Decentralized Physical Infrastructure Networks — DePIN — have crossed from conference-stage speculation into revenue-generating territory. The sector's combined market capitalization now exceeds $19 billion across nearly 250 tracked projects, up from $5.2 billion just twelve months ago. More impo...
"DePIN is being forced into fundamentals. 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." — Markus Levin, Co-Founder, XYO
Decentralized Physical Infrastructure Networks — DePIN — have crossed from conference-stage speculation into revenue-generating territory. The sector's combined market capitalization now exceeds $19 billion across nearly 250 tracked projects, up from $5.2 billion just twelve months ago. More importantly, a subset of networks is posting real, recurring revenue from enterprise customers who are paying for compute, connectivity, and data — not farming airdrops.
Yet DePIN tokens remain deeply depressed. Many are 94–99% below all-time highs. This disconnect — rising revenues, falling token prices — makes DePIN the most critical economic test case in crypto today. If these networks can sustain cash flows without subsidies, they become the first sector to validate the blockchain industry's decade-old promise of decentralized infrastructure replacing centralized incumbents. If they cannot, the $19 billion in market cap will join the graveyard of narratives that never generated a dollar of organic demand.
This report examines who is actually generating revenue, where enterprise demand is real, and whether DePIN's economics can survive without the crutch of token emissions.
The DePIN sector generated an estimated $72 million in on-chain revenue during fiscal year 2025, with leading networks now trading at 10–25x revenue — a dramatic compression from the 1,000x+ multiples seen during the 2021 cycle. This shift toward fundamental valuation marks a structural maturation.
The revenue hierarchy reveals extreme concentration:
| Project | Annualized Revenue | Primary Revenue Source | Enterprise Clients | |---------|-------------------|----------------------|-------------------| | Aethir | $147M ARR | GPU compute for AI/gaming | 150+ clients | | Grass | $33M | AI training data from bandwidth | AI companies | | Helium | $13.3M | Wireless data offloading | T-Mobile, AT&T, Telefónica | | Render | ~$8–12M est. | GPU rendering & AI inference | Studios, AI developers | | Hivemapper | $3M ARR | Mapping data licensing | Enterprise mapping clients |
Aethir's $147 million annualized revenue run rate — generated from over 150 enterprise AI, Web3, and gaming clients across 435,000 GPU containers in 93 countries — is the standout number. At $1.25/hour for NVIDIA H100 GPUs versus AWS's $6.04/hour, Aethir is undercutting centralized cloud by 79% while achieving Q3 2025 revenue of $39.8 million alone.
In January 2026, Solana-based DePIN protocols collectively hit an all-time high in monthly revenue, with Helium and XNET delivering over 600% revenue growth year-over-year. The total for Helium, Render, Hivemapper, UpRock, NATIX, XNET, and Geodnet reached $2.6 million in a single month, exceeding the previous record from September 2025.
The AI infrastructure boom has handed DePIN its most compelling demand-side narrative. U.S. tech companies are expected to invest approximately $650 billion in AI infrastructure by 2026, according to Bridgewater Associates. NVIDIA has publicly stated that demand will exceed supply for several quarters into fiscal 2026. This structural GPU shortage is DePIN's window.
Decentralized GPU networks have shown 428% year-over-year growth in usage, with utilization rates above 80% heading into 2026. The cost advantage is substantial:
| Provider | H100 GPU ($/hr) | Discount vs. AWS | |----------|-----------------|-----------------| | AWS | $6.04 | — | | Azure | $5.80 | 4% | | Aethir | $1.25 | 79% | | Akash | $1.50–2.50 | 59–75% | | Hyperbolic | ~$1.50 | 75% |
Akash Network uses a reverse auction model where GPU providers compete for workloads, driving costs down while maintaining quality. Render Network, which launched its dedicated AI compute subnet "Dispersed" in December 2025, has processed over 65 million cumulative frames and currently operates 5,600 active GPU nodes with 85–95% utilization rates.
