin raised $257 million through an initial coin offering in 2017, one of the largest ICOs of the era.

Filecoin set out to change how data storage worked. Instead of paying Amazon, Google, or Microsoft to store data, users could pay independent storage providers that competed on price and reliability. The protocol used two cryptographic proofs: Proof of Replication (PoRep) and Proof of Spacetime (PoSt). Together, they verified that providers were storing the data they claimed to store. As a result, users no longer had to trust any single provider.

As documented in our RWA tokenization research, using cryptographic verification to replace institutional trust is a recurring theme across blockchain’s most durable applications.

Wireless Networks Join In

Helium launched its blockchain in 2019 after several years of hardware development. It built a decentralized Internet of Things (IoT) network where people ran hotspots from their homes and offices. Those hotspots acted as small wireless transmitters, and operators earned HNT tokens for providing LoRaWAN coverage to nearby IoT devices.

By the time the blockchain launched, Helium had already built a meaningful wireless network. It did that without spending billions on infrastructure. Instead, thousands of hotspot operators each invested a few hundred dollars in hardware.

Render Network launched in 2017 to help GPU owners earn money from idle computing power by renting it to 3D artists and studios.

Its first customers were film studios, game developers, and visual effects companies that needed large amounts of GPU capacity for short periods without buying dedicated hardware. Render’s distributed network provided that capacity at a fraction of the cost of traditional data centers.

Then AI training created an even larger market for GPU compute. Five years after launch, Render was serving demand far beyond its original use case.

One Name for Many Projects

The name DePIN came later. Between late 2022 and early 2023, investment firms such as Multicoin Capital helped popularize Decentralized Physical Infrastructure Networks as the name for this category. Since then, it has become the umbrella term for projects that combine physical infrastructure with blockchain and token incentives.

More specifically, analytics firm Messari introduced the name DePIN in late 2022 through a public poll that brought together several competing labels, including MachineFi, Token Incentivized Physical Infrastructure Networks (TIPIN), EdgeFi, and Proof of Physical Work.

Before that, the same model was known by several names. That made the category harder to understand and follow than it needed to be.

How DePIN Works: The Flywheel Model

Every DePIN project is built around a flywheel. It is a cycle that keeps itself going as it grows. Understanding that cycle explains both DePIN’s potential and its biggest risks.

The cycle begins with token incentives. A new DePIN project launches a protocol and distributes tokens to people who deploy hardware such as hotspots, storage drives, GPU rigs, dashcams, and sensors. In return, they provide the service that hardware makes possible.

At first, the tokens derive most of their value from expectations about the project’s future. That gives early contributors a way to earn rewards before the network has paying customers.

From Contributors to Customers

Token incentives encourage more people to deploy hardware. As more hardware comes online, the network gains coverage, capacity, and better service. Better service attracts users and enterprise customers who pay in fiat currency or stablecoins.

That demand generates protocol revenue. The protocol can use that revenue to buy and burn tokens or distribute additional rewards to contributors. At that point, the token is supported by both future demand and network revenue. More contributors join, more hardware comes online, and the cycle continues.

Traditional infrastructure projects require significant upfront capital expenditure (CapEx) before they can offer a service. DePIN works differently. Token incentives encourage individuals and organizations to deploy infrastructure one device at a time.

That gives DePIN an advantage over traditional infrastructure models. Instead of one company carrying all the capital risk, thousands of contributors each carry a small share. The protocol funds the incentives; contributors pay for the hardware.

Demand Has to Catch Up

The flywheel depends on one thing above all else. Demand must eventually grow enough to support the infrastructure that token incentives create.

Hivemapper shows what happens when that does not occur. Supply grew faster than customer demand. According to Messari’s Q1 2025 DePIN report, the strongest projects generate revenue before they scale infrastructure. Hivemapper’s monthly revenue fell from $195,000 to $6,000-$12,000 between December 2024 and March 2025. It showed what happens when token incentives create infrastructure faster than enterprise customers are willing to pay for it.

