DePIN had the rare honour of being the loudest narrative of 2024 and the quietest narrative of 2026. The networks are still here. The tokens, mostly, are not. The thesis below: this is the first time the sector has been worth analysing on its actual cash flows.
What is the DePIN Revenue Divergence in 2026?
DePIN, short for Decentralized Physical Infrastructure Networks, covers projects that pay token rewards to operators who supply real-world resources: GPUs, wireless coverage, storage, sensors, mapping data. The pitch was simple. Bootstrap a physical network with token incentives, sell the service, watch the token capture the value. The pitch did not survive contact with the price chart.
According to Decrypt's January 2026 analysis on DePIN fundamentals, tokens launched between 2018 and 2022 are now lagging their all-time highs by 94 to 99 percent. Combined market capitalization sits around $10 billion. On-chain revenue, the actual money paid by users for the actual service, came in at roughly $72 million for full-year 2025.
So here is the divergence. Tokens lost most of their notional value while networks went from selling nothing to selling something measurable. Leading projects now trade at 10 to 25 times revenue, down from 1,000-plus times during the 2021 cycle. The numbers say yes. The panda raises an eyebrow.
The thesis: the price collapse is not the story. The story is that, for the first time since the category existed, you can build a valuation model with a positive denominator. Whether that valuation is generous or stingy is a different debate. The fact that it is possible at all changes the conversation.
The Token Charts and the Revenue Receipts
Let us walk through the receipts.
RENDER trades near $1.82 with a market cap of roughly $944 million, down about 86 percent from its $13.60 high in March 2024. Network revenue is roughly $207,000 monthly, with the Render Foundation targeting $5 million monthly by Q4 2026. That is a 24x ramp asked from a sector that does not get the benefit of the doubt anymore. According to BanklessTimes on May 24, 2026, one analyst sees a wedge breakout setup toward $7.70. The chart says maybe. The fundamentals say earn it.
Bittensor (TAO) sits in a different bracket. As of early May 2026, TAO traded around $289 with a market cap of $2.77 billion. The ecosystem has crossed $1.5 billion in combined subnet market cap and generated over $43 million in real AI usage revenue in Q1 2026 alone. The network had 128 live subnets by late 2025 and just upgraded its reward system to concentrate emissions on top-performing subnets, cutting fluff from the bottom. Bittensor is doing the boring thing infrastructure projects are supposed to do: kill weak subgraphs, reward strong ones, repeat.
Helium Mobile is the most surprising entry on the chart. According to SolanaFloor's May 2026 report, Helium Mobile has accumulated over $14 million in cumulative revenue on Solana and crossed 500,000 sign-ups. The wireless sector offloaded 37,000 terabytes of data, a 12 percent jump from January 2026. Since mid-August 2025, 100 percent of Helium Mobile subscriber revenue is used to buy and burn HNT from the open market. The boring engine works.
Akash Network is the GPU compute side. GPU utilization peaked above 79 percent. Daily network revenue hit a record around $9,000. Quarterly lease income reached $851,700 by Q3 2025 per Messari's State of Akash Q3 2025 report. These numbers are small compared with AWS. They are also not zero, which is more than DePIN could say in 2022.
Filecoin is the storage tail of the same story. The token has bled steadily from its 2021 highs, but the network now stores meaningful enterprise and academic datasets, including chunks of climate research and large-language-model training data. Storage utilization tells you nothing exciting on a daily chart. It tells you everything about whether the network has a reason to exist in five years. The chart and the utilization are not on speaking terms. They rarely are in this sector.
Look at the four projects above. The pattern is consistent. Token price: somewhere between flat and disaster. Network revenue: small in absolute terms, growing in relative terms, increasingly tied to AI compute demand rather than crypto-cycle reflexivity. Token mechanics: most have been quietly rewired with buyback-and-burn programmes, concentrated emissions, or reduced inflation schedules. The 2022 tokenomics that paid operators in unlimited dilution are mostly retired.
Per the same Decrypt analysis, the sector raised approximately $1 billion in venture funding during 2025 even as token prices stagnated. VCs are not pricing the chart. They are pricing the customer pipeline. Worth remembering: the same investors who funded the speculative cycle are now writing cheques on revenue multiples. That is either capitulation or conviction, depending on your priors.
The broader macro frame matters. According to CoinGecko's global market data, total crypto market capitalization sat at $2.55 trillion on May 29, 2026, with Bitcoin dominance at 57.67 percent. DeFi total value locked, per DefiLlama, stands at $79.81 billion, still 60 percent below its 2021 peak. The wider crypto market is back near cycle highs. DePIN is the small, quiet appendix of this story. Few people read appendices. The ones who do tend to be the ones who matter.
Why the Tokens Lagged So Badly
Four reasons, in descending order of charity.
