What Is a Perp DEX? Points, Volume, and the Fight for Traders

• A perp DEX is an on-chain venue for perpetual futures, contracts that never expire and hold near spot through funding payments.
• a16z crypto puts 2025 decentralized perp volume at $6.7 trillion against $86.2 trillion on centralized exchanges, “roughly 7.8% of CEX volume, up from about 2.5% just a year earlier.”
• Hyperliquid sent 31 percent of HYPE supply, 310 million tokens, to points holders in November 2024, and every program since is judged against that payout (CoinDesk).
• CoinGecko’s State of Crypto Perpetuals Report 2026 puts perp DEXs at 13.5 percent of crypto open interest against 10 percent of volume in April 2026, so the two metrics rank venues differently.
• US persons stay shut out of most perp DEXs. The CFTC approved KalshiEX’s bitcoin perpetual on May 29, 2026, and the Senate rejected cloture on the CLARITY Act by 49 votes to 50 on September 15, 2026.

• Key Takeaways
• What Does Perp DEX Actually Mean in 2026?
• How Does a Perp DEX Work?
• How Do Perp DEX Points Programs Actually Work?
• Why Is Perp DEX Volume So Easy to Inflate?
• How Is a Perp DEX Different From a CEX Perpetuals Market?
• Who Is Building Perp DEXs in 2026?
• The Perp DEX Launch Playbook
• What Are the Challenges and Limitations?
• What Are the Regulatory Considerations?
• How Surgence Labs Does This
• Conclusion
• FAQs

A perp DEX is the on-chain venue for listing, matching, and settling perpetual futures for traders who keep custody of their margin, with funding payments replacing an expiry date. It is a derivatives venue, not a spot DEX with leverage bolted on. The price comes from an order book the venue runs or an outside feed it reads. Margin sits in a smart contract, on a shared chain or one the venue built. Funding moves between longs and shorts and drags the price toward spot. Depth pulls traders, traders pay fees, and fees plus token supply buy the next round of depth. Depth wins. Points only rent it.

Two things have changed since 2023. App-chain order books proved that on-chain matching can be as fast as matching on a centralized exchange, and points seasons became how venues buy their first traders.

Use liquidity, volume, and total value locked as synonyms, and you will overpay for the wrong venue. Liquidity is how much you can trade near the current price. Volume is what has already traded. Total value locked is deposits sitting still, so read what liquidity means in a token launch before signing a depth deal.

Traders post margin, open a long or short perpetual position, pay or receive funding each interval, and get liquidated if margin falls below the maintenance level. Every design argument in perpetual futures crypto is a variation on one of those four steps.

A perpetual future is a futures contract that never expires. Instead of settling on a date, it charges longs or shorts a funding payment that pulls its price toward spot. a16z crypto puts it in five words, “futures contracts that never expire.”

Whichever side is more crowded pays the other. If most traders bet the price will rise, the contract trades above the index, and those longs pay the shorts. Switch the crowd, and the payment switches with it. Hyperliquid’s docs call the funding rate “a mechanism that is used to ensure the price of the contract stays close to the underlying asset’s price.”

The quiet killer in perp trading is not the matching engine. It is how the venue closes losing positions and who covers the shortfall when the insurance fund is called on. Liquidations run against a mark price, not the last trade, so a thin book and a jumpy mark can close positions that were never losing money. One volatile hour costs a venue every trader its season paid for.

Three designs dominate, each buying something at a named cost. Read the CANNOT column first because vendors leave it out.

A useful heuristic. When the price comes from an oracle rather than a book, ask who loses the money if every trader bets the same way. On a pooled venue, it is whoever funded the pool.

Pick the design that fits the markets you will list. Changing it later means rebuilding the venue.

Perp DEX points are an off-chain ledger a venue keeps of your trading, deposits, and referrals, and it may later convert them into tokens at a ratio it sets and can revise. Points are not tokens, and tokens are not an airdrop. The airdrop is the later distribution event, and most decks blur all three.

Look at how points work from the venue’s side, not the farmer’s. Points buy order flow today, and you pay in token supply later, a liability on the cap table rather than a marketing line. Before deciding points are cheap, check what a crypto trader costs to acquire through paid channels.

The farmers grinding a season have the better view. Four months of work, and the payout does not cover the fees. The venue ran it anyway because a book with no depth attracts nobody.

Hyperliquid’s points campaign ran from November 2023 into late 2024, and at the token generation event on November 29, 2024, the venue sent 31 percent of HYPE supply to points holders.