The Two Types of DePIN Networks

DePIN protocols fall into two groups with very different economics: Physical Resource Networks (PRNs) and Digital Resource Networks (DRNs).

Physical Resource Networks (PRNs) include Helium for wireless coverage, Hivemapper for road mapping, and DIMO for vehicle telemetry. These networks depend on location. Their services only have value in the places where they are available. A Helium hotspot in Manhattan provides coverage for IoT devices in Manhattan. The same hotspot in an uninhabited area has little or no value.

Digital Resource Networks (DRNs)

Digital Resource Networks (DRNs) include Filecoin for storage, Render and Akash for GPU compute, and Theta for bandwidth. These networks provide services that are not tied to a specific location. A GPU in Seoul can train an AI model for a company in London just as effectively as a GPU in London. Likewise, storage on a node in Lagos can serve a user in New York as effectively as storage on a node in New York.

The two groups also face different challenges. PRNs have a harder time getting started because they need enough contributors in the same locations before the network becomes useful. DRNs do not have that problem. Instead, they compete in markets where compute, storage, and bandwidth are more easily substituted. That makes price and reliability their biggest advantages.

The Projects That Defined the Category

Wireless

Helium: The Project That Proved the Model

HNT · Market cap: ~$318M · 900,000+ IoT hotspots · 120,000+ mobile subscribers

Helium is the project that put DePIN on the map. Its history includes both one of the strongest examples of what the model can achieve and one of its clearest warnings.

At its peak in 2021 and 2022, Helium had deployed more than 900,000 hotspots across over 160 countries. No traditional wireless operator had built comparable coverage at a similar cost. Hotspot operators earned HNT rewards for providing LoRaWAN IoT coverage. The network was global, decentralized, and useful for applications that needed low-power, wide-area connectivity.

When Growth Outpaced Demand

The challenge began when Helium’s token economy pulled ahead of network usage. The network had built an enormous supply of hotspots, but demand for IoT services did not grow at the same pace. Contributors continued earning tokens, but token demand depended more on new investors than on revenue from paying IoT customers.

When the broader crypto market fell in 2022, HNT lost much of its value. As a result, rewards dropped, and many hotspot operators no longer had enough incentive to keep their devices online.

Helium responded by moving from its own blockchain to Solana in 2023. The move reduced infrastructure costs and gave the project access to Solana’s developer ecosystem. It also launched Helium Mobile, a consumer cellular service that uses Helium hotspots to offload traffic from traditional mobile carriers.

By 2025, Helium Mobile had passed 120,000 subscribers through partnerships with AT&T and Telefónica. More importantly, those subscribers generated paying customers and recurring revenue instead of relying only on token rewards.

The August 2025 halving reduced annual HNT emissions from 15 million to 7.5 million tokens. In October 2025, Helium recorded its first deflationary month as token burns from Helium Mobile subscription revenue exceeded new emissions. Earlier that year, in April, the SEC dismissed its lawsuit against Helium, giving the project greater legal clarity.

Those two milestones separated a network supported by paying customers from one supported mainly by token demand.

Storage

Filecoin: The Largest Decentralized Storage Network

FIL · Market cap: ~$2.1B · 2.1 exbibytes secured · 9% utilization

Filecoin is the oldest and largest DePIN project by most measures. Its history shows both how far the model can scale and the gap between deployed capacity and actual usage that continues to challenge many DePIN networks.

Filecoin has secured more than 2.1 exbibytes (EiB) of storage capacity, putting it on a scale comparable to major cloud providers. When Protocol Labs published the Filecoin whitepaper in 2014, infrastructure of this size seemed far out of reach.

The network uses cryptographic proofs to verify that storage providers are storing the data they claim to store. That removes the trust requirement that limited earlier decentralized storage networks.

Capacity Is Not the Same as Demand

FIL trades at around $3.80 per token, giving the network a market capitalization of roughly $2.1 billion. It has more than 2 exbibytes of raw storage capacity, yet only about 9% of that capacity is in use.

