01 The set
What is on this board
One tag: decentralised physical infrastructure networks. The top 1,000 by market cap holds 65 of them, and all 65 are on the board.
This board overlaps the AI board heavily, and most of all at the top. The largest names here rent out GPUs, storage or bandwidth, which is simultaneously a physical-infrastructure business and an AI one, so they carry both tags and are drawn on both pages. If the two look similar in their largest bubbles that is not a bug in the filter — it is the same companies, seen from two directions. What differs is the question each page asks about them.
Everything else works the way it does on the whole-market board: size is market cap, color is the price change over the period on the buttons, and grey means a coin barely moved. The longer version, with the exact color thresholds, is in the guide.
02 Movers
Biggest moves in the sector, last 24 hours
▲ Gainers
| Coin | Price | 24h | Volume |
|---|---|---|---|
| GEODGEODNET | $0.2870 | +25.74% | $21M |
| IOio.net | $0.1433 | +14.66% | $29M |
| FILFilecoin | $0.7740 | +12.37% | $169M |
| PHAPhala Network | $0.0254 | +10.87% | $12M |
| 0G0G | $0.1705 | +10.71% | $21M |
▼ Losers
| Coin | Price | 24h | Volume |
|---|---|---|---|
| RIFRootstock Infrastructure Framework | $0.0789 | +2.59% | $10M |
| ZBTZEROBASE | $0.0801 | +2.76% | $11M |
| RENDERRender | $1.47 | +4.78% | $43M |
| TAOBittensor | $223.53 | +4.98% | $300M |
| ICPInternet Computer | $2.46 | +4.99% | $76M |
Both tables are drawn from this board only, and every coin in them traded at least $10M in the past 24 hours. That volume floor is the same one the whole-market board uses, and it matters more here: a sector is a slice, so fewer coins clear a fixed dollar amount, and the ones that do not are exactly where a spectacular percentage turns out to be one buyer on a quiet market.
03 Background
What is DePIN?
DePIN stands for decentralised physical infrastructure networks: paying people in a token to go and install real hardware, so that a network gets built without a company raising the capital to build it. The hardware is genuinely physical — wireless hotspots on rooftops, GPUs in spare rooms, hard drives, dashboard cameras mapping streets, sensors reporting weather or air quality, antennas receiving satellite positioning corrections.
The token solves a specific problem, and it is worth naming precisely. Infrastructure has a chicken-and-egg failure: nobody pays for coverage that does not exist yet, and nobody builds coverage nobody is paying for. Traditionally an investor bridges the gap and waits years. Here the network issues a token to whoever installs the hardware, so the buildout is paid for by people who expect the token to be worth something once demand arrives. It is a way of financing capital expenditure from thousands of small operators instead of one balance sheet, and when it works it is genuinely faster and cheaper than the alternative.
The failure mode is built into the same mechanism. Emissions reliably buy supply — hardware shows up, coverage maps fill in, the device count goes up and to the right. Emissions do not buy demand. Nothing about paying people to install antennas causes anyone to want the service, and the sector's history is full of networks with impressive coverage and almost no paying customers. So the number that matters for one of these is revenue from users, not the size of the network, and the two are routinely confused — including by the projects themselves, whose dashboards tend to lead with the device count.
Ask which side of the ledger the yield is coming from. If operators are being paid mostly in freshly issued tokens, the network is subsidising itself and the payouts fall as issuance tapers, regardless of how well it is running. If they are being paid out of what customers pay, it is a business. Most of this board is somewhere between, and moving along that line is the whole plan. None of it is visible here.
The hardware makes this sector behave unlike the rest of the site. An operator paid a real electricity bill and bought a real device that loses value every month. That gives the sector something almost nothing else in crypto has — a genuine cost floor and a genuine reason for participants to leave when the token falls below it. It also makes the buildout hard to reverse and hard to fake: an antenna either exists on a roof or does not. When these networks fail they tend to fail slowly, as operators unplug, rather than at the speed a purely on-chain economy unwinds.
Geography is a constraint the picture cannot show. Coverage is worth something only where people are. A network with a hundred thousand devices concentrated where bandwidth is already cheap has less to sell than one with far fewer in the right places. The same applies to compute: GPUs matter by model and by where they sit relative to the data. Two bubbles of the same size can be networks with completely different value, and market cap does not distinguish them.
What DePIN is competing with is worth remembering. The incumbent is usually a large, efficient, heavily capitalised operator — a carrier, a cloud provider, a mapping company — with economies of scale that decentralisation gives up. The case for these networks is that coordination costs, idle capacity and the last mile are where those incumbents are weakest. That is a real argument, not a sure one, and this board is a record of the market's changing opinion about it rather than of how it is going.
04 The app
The same filter, over the whole market
This page is one fixed slice: one sector, the largest names in it, four periods, prices in dollars. The app it was cut from carries all 1,000 coins and lets you set the filter yourself — by category, by chain ecosystem, by exchange — then keep what you found as favorites. Free, no ads, no account, and it runs in a browser tab as well as on a phone.
Open the full board →