DePIN (Decentralized Physical Infrastructure Networks)
Executive summary
DePIN, the practice of bootstrapping physical infrastructure networks through token incentives, evolved from a curiosity in the 2017-2020 era (early Helium, Filecoin) into a credible category by 2026. Helium pivoted from LoRaWAN-only into a U.S. mobile carrier on T-Mobile's network, Render and io.net consolidated the decentralized GPU compute story into a real AI-inference revenue base, Hivemapper crossed enough hardware-collected road imagery to seriously compete with mapping incumbents, and DIMO scaled vehicle-data subscriptions to several hundred thousand connected cars. Aggregate DePIN token market cap settled in the $15-25B range, with real revenue running at multi-hundred-million-dollar annualized run-rates across the category. The fundamental DePIN dynamic - that token incentives can frontload capex that would never close on a traditional venture-backed P&L - has been validated, but the second-order question of whether these networks can sustain unit economics once token subsidies fade is the binding constraint through 2027.
Origin and the DePIN flywheel mechanism
DePIN's core thesis is that hardware deployment economics, particularly in long-tail or geographically distributed infrastructure, are dramatically better when capex is funded by token issuance to operators rather than corporate balance-sheet debt. The flywheel: a protocol issues tokens to participants who deploy hardware (a hotspot, a GPU, a dashcam, an air-quality sensor); deployed hardware produces a service (wireless coverage, GPU compute, mapping imagery, sensor data); customers pay in fiat or stablecoins for that service; revenue is captured by the protocol and either burned, redistributed to operators, or used for buybacks. The model works when the unit economics of hardware-plus-token-incentive are favorable enough to bootstrap, and when end-customer demand exists at price points the network can support. Critics argue that DePIN networks frequently fail the second test: hardware operators are rewarded for deployment regardless of utilization, leading to oversupply and tokens that trade as inflation rather than yield. By 2026, the surviving DePIN projects are differentiated from failed peers primarily on the utilization axis - networks where deployed hardware is actually being used, not just present.
Helium: from LoRaWAN to mobile carrier
Helium is the foundational DePIN case study. Originally launched in 2019 as a LoRaWAN network for low-bandwidth IoT, Helium grew to nearly one million hotspots at peak before facing a reckoning when actual data revenue lagged dramatically behind hotspot deployment incentives. The network restructured aggressively: migrated from a custom L1 to Solana in 2023, restructured tokenomics around HNT and the IOT/MOBILE subDAOs, and pivoted toward 5G mobile through Helium Mobile. The Helium Mobile carrier launch in 2023, riding on T-Mobile's network with a hybrid offload model where Helium 5G hotspots provide local capacity to T-Mobile in exchange for revenue share, was the strategic inflection. By 2026, Helium Mobile reached low-hundred-thousands of paid subscribers at a $20-30/month price point, generating real ARR in the tens of millions of dollars. The hotspot operator base shrank meaningfully from the 2022 peak as unprofitable LoRaWAN deployments rationalized, but the mobile-side hotspot density grew in target metros. The structural question for Helium remains whether the carrier offload economics scale to compete with traditional MNO infrastructure, or whether the model finds a niche as a coverage augmentation layer rather than a primary network. The Nova Labs / Helium team's ability to negotiate with T-Mobile was the political moat.
Render, io.net, and the GPU compute unbundling
Render Network started in 2017 as a decentralized GPU rendering protocol for 3D and visual effects, with the original RNDR token rewarding GPU operators who provided rendering capacity to creators using Octane and similar tools. The 2023-2024 AI inference wave reframed Render's value: the same GPU fleet that could render frames could run AI inference, particularly for image and video generation workloads where latency tolerance was higher than for production LLM inference. Render migrated from Ethereum to Solana in 2023 and rebranded its token to RENDER, with consistent expansion of supported workloads. By 2026, Render's GPU operator base served a mix of rendering and AI inference, with revenue running at multi-tens-of-millions annualized. io.net took a different angle: a marketplace aggregating distributed consumer and enterprise GPUs into clusters that could be rented by AI labs for training and inference. io.net's IO token launched in 2024 and the network onboarded hundreds of thousands of GPUs onto its scheduler, though utilization rates have been a persistent question. By 2026, io.net's positioning shifted toward serving cost-sensitive AI startups and Web3-native AI workloads rather than competing head-on with hyperscalers. Akash Network, the older of the decentralized cloud players, scaled GPU support through 2024-2025 and found steady niche traction. The collective story across Render, io.net, and Akash is that decentralized GPU compute is real and growing but structurally smaller than centralized hyperscaler capacity, with the addressable market being cost-sensitive workloads that tolerate higher latency or reliability variance.
Hivemapper, DIMO, and physical-data DePIN
Hivemapper deployed dashcams to drivers in exchange for HONEY token rewards proportional to imagery contribution, building a competing road-imagery dataset to Google Street View and Mapbox. By 2026, Hivemapper's coverage of major U.S. roads exceeded major incumbents on freshness (with most coverage being weeks rather than years old), and the network signed enterprise customers in fleet, mapping, and autonomous-vehicle data licensing categories. Hivemapper revenue grew into the multi-millions annually, though the path to scale required negotiating with privacy regulators around facial and license-plate blurring. DIMO took a related angle for connected-vehicle data: drivers install a DIMO device or app in their car, opting into anonymized driving and vehicle telemetry data sharing in exchange for DIMO token rewards. By 2026, DIMO had crossed several hundred thousand connected vehicles and signed early enterprise customers in insurance, fleet management, and EV-charging optimization. WeatherXM (decentralized weather stations), GEODNET (high-precision GNSS), Foam (location attestation), and several other physical-data DePIN networks scaled similarly, each carving out a defensible data niche. The unifying thesis: hardware-collected data with tokenized incentives can build datasets that traditional venture-backed companies could not afford to collect, particularly in long-tail geographies.