That 9% utilization rate is the number many discussions about Filecoin overlook. The network succeeded in attracting storage providers and building capacity. However, demand has not grown at the same pace. Token incentives solved the supply side. Building enough demand to fill that storage has taken longer.

The Filecoin Virtual Machine (FVM), launched in 2023, added smart contract support to Filecoin’s storage layer. It made programmable storage deals, data DAOs, and decentralized finance (DeFi) applications built on storage possible.

The FVM is Filecoin’s biggest effort to expand beyond storage into a broader application platform. Growing adoption among AI data providers and archival institutions suggests the utilization gap could narrow as more applications are built on the network.

GPU Compute

Render Network: GPU Compute Meets the AI Boom

Token: RENDER · Market cap: ~$3.2B · Largest DePIN project by market cap · AI training and rendering

Render Network has the clearest demand driver of any major DePIN project. That demand did not come from the project’s design alone. It came from timing. The AI boom that began in 2023 and accelerated through 2024 and 2025 created a global shortage of GPU compute. As a result, Render’s distributed GPU network became valuable to enterprises that could not secure enough capacity from traditional cloud providers.

The biggest story in DePIN today is not a token launch or a protocol upgrade. It is the global shortage of AI compute.

SK Hynix and Micron have both confirmed that their entire 2026 supply of High Bandwidth Memory (HBM) is already sold out. HBM is a critical component in AI training hardware. As supplies tighten, more companies are looking beyond traditional cloud providers and turning to decentralized networks to expand their compute capacity.

As of September 2025, Render Network had a market capitalization of about $2.03 billion, with 518 million RENDER tokens in circulation. The network has enough compute capacity to run between 300 and 1,000 AI models at the same time. It has also expanded beyond crypto through work on VR and film projects, including Batman: The Animated Series and Westworld.

Akash Shows What High Utilization Looks Like

Render’s biggest competitor is Akash Network. While Render benefits from AI demand, Akash stands out for a different reason. It has achieved some of the strongest utilization rates in the DePIN sector.

Akash reported 428% year-over-year growth in usage, with utilization above 80% heading into 2026. Many distributed compute networks struggle to match supply with demand. Akash has largely avoided that problem.

An 80% utilization rate means most of the available compute on the network is being used by paying customers. That supports continued infrastructure growth and gives the token value based on network activity, not just market sentiment.

Mapping

Hivemapper: Building a Fresher Google Maps With Dashcams

HONEY · 37% global road coverage · Lyft and Volkswagen among enterprise customers

Hivemapper shows both the promise of DePIN and its biggest risks. In its case, both played out within a single year.

The idea is easy to understand. Drivers install AI-powered 4K dashcams and earn HONEY tokens for capturing street-level imagery while driving their normal routes. That data is used to build a map that stays more up to date than Google Street View, where images can be years old.

Fresh mapping data has real value for autonomous vehicle companies, logistics platforms, and businesses that need current information about road conditions, traffic signs, and infrastructure.

Hivemapper has mapped 37% of the world’s roads, and companies such as Lyft and Volkswagen use its data for autonomous vehicle navigation and route optimization. Four autonomous vehicle companies also pay for Hivemapper’s mapping data, generating $4 million in revenue during the first quarter alone.

The network also created new income opportunities for contributors. In 2025, one driver in Nairobi earned $18,000 by driving his regular delivery route with a Hivemapper dashcam.

Those numbers capture the DePIN model at its best. Enterprise customers pay for the data, while contributors earn income from hardware they were already using.

When Demand Fell Behind

Hivemapper also provides one of the clearest examples of DePIN’s cold-start problem.

Despite mapping 37% of the world’s roads and signing enterprise customers such as Lyft, monthly revenue fell by 94%, from $195,000 in December 2024 to between $6,000 and $12,000 by March 2025. The network had built the supply. Demand was not growing fast enough to support the token rewards that attracted contributors.