The Hyperliquid airdrop reset what the category expects, so late programs now compete on the promise of a similar drop rather than on the product. That expectation warps airdrop economics generally. A promise that size with no product underneath buys farmers, not traders.

CoinGecko’s perpetuals report notes that Pacifica, Extended, and Variational each gained share in April 2026, reaching 4 percent, 4 percent, and 3 percent. All three run points programs. A fixed allocation spread over more points is a smaller allocation per point, and every new season raises what farmers expect.

Score the design, not the venue. Rate each criterion from one to three and read the total before committing budget.

A program rewarding raw volume buys wash trading. One rewarding open interest held over time, maker depth, or net maker and taker fees buys a trader. A season with no published sybil rules, the filters that catch fake accounts, is one you will spend a month arguing about, so read designing incentives that do not get farmed first.

A free perp DEX launch audit. Free 30-minute call. No pitch deck. No obligation.

Perp DEX volume counts both sides of every trade, costs almost nothing to generate when fees come back as points, and says nothing about whether a position stayed open. It is the easiest number in crypto to move and the one every ranking page leads with.

Which perp DEX has the highest volume? The answer changes month to month, and the question is the wrong one. Ask instead what happens when one venue’s season runs and another’s does not.

21Shares, writing in early October 2025, put Aster at “nearly 70%” of perp DEX volume while Hyperliquid’s share “slipped to about 10%.” Hyperliquid, meanwhile, carried $13.5 billion of open interest against Aster’s $3 billion. Aster had the volume. Hyperliquid had the positions.

Three numbers tell you more, and all three are harder to fake. Open interest shows what is at risk now. Maker and taker fees paid in money, not rebated in points, show what the venue earns. Traders still active thirty days after a snapshot show what the season bought.

Here is the live picture as of September 21, 2026, from DefiLlama, with positions from its open interest view.

Read the right-hand column before the left. For every dollar of open interest, ApeX trades about nine times more than Hyperliquid does, and a gap that far above the rest of the table is a question to ask, not a verdict to publish. The other finding is Hyperliquid’s share, about 35 percent of the $617.74 billion DefiLlama recorded for the category in those thirty days. A venue picked on headline volume alone hands you a book that empties the week its season ends.

A free perp DEX launch audit. Free 30-minute call. No pitch deck. No obligation.

Custody, price formation, verification, access, and liquidation transparency all differ, while depth still favors the largest centralized exchanges.

The scale gap is real, and the sources disagree on its size. a16z crypto puts 2025 centralized perp volume at $86.2 trillion against $6.7 trillion on decentralized venues, while CoinGecko records “$6.38T in trading volume” across the top twelve perp DEXs for 2025. They counted different venues, and a16z cites CoinGecko for that centralized figure.

Are perps better than futures? Neither is better. A perpetual suits a trader who wants continuous exposure without a quarterly roll, and a dated future suits a hedger who needs a fixed settlement date. A perp DEX wins on transparency long before it wins on depth, and venues that chase depth alone end up with neither.

The field falls into five groups. The first three are defined by how a venue forms a price, and the last two cut across them. Each carries a real limitation.

• App-chain order books. Hyperliquid, dYdX, and Aster each run their own chain, so orders, cancels, and liquidations settle in the venue’s own blocks. The limitation is governance, because owning the chain means owning the validator set and every dispute that reaches it.
• Layer 2 and zk order books. Lighter, Paradex, the edgeX perp DEX, and Extended settle to Ethereum while matching off-chain, and ApeX runs a similar model on zkLink. Those proofs cover settlement, not the price you got.
• Pool and oracle models. Jupiter Perps runs on Solana, and GMX on Arbitrum and Avalanche. Both price from oracle feeds and trade against pooled liquidity, so a market with no reliable price feed cannot be listed.
• Points-era entrants. Pacifica, Extended, and Variational arrived with incentives running, and none has traded a full year past a season.
• Forkable stacks. The dYdX v4 codebase is open source, and vendors resell it as a launch kit, which is why whole businesses selling clones rank on Google. A fork gives you software, not traders.

Pick the group that fits your listing plan, because building for two price models doubles the engineering work.

Hyperliquid is a perp DEX running on its own layer 1 blockchain, with what its docs call “fully onchain perpetual futures and spot order books” where “every order, cancel, trade, and liquidation happens transparently with one-block finality.” Nothing about the price you get is private, including when it goes against you.

