The DePIN (Decentralized Physical Infrastructure Networks) sector generated $72 million in on-chain revenue in 2025 across 650+ projects, according to Messari's State of DePIN 2025 report. That figure is projected to exceed $100 million in 2026 as enterprise AI teams turn to decentralized compute...
"The market stopped rewarding projects for their narratives and started demanding actual metrics: revenue per node, utilization rates, and paying customers." — Carlos Lei, CEO, Uplink
The DePIN (Decentralized Physical Infrastructure Networks) sector generated $72 million in on-chain revenue in 2025 across 650+ projects, according to Messari's State of DePIN 2025 report. That figure is projected to exceed $100 million in 2026 as enterprise AI teams turn to decentralized compute for overflow capacity. In January 2026 alone, leading DePIN networks reported roughly $150 million in aggregate monthly on-chain revenue from storage deals, compute jobs, data credits, and mapping services.
The numbers tell two stories simultaneously. On one hand, total sector market capitalization has stabilized at approximately $10 billion, with DePINscan tracking over 8.8 million active devices across 199 countries. On the other hand, revenue concentration is severe: fewer than 20 of the 650+ projects generate meaningful non-token revenue, and just seven have crossed $10 million in annual recurring revenue. The average DePIN project generates approximately $110,000 per year. For every Aethir reporting $166 million in ARR, there are dozens of sensor networks and bandwidth projects burning through treasury faster than they acquire paying customers.
This report examines the economic structure of the DePIN sector in mid-2026, comparing revenue models, tokenomics mechanisms, and utilization rates across the four dominant subsectors: compute, wireless, storage, and sensors.
Messari's State of DePIN 2025 report, published in January 2026, catalogued a sector with a $10 billion combined market capitalization generating $72 million in annual on-chain revenue. The ratio — roughly 139x revenue — represents a compression from over 1,000x multiples seen during the 2021 cycle. Leading networks now trade at 10-25x revenue, according to the same report.
The fundraising pipeline remains active. DePIN projects raised approximately $1 billion in 2025, up from $698 million in 2024, according to Messari. That capital, however, flows unevenly. Compute-focused projects — those serving AI training, inference, and rendering workloads — captured the majority of investor interest.
Token prices have diverged from revenue trajectories. Helium's HNT token declined 77% from December 2024 to December 2025, while its on-chain revenue increased roughly eightfold over the same period. GEODNET's GEOD token fell 41% while revenue grew 1.7x. As Markus Levin, co-founder of DePIN project XYO, told Decrypt: "DePIN is being forced into fundamentals."
GPU infrastructure spending is projected to reach $353 billion by 2030, up from $83 billion in 2025, according to industry estimates. Decentralized compute networks are positioning themselves to capture overflow demand from enterprises priced out of hyperscaler queues.
Aethir leads the subsector. The platform reported $166 million in annualized recurring revenue as of Q3 2025 — a 13x year-over-year increase — generated from enterprise customers across 94 countries and over 200 locations. Aethir operates 440,000+ containers running primarily on NVIDIA H100 hardware, serving AI training, inference, and cloud gaming workloads.
Render Network recorded $38 million in revenue in January 2026. Token burns tied to compute purchases rose approximately 279% year over year. By Q2 2026, Render's infrastructure reporting indicated demand for GPU processing power was outstripping available capacity. However, Render's USD revenue is not disclosed in a standardized, audited format, complicating direct comparisons.
Akash Network crossed $5 million in cumulative compute spend in Q1 2026, its highest quarter on record. Lease revenue of $253,250 accounted for 98% of the $257,580 in total network fees during Q1 2026, according to Messari's State of Akash Q1 2026 report. The network aggregated approximately 250 active GPUs across 73 providers as of May 2026, generating roughly $4.3 million in annualized revenue. Notably, GPU capacity contracted 57.5% quarter-over-quarter in Q1 2026, alongside declines in CPU (-46.5%), storage (-37.5%), and RAM (-41.7%) capacity — suggesting provider economics may not sustain participation at current token prices.
io.net closed $8 million in enterprise deals in Q1 2026. The network claims access to over 100,000 GPU devices. Independent data, however, placed verified-active GPUs at approximately 2% of registered supply, according to an analysis cited by Own Your Mind. That utilization gap raises questions about the difference between stated capacity and actual economic activity.