Revenue later recovered. Annualized revenue increased from about $500,000 in August 2025 to roughly $18 million by early 2026, a 36-fold increase driven by enterprise customers that needed fresher street-level mapping data than Google could provide.

The recovery supports the idea behind Hivemapper. At the same time, the collapse showed how vulnerable the flywheel can be when infrastructure grows faster than customer demand.

Vehicle Data

DIMO: Giving Vehicle Owners Control of Their Own Data

Token: DIMO · 220,000+ connected vehicles · Revenue from insurers, OEMs, and fleet operators

DIMO addresses a problem most vehicle owners do not realize exists. Every time you drive a modern car, it generates a constant stream of telemetry data, including location, speed, fuel consumption, maintenance diagnostics, and driving patterns. Manufacturers collect that data and use it to improve products, support insurance pricing, and train autonomous driving systems. The people generating the data receive none of the value.

DIMO connects vehicles to its network through a hardware device or software integration with compatible vehicles. It gives owners an on-chain record of their vehicle’s data and lets them choose who can access it. That includes insurers looking for verified driving data, OEMs that need anonymized fleet data for engineering, and fleet operators that rely on real-time vehicle health and location data.

DIMO now has more than 220,000 connected vehicles. Earlier in its development, the network had about 80,000 connected vehicles. That was enough to provide useful insights for insurers and fleet operators because the network had reached the scale needed for meaningful analysis. It has continued to grow since then.

Insurers, OEMs, and fleet operators pay for the data. That shows DIMO is building a business around paying customers instead of depending only on token demand

The AI Connection: Why DePIN Became Relevant in 2024–2025

DePIN became one of crypto’s most talked-about sectors in 2024 and 2025, but that was not only because of its own progress. The AI boom created demand for GPU compute, storage, and data at a time when traditional cloud providers struggled to keep up.

GPU compute is one part of that story. Training AI models requires large amounts of processing power, and shortages at centralized cloud providers made decentralized GPU networks such as Render and Akash more attractive.

However, AI needs more than compute. It also needs large amounts of high-quality data.

AI Needs Data Too

Hivemapper’s street-level imagery helps train autonomous vehicles. DIMO’s vehicle telemetry supports AI systems used by insurers and fleet operators. Grass Network pays people for unused internet bandwidth so AI companies can collect publicly available web data at scale without running into IP blocks.

The AI boom changed how many people viewed DePIN. Hard drives, GPUs, Wi-Fi hotspots, dashcams, and sensors were no longer just hardware connected to token rewards. They became infrastructure that AI companies were willing to pay for.

The connection between AI and DePIN is also visible in the growth of leading projects. By mid-2025, Bittensor (TAO), Render (RENDER), and Filecoin (FIL) each had market capitalizations above $1 billion. Grass Network also grew from 200,000 users to 3 million in a single year.

Bittensor takes that connection a step further. Instead of providing compute or data for AI, it uses blockchain to coordinate and reward AI model development. It shows how DePIN and AI can extend beyond infrastructure into decentralized AI networks.

As covered in our RWA tokenization research, blockchain projects have been most successful when they solve problems that institutions are willing to pay for. DePIN’s connection to AI follows the same pattern. Enterprises with AI budgets need compute, storage, and data, and they are paying DePIN networks for those services instead of relying only on token demand.

What Makes a DePIN Project Sustainable and What Kills It

The DePIN sector includes some of crypto’s strongest utility projects. It also includes projects that used token incentives to build supply but never attracted enough customers to support it. Telling the difference is one of the most important questions for anyone evaluating the sector.

Many DePIN projects in 2026 are still relying on token incentives while generating little revenue. The projects that have built sustainable businesses share a few common characteristics.

1. Enterprise Customers Pay in Fiat

The first is enterprise customers paying in fiat currency. When Render Network completes a paid GPU compute job, when Hivemapper receives payment from an autonomous vehicle company, or when Helium Mobile collects a subscription fee, those payments come from customers buying a service, not from people buying a token.