CoinGecko records $190.28 billion of Hyperliquid volume in April 2026, about 3.9 percent of all perpetual volume including centralized venues. The table above puts it at roughly 35 percent of decentralized perp volume. A third of the category’s activity depends on one validator set and one team’s uptime.

Six windows decide whether a perp DEX still has a book ninety days after its season ends.

• T minus 90 days. Pick the design and the chain, then contract depth. Contracting market makers to a written minimum inside named price bands is the step teams postpone, and postponing it means launching into a book that looks empty on day one.
• T minus 30 days. Design the points program against the scorecard above and publish conversion rules you can keep. A rule revised mid-season costs more trust than the allocation is worth.
• T minus 0. Verify depth through your own monitoring, not the market maker’s report, and launch creator-led onboarding the same week.
• First 7 days. Watch open interest, fee revenue, and liquidation behavior daily. Kill wash trading patterns before an integration partner finds them, because that is the first thing partners check.
• First 30 days. Read retention, not volume. If the number of active traders falls while volume holds, the points are doing the work.
• Season end. Plan the post-snapshot week before the season starts and decide now whether a token launch follows or the program continues. This is when most of the volume disappears.

The last window is the one teams skip, which is why so many of them run the whole playbook twice.

Perp DEXs have closed most of the gap with centralized venues on speed, fees, and the number of markets listed. What is left is built into the model, and no budget fixes it.

The most expensive proof is recent. On April 1, 2026, attackers drained Drift Protocol, then the largest DeFi protocol on Solana, of $285 million, “wiping out more than 50% of its total value locked,” according to Chainalysis.

The mechanism was not a pricing bug. The attackers spent months building trust, then got real members of the security council to sign pre-approved administrative transactions. Nobody checks who can approve a transaction when comparing venues.

• Mercenary liquidity. Depth bought with points leaves when the points stop, faster than it arrived.
• Oracle and liquidation risk under stress. An insurance fund sized for a calm quarter is not sized for the hour that matters.
• Geography. Most venues geoblock US persons, so a real share of the market cannot legally trade there.
• Governance concentration. On an app-chain, the venue is also the validator set, so every dispute is its own to settle.

• Paying for volume instead of open interest
• Publishing a points table you later change
• Launching without contracted depth
• Running one season with no retention plan
• Treating total value locked as liquidity
• Marketing leverage into jurisdictions you geoblock

Those mistakes are cheap to avoid before launch and expensive to correct afterward, which is why you plan backward from launch day.

In the United States, true perpetual futures reach US persons only through a few CFTC-registered routes, and most perp DEXs block US users outright. None of this is investment advice, perpetual futures carry liquidation risk, and availability depends on where you are.

Three recent actions set the position. On May 29, 2026, the CFTC approved KalshiEX’s BTCPERP contract, a cash-settled bitcoin contract with, in the Commission’s words, “an indefinite (or ‘perpetual’) term.” CME then sued the CFTC over that approval on June 18, 2026, arguing that perpetual contracts are swaps rather than futures, and the case was still pending in late September 2026. On the same day, CFTC Letter 26-17 let Coinbase Financial Markets route US customers into Deribit’s perpetuals through Coinbase’s own Bermuda-based broker affiliate. That is offshore execution through a registered intermediary, not access to a US-listed contract.

Then the legislation stalled. On September 15, 2026, the Senate rejected cloture on the CLARITY Act by 49 votes to 50, against the sixty required, though a motion to reconsider leaves another vote open. Build so the plan works either way.

Hyperliquid in the US has a clear answer, and the venue gives it in its terms of service. Clause 1.6 classifies anyone residing, located, or incorporated in the United States or Ontario, Canada, as a restricted person, “strictly prohibited from accessing or using the Interface.”

Europe judges a product by what it does, not what you call it. ESMA’s guidelines on qualifying crypto-assets as financial instruments state that “perpetual futures should be treated as derivative contracts” assessed against MiFID II, and that a crypto-asset meeting that test “will be subject to MiFID II, not MiCA.” ESMA added on February 24, 2026, that a firm’s commercial name “is irrelevant” to whether national CFD intervention measures apply, and that reading pulls leverage caps and risk warnings into scope.

Promoting points or leverage into a country you geoblock is a marketing problem with legal consequences, not a legal problem with marketing side effects, and compliance-safe crypto marketing is cheaper built in than retrofitted. Get the geography right, and the program scales. Get it wrong, and your best channel becomes your largest liability.