Hardware token-incentive design and the inflation problem
Every DePIN network faces a fundamental design tension: rewards must be high enough to incentivize hardware deployment, but not so high that tokens trade as pure inflation absent corresponding revenue. The 2020-2022 era of Helium illustrated the failure mode - hotspot deployment far outstripped data demand, HNT inflation dramatically exceeded HNT-burn-from-data-credits, and the token price collapsed. The 2024-2026 generation of DePIN tokenomics learned from this. Modern designs include burn-and-mint equilibrium (where service consumption burns tokens at a rate calibrated to operator emissions), proof-of-coverage with utilization weighting (where rewards are partially tied to actual data served, not just presence), and dynamic emissions schedules (where issuance compresses as token velocity from real demand grows). Helium's MOBILE subDAO restructure was the highest-profile example. Render's BME (burn-mint-equilibrium) model became a template that several newer DePIN networks adopted. The structural challenge remains that during the bootstrapping phase, when end-user demand is small and hardware deployment is being subsidized, the token cannot be priced rationally on a discounted-cash-flow basis; investors are pricing the option that demand materializes. This is similar to early SaaS pricing rounds and equally prone to overshoot and undershoot.
Recent milestones (2024 to 2026)
Major milestones across the DePIN category between 2024 and 2026: Helium Mobile passed multi-hundred-thousand subscribers and announced expanded T-Mobile partnership terms in 2025; Render scaled inference workloads alongside rendering through 2025; io.net launched its IO token in 2024 and onboarded major GPU operator partners through 2025-2026; Hivemapper crossed enterprise-grade U.S. coverage and licensed data to multiple Tier-1 customers; DIMO crossed several hundred thousand connected vehicles by 2026; Foam, GEODNET, and WeatherXM scaled their respective data-collection networks; and several newer DePIN categories emerged in 2025-2026, including decentralized energy (PowerPod, Daylight), bandwidth-sharing (Karrier One, Pollen Mobile), and decentralized storage (Filecoin, Arweave, Walrus on Sui). Solana became the dominant DePIN settlement chain by 2026, hosting Helium, Render, Hivemapper, and DIMO; Solana's low fees, high throughput, and DePIN-friendly tooling (the Helium-pioneered HotSpot Onboarding Server pattern) created a network effect. Aggregate DePIN token market cap by Q1 2026 settled in the $15-25B range, with Bittensor and Filecoin as the largest by individual market cap among adjacent crypto-AI and decentralized storage categories. Real revenue across the top 10 DePIN protocols was estimated at $200-400M annualized run-rate by Q1 2026.
Risks, regulatory treatment, and outlook through 2027
DePIN's structural risks fall into three buckets. First, demand-side risk: most DePIN networks are still subsidizing supply with token rewards, and the existential question is whether organic demand can eventually fund the network. Helium Mobile's carrier ARR is the most concrete proof point but still requires growth. Second, regulatory risk: physical-infrastructure DePIN often touches regulated industries (telecom, mapping, vehicle data, energy, weather) where licensing, privacy, and compliance frameworks were not designed with token-incentivized networks in mind. Helium's mobile carrier required partnerships and licenses; Hivemapper navigates GDPR and California privacy law on imagery; DIMO addresses vehicle-data privacy. Third, hardware-supply-chain risk: many DePIN devices were manufactured during chip-shortage periods, with operators paying premiums that token rewards must justify on a multi-year time horizon. Looking forward to 2027, the realistic outlook is bifurcation. The two or three largest DePIN networks (Helium Mobile, Render, possibly DIMO and Hivemapper) will scale into recognizable real-revenue businesses with token-economic stability; the long tail of smaller DePIN networks will struggle as token-subsidy capital tightens. Solana's DePIN ecosystem advantage will likely deepen, with most new DePIN launches choosing Solana by default. The category as a whole will increasingly be measured by revenue rather than market cap, which is the maturity test.
Watch points
- Helium Mobile subscriber growth and T-Mobile partnership terms
- Render and io.net AI inference revenue trajectory and operator utilization rates
- Hivemapper enterprise data licensing revenue and DIMO connected-vehicle count
- Aggregate DePIN protocol revenue versus token issuance run-rate
TL;DR
DePIN networks (Helium Mobile, Render, io.net, Hivemapper, DIMO) graduated from supply-side speculation into real-revenue infrastructure by 2026, with $15-25B aggregate token market cap and multi-hundred-million annualized revenue, though sustained unit economics post-token-subsidy remains the binding test through 2027.
Get DeFi Intel research in your inbox
Weekly long-form coverage of papers, incidents, jurisdictions, chains, tokens and the people building them. Free tier covers headlines; Pro adds the analyst-grade breakdowns.