The key limitation remains synchronous training workloads. Centralized providers retain an advantage for training massive foundational models requiring ultra-low-latency interconnects. DePIN excels at asynchronous tasks: AI inference, 3D rendering, distributed fine-tuning, and edge machine learning. By 2026, inference workloads are projected to account for two-thirds of total AI compute usage — precisely the segment where decentralized networks are most competitive.
Aethir is planning to more than double its compute network by Q1 2026, establishing a Strategic Compute Reserve (SCR) to accelerate GPU expansion and enterprise deals. This is infrastructure scaling driven by signed contracts, not speculative capacity building.
Helium represents DePIN's most mature enterprise integration story. The network has surpassed 500,000 subscriber sign-ups for Helium Mobile and maintains partnerships with T-Mobile, AT&T, and Telefónica for wireless data offloading.
The economics are tangible: daily Data Credit burns consistently hit $50,000, driven by actual data offloading. In January 2026, Helium Mobile added 29,000 new subscribers (a 31% month-over-month decline, signaling potential subscriber fatigue). Meanwhile, Dabba Network, Helium Mobile, and XNET collectively delivered 33,000 terabytes of offloaded data in January 2026, up 22% from the prior month.
Helium's transition to a unified HNT token model — ceasing IOT and MOBILE emissions in January 2025 — simplified its economics. The burn-and-mint mechanism creates deflationary pressure: users burn HNT to create Data Credits for network fees, linking token value directly to network demand. This is exactly the kind of self-reinforcing economic loop that the broader crypto industry has failed to achieve at most Layer-1 networks, where subsidy-driven flows account for 85–90% of total value.
The critical question is whether $13.3 million in annualized revenue can justify Helium's ~$320 million market cap — a roughly 24x revenue multiple. For a growth-stage telecom infrastructure play with contracted enterprise partners, this multiple is defensible. For a network still dependent on token emissions for node operator incentives, it warrants scrutiny.
The convergence of AI and DePIN — termed "DePAI" (Decentralized Physical AI) by the World Economic Forum — has created an entirely new demand category: decentralized data collection and distribution for AI model training.
Grass has emerged as the sector's breakout data play, monetizing unused internet bandwidth from 8.5 million users to generate $33 million annually by selling AI training data. The model is elegant: users contribute idle bandwidth, Grass aggregates it into commercially valuable datasets, and AI companies pay for access. No token farming gimmick — actual enterprise procurement.
Hivemapper boosted its 30-day ARR from $500,000 to $3 million in late 2025 by lowering entry barriers for data suppliers and introducing region-based staking. The network has generated approximately $600,000 in cumulative revenue since January 2025, with mapping data licensed to enterprise clients.
Bittensor (TAO) operates the most ambitious decentralized AI marketplace, with over 128 active subnets specialized in tasks from image generation to prediction models. Its first halving on December 15, 2025 cut daily emissions from 7,200 to 3,600 TAO, reducing inflation from 26% to 13% annually. With a $2.7 billion market cap, Bittensor is the largest pure-play DePIN AI asset — though its revenue generation remains more opaque than compute or connectivity networks.
Filecoin is executing a strategic pivot from capacity growth to monetization. Operating at exbibyte scale (2.1 EiB of secured data, 7.6 EiB raw capacity), Filecoin launched its Onchain Cloud in November 2025, positioning itself as a programmable cloud layer with verifiable storage, automated payments, and compute capabilities. The 2026 priority is clear: convert infrastructure into paid deals and onboard flagship enterprise production workloads.
The foundational question for DePIN — consistent with the broader blockchain economic reality — is whether these networks can sustain themselves without token emission subsidies.