We run liquidity-first go-to-market for perp DEXs. That means trader acquisition, creator-led onboarding, incentive design that survives a post-season audit, and cross-chain positioning.

For StandX, that work supported “$200 million+ peak TVL,” “$2.038 billion+ cumulative perps volume,” “$1.947 billion+ 30-day perps volume,” and “20,000+ registered traders” across BNB Chain and Solana, as published in the StandX perps launch case study.

A free perp DEX launch audit. Free 30-minute call. No pitch deck. No obligation.

We have worked with Binance, Animoca Brands, Polygon, Falcon Finance, Incentiv, Rails, Biptap, SendIt, Black Mirror, and Boost.

The fight for traders is won on depth and retention, not on points. Points buy one season. What keeps a book alive after that is real depth, a liquidation engine that holds under stress, and a fee model that still works once the rebates stop. All three take longer to build than any program takes to announce.

The durable version of a perp DEX looks boring in month thirteen. Open interest holds, fees arrive as money rather than rebates, and the traders who came for the allocation stay because the venue is the best place to open the trade. Build toward that month and the season becomes the start of the business, not all of it.

A perp DEX, also written perps DEX, is a decentralized exchange built for perpetual futures, contracts with no expiry that track an asset through funding payments. Traders keep custody of their margin, and positions settle on-chain or on a blockchain the venue runs. It is a derivatives venue, not a spot exchange.

Perps are perpetual futures, leveraged contracts that let a trader go long or short an asset with no expiry date. A funding rate paid between longs and shorts keeps the contract price close to spot. They trade on both centralized exchanges and perp DEXs, though custody and price formation differ.

Perpetual futures are futures contracts that never expire. Instead of settling on a fixed date, they charge a periodic funding payment that pulls the contract price toward the underlying index. That choice allows a position to stay open indefinitely, and it is why funding replaces settlement in every perpetual market.

A trader posts margin, opens a leveraged long or short, and pays or receives funding each interval depending on which side is crowded. If margin falls below the maintenance level, the position is liquidated against a mark price. Perp DEXs run that loop with margin held on-chain, not at an exchange.

The answer changes month to month, and points seasons swing it fast. In early October 2025, 21Shares put Aster near 70 percent of perp DEX volume while Hyperliquid, at about 10 percent, held more than four times Aster’s open interest. Check DefiLlama for the current ranking and compare open interest, not volume alone.

Neither is better, and they suit different needs. Perpetuals fit traders who want continuous exposure without rolling a contract each quarter, since there is no expiry to manage. Dated futures fit hedgers who want a fixed settlement date, a known cost, and no funding payments to model over a long holding period.

Yes. Leverage magnifies losses, funding can run against a position for days, and a liquidation can close an account within minutes of a sharp move. On a perp DEX, a trader also carries smart contract risk, oracle risk, and the operational risk of the venue, as recent incidents have shown.

Perp DEX points are a score a venue keeps off-chain for trading, deposits, and referrals. The venue may later convert that score into tokens at a ratio it sets and can revise, which is why points are a promise, not an asset. Only the venue decides what a point is worth.

Only through a small number of CFTC-registered routes. The CFTC approved KalshiEX’s bitcoin perpetual on May 29, 2026, and gave Coinbase Financial Markets relief to route customers into Deribit’s perpetuals through a Bermuda affiliate. Most perp DEXs still block US persons, so check each venue’s own terms carefully before trading.

No. Hyperliquid in the US is blocked at the interface, and its terms classify anyone residing, located, or incorporated in the United States or Ontario, Canada, as a restricted person, “strictly prohibited from accessing or using the Interface.” A US trader wanting perpetual exposure needs a CFTC-registered venue, not a workaround.

Hyperliquid is a perp DEX running on its own layer 1 blockchain, with perpetual futures and spot order books held fully on-chain. Every order, cancel, trade, and liquidation settles in the venue’s own blocks. It sent 31 percent of its HYPE token supply to points holders at its November 2024 token generation event.

Surgence Labs runs liquidity-first go-to-market for perp DEXs, covering trader acquisition, creator-led onboarding, incentive design, and cross-chain positioning. For StandX, that work supported $200 million+ peak TVL and $2.038 billion+ cumulative perps volume across BNB Chain and Solana. The starting point is a free launch audit and a 30-minute call.

• Crypto Exchange Marketing Playbook
• Trading Competition Design for Exchanges and Perp DEXs
• Crypto Market Makers Explained
• How to Launch a Token