Another difference is transparency. Transactions such as Render’s GPU compute jobs or Filecoin’s storage deals are recorded on-chain, making the revenue publicly verifiable.

2. High Utilization

The second is utilization rates that make economic sense.

A storage network with 9% utilization is paying contributors to maintain capacity that sits idle. A compute network with 80% utilization is paying contributors to provide resources that customers are actively using. That difference has a direct impact on whether the network can support itself over time.

3. Network Usage Supports the Token

The third is a burn mechanism that links network usage to the token.

When Helium burns HNT based on subscription revenue, network growth reduces the circulating supply. When Render uses protocol revenue to buy and burn RENDER, demand for GPU compute affects the token’s supply. These mechanisms connect network activity to token value.

4. Customers Came First

The fourth, and the most important, is demand.

The strongest DePIN projects solved a problem that customers were already willing to pay for. They built infrastructure to meet that demand. The weakest projects built infrastructure with token incentives first and then searched for customers afterward.

As documented in our research on why crypto projects fail, product-market fit remains the strongest indicator of long-term survival. DePIN is no different.

$3.5T

World Economic Forum estimate of DePIN’s market potential by 2028

The World Economic Forum’s $3.5 trillion estimate comes with an important condition. It applies only to projects that build sustainable revenue models. Telecommunications, cloud compute, data storage, mapping, and energy represent massive markets. How much of that opportunity DePIN captures depends on whether projects can build lasting customer demand before token incentives lose their effect.

The Honest Counterarguments

The case for DePIN deserves a closer look. So do the strongest arguments against it.

The Cold-Start Problem

Every DePIN network faces the same challenge. The service becomes more valuable as the network grows, but building that network requires contributors before the service has enough customers.

Token incentives help during the early stages. They reward contributors before enterprise customers arrive. However, those incentives do not last forever. If customer demand does not grow before rewards lose their appeal, the network faces a difficult transition.

Hivemapper’s revenue decline in 2025 is a good example. The network had mapped 37% of the world’s roads, an impressive supply milestone. However, demand for that mapping data was not growing fast enough to support contributor rewards. Revenue later recovered as enterprise adoption increased, but the decline showed how fragile the early stage of the flywheel can be.

Token Incentives Can Create the Wrong Behavior

Token rewards attract contributors who want exposure to the token as well as contributors who want to support the network. When token prices rise, more hardware comes online, even if it adds little value. When prices fall, contributors leave, even if their hardware is still useful.

That creates swings in supply that make DePIN networks harder to sell to enterprise customers, who expect reliable and consistent service.

Physical Resource Networks face another challenge. Contributors naturally deploy hardware where installation is easiest, not where coverage is needed most. A hotspot in Manhattan can earn the same rewards as one serving a rural community in West Africa, even though the rural deployment may contribute more to the network. Designing incentives that reward useful coverage instead of convenient coverage is still one of DePIN’s biggest challenges.

The Regulatory Question

DePIN projects that provide communications services, including Helium Mobile and XNET, operate in heavily regulated industries. Spectrum allocation, carrier agreements, and consumer telecommunications rules apply whether the network uses blockchain or not.

Helium’s legal battle with the SEC, which ended when the case was dismissed in April 2025, showed the regulatory uncertainty DePIN projects face when token models resemble securities offerings.

The regulatory environment in 2026 is different from what it was in 2022. The GENIUS Act, passed in July 2025, established a framework for token-based financial instruments. Several DePIN projects have also completed regulatory reviews and now operate with more certainty about their legal status.

DePIN also operates across telecommunications, energy, finance, and data privacy. Meeting the rules that govern those industries will continue to be part of building and operating these networks.