The data presents a mixed picture. Real revenue exists, but for most networks, token emissions still dwarf fee income:
| Network | Annual Revenue | Estimated Annual Emissions | Revenue/Emissions Ratio | |---------|---------------|--------------------------|------------------------| | Aethir | $147M | $200–400M (est.) | 37–74% | | Grass | $33M | Undisclosed | N/A | | Helium | $13.3M | $50–100M (est.) | 13–27% | | Render | $8–12M | $30–60M (est.) | 13–40% | | Filecoin | Pivoting | $500M+ | <5% |
Aethir stands closest to sustainability, with revenue potentially covering a significant portion of its emission costs. Helium and Render are in the "promising but subsidized" category — generating real demand but still relying on inflation to compensate node operators. Filecoin, despite its massive infrastructure scale, remains firmly in the subsidy-dependent zone.
This pattern mirrors the broader blockchain economy where approximately 85–90% of ecosystem value flows are subsidy-driven. The difference is that DePIN subsidies fund physical infrastructure with measurable utilization, rather than abstract network security with minimal fee generation.
As Markus Levin put it: "When investors can point to real demand, recurring revenue, and clearer paths to scaling capex, they write bigger checks." DePIN startups raised approximately $1 billion in 2025, primarily at seed and Series A stages, signaling that private-market conviction persists even as public token prices compress.
Three institutional developments signal that DePIN has crossed the credibility threshold:
SEC No-Action Letter for DoubleZero (September 2025): The SEC issued a no-action letter for DoubleZero's 2Z utility token — a landmark moment that legitimized DePIN utility tokens and opened the door for institutional capital without regulatory ambiguity.
Grayscale Bittensor Trust (December 2025): Grayscale launched GTAO, giving accredited investors exposure to decentralized AI compute infrastructure through a traditional fund wrapper. This is the first institutional-grade DePIN product from a major asset manager.
Venture Capital Concentration: DePIN attracted the highest share of crypto-VC inflows in early 2025 according to Messari research, with total crypto VC funding surging to $30 billion in Q4 2025. The institutional thesis is clear: DePIN offers cash-flow-based crypto exposure in a market starved for fundamentals.
The World Economic Forum projects the DePIN sector could reach $3.5 trillion by 2028 — implying a 375% compound annual growth rate. Even at half that trajectory, DePIN would represent one of the fastest-growing infrastructure sectors across any industry.
Revenue is real but concentrated. Aethir ($147M ARR), Grass ($33M), and Helium ($13.3M) generate the vast majority of DePIN's on-chain revenue. The long tail of 250 projects is largely pre-revenue.
The GPU shortage is DePIN's structural tailwind. With AI inference projected to consume two-thirds of compute by 2026, and decentralized networks offering 60–80% cost savings over AWS, DePIN compute has a clear product-market fit for asynchronous AI workloads.
Token prices and fundamentals have decoupled. DePIN tokens are down 94–99% from all-time highs while revenues grow. Networks now trade at 10–25x revenue versus 1,000x+ in 2021 — a repricing toward fundamentals that should attract institutional allocators.
Subsidy dependence persists. Most DePIN networks still rely on token emissions to compensate infrastructure operators. Only Aethir approaches a potentially self-sustaining revenue-to-emissions ratio.
Institutional validation is accelerating. The SEC no-action letter for DoubleZero, Grayscale's Bittensor Trust, and concentrated VC inflows signal a maturing sector that is crossing the credibility threshold for traditional finance.
DePIN stands at the intersection of crypto's greatest promise and its most persistent failure mode. The promise: decentralized networks replacing centralized infrastructure through token-coordinated incentives. The failure mode: subsidized economics masquerading as organic demand.
What makes DePIN's 2026 moment different is that the demand side is no longer hypothetical. T-Mobile is paying Helium for data offloading. AI companies are paying Grass for training data. Enterprises are paying Aethir for GPU compute at 79% below AWS rates. These are procurement contracts, not airdrop campaigns.
The sector's $19 billion market cap and $72 million in annual on-chain revenue imply an aggregate 263x revenue multiple — still elevated by traditional standards, but a universe away from the infinite multiples of networks generating zero fee revenue. As the subsidy-to-revenue ratio compresses and AI infrastructure demand accelerates, DePIN may become the first crypto sector to prove that decentralized coordination can build real infrastructure at competitive costs.
The tokens may be down 99%. The infrastructure is just getting started.