Helium Mobile reached $2.5 million in monthly revenue in March 2026, a 14% increase from the prior month, according to Solana Floor. Q1 2026 mobile revenue nearly matched full-year 2025 figures. The network reports over 500,000 mobile subscribers at a $20/month consumer price point.
The business model has shifted. Carrier offload — where major U.S. carriers route traffic through Helium's community-operated hotspots — now accounts for 57% of total revenue, up from a minority share in 2025. Subscriber fees represent the balance. This dual-revenue structure reduces dependency on retail acquisition costs and aligns Helium's economics with wholesale telecom pricing.
Helium completed its migration to Solana, consolidating its IoT and Mobile networks onto a single settlement layer. The move reduced infrastructure costs and improved data credit throughput. Whether the carrier offload model scales beyond its current geographic footprint — concentrated in U.S. metro areas — remains an open question. Major carriers have not publicly disclosed the volume or pricing terms of their offload agreements with Helium.
Filecoin's 2026 network strategy, published by the Filecoin Foundation, explicitly frames the year as a transition from protocol subsidies to paid storage revenue. Active storage from paid deals has stabilized at approximately 1,110 PiB, with the network targeting 1 exbibyte in paid deals by year-end.
The revenue context is sobering. Filecoin generated $180,700 in protocol fees over a 30-day period in early 2026, ranking eighth among Layer 1 blockchains. Against roughly $725 billion in hyperscaler AI infrastructure spending projected for 2026, Filecoin's revenue remains a rounding error. The network's go-to-market focus has narrowed to AI dataset archiving, chain data, and real-world asset documentation — verticals where decentralized storage offers compliance or redundancy advantages that centralized providers do not.
The broader question for Filecoin is whether its storage-provider economics can sustain participation as block rewards decline. Providers must commit hardware and energy costs against uncertain future paid-deal revenue. If utilization of committed capacity does not increase, provider attrition could reduce network reliability precisely when enterprise adoption would need to accelerate.
Hivemapper represents the strongest revenue growth story in the sensor subsector. Annualized revenue rose from $500,000 in August 2025 to approximately $18 million by early 2026 — a 36x increase — driven by enterprise customers purchasing fresh street-level mapping data. The network has mapped 28% of the world's roads, up from 10% in 2024. Drivers earn HONEY tokens by capturing imagery with standard dashcams.
GEODNET reported Q3 2025 revenue of $1.23 million, a 216% year-over-year increase. The network operates 21,000 active stations globally, serving agriculture, construction, and autonomous vehicle companies that require centimeter-accurate positioning data.
DIMO has connected over 165,000 vehicles, with drivers logging more than 100 million miles. The revenue model relies on selling anonymized mobility data to insurers, original equipment manufacturers, and fleet operators. Specific revenue figures for 2026 have not been publicly disclosed.
These sensor networks share a common structural advantage: the data they produce has no centralized equivalent at comparable cost. Google Street View updates on multi-year cycles; Hivemapper contributors update daily. Traditional RTK positioning networks charge thousands per base station per year; GEODNET offers equivalent accuracy at a fraction of the cost. Whether that cost advantage translates to sustainable token-denominated revenue — and whether data buyers will accept decentralized networks as reliable counterparties — remains unresolved.
The dominant tokenomic innovation in DePIN during 2026 is the shift from inflationary emission schedules to demand-driven burn mechanisms.