What I’m Watching

The DePIN sector in mid-2026 has reached a stage where the difference between the strongest projects and the rest is becoming easier to measure. A handful of projects have moved beyond token incentives and built businesses with enterprise customers. Most have not. They are still relying on token rewards while trying to prove that customer demand will follow. Over the next 12 to 24 months, that gap will become easier to see as token incentive programs mature and projects either build sustainable businesses or run out of time.

The metric I watch most closely is the ratio of protocol revenue to token emissions. In other words, how much revenue a network earns from paying customers compared with how much value it distributes through token rewards.

If protocol revenue grows faster than token emissions, the network’s economics are improving. If token emissions continue to exceed protocol revenue, the project still depends more on token demand than customer demand.

What Will Decide the Next Phase?

DePIN in mid-2026 already has working infrastructure, enterprise revenue, and growing demand from AI companies. At the same time, most of the sector’s 650-plus projects have not shown they can build lasting demand for their services.

Projects such as Aethir, Helium, Akash, Render, Hivemapper, and GEODNET stand out because they have customers paying for the services they provide. Many other projects are still betting that customer demand will catch up with the infrastructure they have built.

The second thing I am watching is the AI compute shortage. SK Hynix and Micron have confirmed that their entire 2026 output of High Bandwidth Memory (HBM) is already sold out. As long as demand for AI training exceeds the capacity available from traditional cloud providers, networks such as Render and Akash have customers looking for more compute.

If that shortage eases through new capacity from NVIDIA, AMD, or the major cloud providers, demand for decentralized GPU networks could also slow.

Looking Beyond AI

The third area I am watching is how DePIN develops alongside tokenized real-world assets (RWAs). As our RWA tokenization research shows, blockchain delivers the most value when it solves problems that existing systems cannot solve as efficiently.

DePIN applies the same idea to physical infrastructure. Public blockchains create verifiable records of who contributed infrastructure, how it was used, and how contributors are rewarded. That reduces the need to rely on a central operator and gives every participant access to the same data.

A delivery driver in Nairobi earning $18,000 by driving his regular route. A homeowner in Austin earning $42 a month from a hotspot in the living room. A GPU owner in Seoul renting idle compute to an AI company she has never met. These are not hypotheticals. They are documented, on-chain records of a model that is already working in different parts of the world.

Whether that model grows much further depends on one thing: Customer demand has to grow before token incentives lose their effect. That is the question every DePIN project has to answer.

Key Takeaways

DePIN uses blockchain token incentives to coordinate distributed hardware providers, including hotspot operators, GPU owners, storage providers, and dashcam-equipped drivers. Together, they build physical infrastructure that has traditionally required centralized ownership and large amounts of capital. The token provides the incentive. The protocol coordinates the network. The contributors own and operate it.

The term “DePIN” first appeared in late 2022 when Messari introduced it as a shared name for a category that had been developing since at least 2017. Filecoin, Helium, and Render Network were all building long before the category had a name.

DePIN projects fall into two groups. Physical Resource Networks (PRNs), such as Helium and Hivemapper, provide location-dependent services. Digital Resource Networks (DRNs), such as Filecoin and Render, provide digital services that are not tied to a specific location. PRNs face a harder cold-start problem because they need coverage in the right places. DRNs can scale more easily, but they compete more on price and reliability.

At a Glance

The AI compute shortage has become one of the biggest drivers of DePIN growth. Render Network and Akash Network expanded through 2024 and 2025 as enterprises looked for GPU capacity that traditional cloud providers could not supply. Akash’s utilization rate of more than 80% and its 428% year-over-year growth in usage show strong customer demand.

By mid-2026, projects such as Helium, Akash, Render, Hivemapper, Filecoin, and Aethir were generating tens of millions of dollars in annual revenue from enterprise customers. Most of the more than 650 DePIN projects still rely on token incentives while working to build lasting customer demand.

One metric stands above the rest when evaluating DePIN projects: protocol revenue compared with token emissions. When revenue grows faster than token emissions, the business is moving in the right direction. When token emissions continue to outpace revenue, the project still depends more on token demand than customer demand.