DeepChain深链
Other explainer 02 / 04  |  app-chains  |  app  |  4 min read
APP

Custom chains: one project, its own chain

Some projects would rather not share a public chain with everyone else, so they build a chain of their own (an app-chain). They set their own performance and rules, and in exchange carry their own security, validators and bridges.

Other · App-chainsData as of 16 Oct 2026, 01:45 UK time
Hyperliquid TVL
$1.08B
dYdX TVL
$70.3M

In one sentence

A public chain is like a shopping centre: anyone can open a shop, but when it gets busy there are queues, and the rules (fees, block speed, transaction ordering) are set by the centre. An app-chain is a project putting up its own building: it runs only its own application, sets its own rules and keeps its own fee revenue.

The price: it has to provide its own security too. It must find its own validators and build its own bridges, and when something goes wrong there is no “shopping centre” to fall back on.

Why build your own chain

  • Performance: no competing with thousands of other apps for block space. A perpetuals exchange needs millisecond matching, which it cannot get on a crowded public chain.
  • Rules: it can change transaction ordering (for example to prevent front-running), let users skip gas, or build a matching engine into the chain.
  • Value capture: fees and MEV (extra profit from ordering) go to its own chain and token rather than to a public chain’s validators.

Common ways to build one

Approach Who secures the chain Examples
Cosmos SDK Its own validators dYdX, Injective
OP Stack / Arbitrum Orbit Settles to Ethereum (or another Layer 2) Many gaming chains and app-specific Layer 2s / Layer 3s
Avalanche L1 Its own validators, using Avalanche’s tooling Some gaming and institutional chains
Fully custom Its own validators and consensus Hyperliquid

The difference between the first two matters: with the Cosmos SDK or custom consensus, the chain’s security depends entirely on its own few dozen validators; a rollup built with the OP Stack or Orbit can at least rely on Ethereum to check that the books are right.

Four examples

Locked value on several app-chains
  1. HyperliquidChain built by a perpetuals exchange$1.08B
  2. dYdXPerpetuals exchange, Cosmos SDK$70.3M
  3. InjectiveFor financial apps, Cosmos SDK$14.4M
  4. RoninThe chain of the game Axie Infinity$11.4M

Data as of 16 Oct 2026, 01:45 UK time · source DefiLlama

Their token market caps
  1. Hyperliquid$18.28B
  2. dYdX$104M
  3. Injective$856.3M
  4. Ronin$49.2M

Data as of 16 Oct 2026, 01:45 UK time · source CoinGecko

  • Hyperliquid: a perpetuals exchange that wrote its own consensus and matching engine from scratch, with a fully on-chain order book running at speeds close to a centralised exchange. It is one of the most successful app-chains so far. But it has few validators, and in the JELLY incident of March 2025 the validators voted to delist a manipulated contract and settle it at a chosen price, sparking a debate over whether that counts as decentralised.
  • dYdX: originally a Layer 2 app on Ethereum, it moved in 2023 to its own chain built with the Cosmos SDK, where its own validators run the order book.
  • Injective: a finance-focused chain built with the Cosmos SDK, with an order book and other modules built in.
  • Ronin: the chain built by the developer of the game Axie Infinity. In March 2022 its bridge had only 9 validators; an attacker obtained the private keys of 5 of them and stole about $600 million, one of the largest crypto thefts in history.
how it worksHow the Ronin bridge was drained
  1. 01The bridge rule

    Withdrawing from the bridge needed signatures from at least 5 of 9 validators.

  2. 02Four keys taken

    Through social engineering the attacker compromised 4 validator nodes run by the same company.

  3. 03The fifth key

    A signing permission granted earlier to help with traffic had never been revoked, and the attacker used it.

  4. 04Straight withdrawal

    With 5 signatures the attacker "legitimately" withdrew about 174,000 ETH and 25.5 million USDC.

  5. 05Noticed six days later

    The team only discovered the bridge had been emptied when a user's withdrawal failed.

    About 6 days

The lesson: the fewer the validators, and the more they sit with the same organisation, the more fragile an app-chain and its bridge become.

Strengths and trade-offs

Strengths
  • Performance and rules can be tailored to the app, with an experience close to a centralised product.
  • Fees and ordering revenue go to the project, giving its token a more direct source of value.
Trade-offs
  • Usually only a few dozen validators, sometimes run by the project and its partners; decentralisation is low.
  • Assets move in and out through bridges, the most frequently attacked component.
  • The project and its validators can pause the chain, delist assets and change rules; when things go wrong, users lack the protections of a public chain.
Investigating a project on app-chain projects? Check these first
  1. 01

    How many validators are there, who runs them, and how much voting power do the top few control?

  2. 02

    Which bridge moves assets in and out, how many signatures release funds, and are the signers from the same organisation?

  3. 03

    Can the chain be paused? Has it ever halted or force-settled anything, and how was it handled?

  4. 04

    The token unlock schedule: team and investor allocations in app-chain projects are often large.

  5. 05

    Is the official block explorer the only one, and can third parties check the data independently?

Data and sources

  • Locked value: DefiLlama; market caps: CoinGecko (hover over a number to see its cut-off time).
  • The Ronin incident is based on Ronin’s official post-mortem; the Hyperliquid JELLY incident on Hyperliquid’s official announcement.
  • This guide is educational and is not investment advice.