Akash launched Burn-Mint Equilibrium (BME) in March 2026, automatically purchasing and burning AKT tokens whenever customers pay for compute. The mechanism directly links token scarcity to network usage, replacing the prior inflationary emission model.
io.net introduced its Incentive Dynamic Engine (IDE), expected to roll out in Q2 2026, replacing fixed emissions with a system that stabilizes GPU provider payouts in USD terms and dynamically adjusts token supply based on real-time revenue and token price.
The theoretical promise of these mechanisms — a self-reinforcing loop where demand burns tokens, reducing supply, supporting price, attracting operators, expanding capacity, and attracting more demand — requires one precondition that most DePIN networks have not met: sufficient demand-side revenue to offset token issuance to operators. When inflation outpaces burns, the loop runs in reverse.
The central economic challenge facing DePIN in mid-2026 is the gap between deployed capacity and utilized capacity. Nearly 9 million devices are registered across 199 countries, according to DePINscan. The proportion of those devices generating revenue from paying customers — as opposed to earning token emissions for being online — is not systematically reported across the sector.
Where data exists, the picture is mixed. io.net's verified-active GPU rate of approximately 2% of registered supply suggests significant overcapacity. Akash's 250 active GPUs across 73 providers, despite claiming a much larger registered base, tells a similar story. Helium's carrier offload model achieves higher utilization in covered metro areas but has limited geographic reach.
The utilization gap matters because DePIN economics ultimately depend on whether the cost of operating a node — hardware, energy, bandwidth, maintenance — can be covered by revenue from paying customers, not just token emissions. As token prices compressed through 2025, operators whose economics depended on emission income exited. Akash's 57.5% GPU capacity contraction in Q1 2026 may reflect this dynamic.
The DePIN sector generated $72 million in on-chain revenue in 2025 across 650+ projects, with Messari projecting $100 million+ for 2026. Revenue is concentrated: seven projects account for the majority, while the average project generates $110,000 annually.
Compute networks lead sector revenue, driven by AI demand overflow. Aethir's $166 million ARR dwarfs peers, but revenue disclosure standards vary widely, complicating comparisons.
Helium's shift to carrier offload (57% of revenue) demonstrates a viable B2B model. Whether U.S. carriers will expand offload volume — and at what price — will determine Helium's revenue ceiling.
Filecoin's revenue remains negligible against hyperscaler competition ($180,700/month vs. $725 billion in annual hyperscaler AI infrastructure spend). The network's value proposition depends on niche verticals where decentralization offers compliance or redundancy benefits.
Tokenomics are evolving from inflationary emissions to burn-mint mechanisms (Akash BME, io.net IDE), but these models require demand-side revenue to exceed operator token issuance — a threshold most projects have not reached.
Utilization rates, where disclosed, suggest significant overcapacity. The gap between registered devices and revenue-generating devices is the sector's most material risk.
DePIN in mid-2026 occupies an unusual position in the crypto landscape: a sector where revenue is growing, token prices are compressed, and the gap between narrative and economics is narrowing but not closed. The $72 million in 2025 on-chain revenue is real. The $150 million monthly figure cited for January 2026 — if sustained — would represent a meaningful acceleration. But the concentration of that revenue in fewer than 20 projects out of 650+ indicates a sector where most participants are not economically viable.
The projects that have found traction share a common characteristic: they sell services to buyers who would otherwise purchase from centralized alternatives. Aethir sells GPU compute to AI companies. Helium sells wireless offload to carriers. Hivemapper sells mapping data to logistics firms. Geodnet sells positioning data to precision agriculture operators. In each case, the buyer's decision is driven by cost, availability, or geographic coverage — not by ideological commitment to decentralization.
The sector's near-term trajectory depends on whether the handful of revenue-generating projects can scale, whether tokenomic reforms can align operator incentives with demand-side economics, and whether the 630+ projects without meaningful revenue can either find product-market fit or wind down without damaging sector credibility. The data available as of mid-2026 suggests that DePIN's economic model works — but only for a small minority of participants.