First, tokenomics. Most DePIN networks emitted aggressively in early years to subsidize operator hardware purchases. Operators sold the rewards immediately to pay back hardware costs. The result: persistent sell pressure with no compensating buy pressure from end users, because there were no end users yet. This was structurally similar to what BSC memecoins eventually confronted (see our memecoin tokenomics cluster), with a few years of lag.
Second, valuation overshoot. At the 2021 peak, the sector traded as if AWS-scale revenue was three quarters away. It was not. Mean reversion did the rest. Markets occasionally overshoot in both directions. The sector spent 2022 to 2024 paying back the overshoot.
Third, narrative collapse. DePIN got grouped with AI crypto in early 2024, rallied with the wider AI-token mania, then fell harder than the rest when the narrative compressed. The good news: when a sector falls outside the spotlight, you get to study it without the FOMO tax.
Fourth, opportunity cost. From 2023 to early 2025, every dollar of crypto capital that could have flowed into DePIN had three louder homes competing for it: Bitcoin spot ETFs, memecoin launchpads, and the AI-token cycle. Infrastructure rarely beats narrative on a one-year clock. It often wins on a five-year one. The longer the audience looks elsewhere, the more time builders get to fix the boring stuff: hardware partnerships, enterprise contracts, regulator conversations.
For context on how this overlaps with the broader AI buildout, see our recent DePIN GPU networks under AI squeeze analysis and the AI agents pillar covering how on-chain compute is starting to migrate to chains with lower fees.
Counterarguments: Honest Objections
A 2,000-word bull thesis deserves an honest section of objections. Here are the strongest.
Objection one: $72 million in revenue across the entire sector is, frankly, rounding error. Cloudflare did $1.7 billion in a single quarter. Equinix did $9 billion last fiscal year. The DePIN thesis asks us to compare seedlings to redwoods and pretend the size difference is incidental. Fair point. The counter: every infrastructure sector started as rounding error. The question is the slope, not the level.
Objection two: the revenue might not be sticky. AI compute spending in 2025 was inflated by inference workloads from a handful of large customers. If those customers move to AWS or build their own GPUs (NVIDIA's DGX Cloud is right there), DePIN revenue collapses faster than it rose. This is real. The strongest defense is workload diversity: Bittensor's 128 subnets, Akash's mix of inference and training, Helium's carrier offload contracts. Single-customer concentration is the actual risk to watch.
Objection three: the burn-and-buyback mechanics are gimmicks. Helium burning HNT only matters if there is enough non-burn demand to absorb the existing float. Render's $4.3 million annualized Salad inflow is real, but it is also one integration. Buybacks without organic demand are just a slower version of dilution. Fair enough. This is the most uncomfortable critique in the set.
Objection four: the entire category may be regulatory poison waiting for a court date. Token-incentivized hardware operators look a lot like an unregistered security distribution under any post-Howey reading. SEC enforcement has been intermittent here, but a single adverse ruling could vaporize half the sector overnight. Spoiler: we saw this one coming.
What to Watch Next
Five concrete things, with dates, that will decide whether the thesis above holds or breaks.
End of Q3 2026: Render must show monthly revenue above $500,000 to keep the $5 million Q4 target plausible. Anything below $400,000 means the Salad integration was the ceiling, not the floor.
By October 2026: Bittensor's spot ETF filing has a decision window. Approval would route institutional flows through the most fundamentals-oriented project in the sector. Rejection would push retail back to short-duration narratives.
End of fiscal 2026: Helium Mobile needs to cross $30 million in cumulative carrier offload revenue (currently around $14 million as of May 2026). The trajectory looks feasible. The carrier deals get easier as the data offload numbers climb. The panda is watching.
February 2027: aggregate sector revenue print. If the $72 million 2025 number doubles to $150 million for full-year 2026, the 10-25x revenue multiple holds and the sector has a defendable floor. If it stalls near $90 million, the multiple compresses and tokens drop another leg.
Q2 2027: at least one SEC or international regulator enforcement action against a top-five DePIN token. The base rate suggests it happens. The question is whether the network keeps running through it. Filecoin survived 2024 scrutiny. Helium has been pressure-tested. Whoever is next gets to find out if they were building infrastructure or selling securities.
DePIN is in its boring years. The hype is gone. The networks still work. The revenue is finally measurable. The tokens have not noticed yet. That last clause is the entire thesis.
This is not investment advice (we are allergic to giving any). It is a structural observation that the sector most easily mocked in 2024 is the one quietly building the income statement nobody asked for. Infrastructure usually wins on a longer clock than crypto can stomach. The clock is still running.
The same logic applies to chains that prioritize cheap, boring, working infrastructure over headline narratives. Lower fees, denser activity, fewer fireworks. BSC has been absorbing more of that traffic than its dominance score suggests. The panda has noticed. The market eventually will.
