How Onchain Perp DEX Volume Is Measured
Onchain perpetuals volume is one of the largest numbers in crypto, and much of it is notional, not cash changing hands. Here is how the figure is built and where it misleads.
Onchain perpetual DEX volume is notional volume: contract size multiplied by price, summed per matched trade. It measures the face value of positions changing hands on decentralized exchanges (exchanges that run on public blockchains rather than being operated by a single company), not the cash traders actually put up. A trader using leverage generates many times their own capital in reported volume, so the headline figure measures contract turnover, not money at risk.
That gap is the single most important thing to understand about this metric. Once you know that most of the number is leverage-inflated notional, the rest of the picture falls into place quickly.
Key takeaways
- Onchain perp DEX volume is notional volume: contract size times price, magnified by leverage, not deposited capital.
- A perpetual future is a derivative with no expiry date, kept tethered to the spot price by a recurring payment between longs and shorts called the funding rate.
- A round trip generates volume on both entry and exit, so volume should not be read as unique capital deployed.
- Because these venues settle onchain, the trades, funding payments and open interest are publicly verifiable, unlike a centralized exchange's internal ledger.
- To compare volume across venues you have to normalize for leverage conventions, how each protocol logs a fill, and whether maker and taker sides are both counted.
A worked example
Suppose a trader posts $10,000 of margin and opens a position at 20x leverage. That is a $200,000 position (contract size times price). Opening it records $200,000 of traded notional. When the trader later closes it, that generates roughly another $200,000. One round trip on $10,000 of real capital produces about $400,000 of gross traded notional.
| Trader capital (margin) | Leverage | Position size (notional) | Reported volume for one open + one close |
|---|---|---|---|
| $1,000 | 1x | $1,000 | $2,000 |
| $1,000 | 5x | $5,000 | $10,000 |
| $1,000 | 20x | $20,000 | $40,000 |
| $1,000 | 50x | $50,000 | $100,000 |
The economic event is identical in every row: one person risked $1,000. But reported volume ranges widely depending only on the leverage setting and the fact that opening and closing each count as a trade. This is why comparing a perp DEX's daily volume to a spot exchange's, or to a stock market's, is misleading unless you adjust for leverage and turnover. A number that peaks during volatile sessions is often measuring liquidations and rapid re-entry, not fresh capital arriving. Allium's guide to reading onchain volume without getting fooled walks through the double-count and wash-trading traps that inflate raw figures.
Why a perpetual is the contract being counted
A regular futures contract is an agreement to buy or sell an asset at a set price on a set date. A perpetual future strips out the date. It never expires, so a trader can hold a leveraged bet on Bitcoin's price indefinitely. The problem that creates is drift: with no expiry to force the contract price back toward the real spot price, the two can wander apart. The fix is the funding rate, a small payment exchanged directly between traders holding long positions and traders holding short positions, typically every hour or every eight hours. When the perpetual trades above spot, longs pay shorts, which nudges the price back down. It works like a tax that always pushes the crowded side to pay the other side.
The perpetual, first popularized on centralized venues, became the dominant crypto derivative because it lets a trader express a leveraged view without ever taking delivery or rolling an expiring contract. When those same contracts moved onchain, every fill, every funding payment and every liquidation became a public transaction. That is what onchain perp DEX volume is counting: the settled record of that activity on the blockchain itself.
Five metrics, five different questions
Volume is one of several numbers a perp venue reports, and each answers a different question. Confusing them is the most common way to misread a venue.
| Metric | What it measures | What it cannot tell you |
|---|---|---|
| Volume | Total notional traded over a period (flow) | How much unique capital is committed |
| Open interest | Total notional of positions still open at a moment (stock) | How actively those positions are being traded |
| Collateral (TVL) | Deposited capital backing positions | The leveraged size of exposure on top of it |
| Liquidations | Positions forcibly closed when margin runs out | Whether traders re-entered afterward |
| Funding | Payments between longs and shorts to hold price near spot | Direct trade volume (funding is separate) |
Reading volume next to open interest is the difference between a headline and an accurate one. A day where volume jumps but open interest is flat usually means the same positions are being churned, not that new money committed.
How a matched trade is actually counted
The measurement question has a precise answer once you look at how records are structured. In Allium's Hyperliquid perp trade table (hyperliquid.dex.trades), one row equals one matched trade. Over the seven days to 2026-09-23, 44,010,086 rows carried 44,010,086 distinct trade_ids, and each row carries BOTH the buyer_address and the seller_address. A matched trade is therefore counted once. The maker and taker are not logged as two separate trades, which is the single most common source of inflated cross-venue comparisons.
USD notional comes from the usd_amount column on that same row, alongside the raw amount and the price. Direction is explicit per side through buyer_dir and seller_dir, so opens and closes are distinguishable rather than inferred. That distinction is what lets you see why notional overstates capital deployed.
Across the window, Hyperliquid recorded $66.78B of traded notional across 44,010,086 trades (Allium data, 7 days to 2026-09-23). The two largest of eight direction labels, which together cover 39.9M of the 44.0M trades:
| Direction | Trades | Notional |
|---|---|---|
| Open Long | 19,930,595 | $31.47B |
| Close Short | 19,991,050 | $31.18B |
Opens and closes are near-symmetric. A round trip contributes on both entry and exit, so notional measures trading activity, not unique capital at risk. The same direction field also distinguishes cross-margin from isolated-margin positions: "Liquidated Cross Long" and "Liquidated Isolated Long" appear as separate directions rather than being lumped together.
One venue, many deployments
Hyperliquid's HIP-3 framework lets builders deploy their own perp markets on the same infrastructure, and each shows up as a distinct venue in the same schema.
The same schema separates the core order book from builder-deployed (HIP-3) venues, so activity on each resolves to the same fields. We are not publishing comparative volumes for individual builder-deployed venues: a single seven-day window is too short a base to size a venue fairly.
This is Hyperliquid and its HIP-3 deployments only. It is not a cross-protocol comparison with dYdX, GMX or Drift, and it should not be read as covering all perp DEXs. It does show how much detail a normalized schema exposes: builder-deployed venues that sit on shared infrastructure but trade at very different scales.
What liquidations look like in the data
Because rows tag the liquidated_user when a position is forcibly closed, the two liquidation paths separate cleanly. Over the window (Allium data, 7 days to 2026-09-23), market liquidations totaled 120,351 trades and $546.0M of notional. The backstop path, the venue's last-resort mechanism, fired just 20 times for $635.5K across 20 users. The backstop is rare by design: it only engages when a position cannot be closed into the open market, which is why it appears 20 times against 120,351 ordinary liquidations. For how forced selling can compound, see Allium's piece on liquidation cascades.
Why counting it consistently across venues is hard
Every perp DEX records trades differently. One protocol emits a single event per fill; another logs maker and taker legs separately, which double-counts if you sum naively. Leverage caps, funding intervals, the denominating asset and how a partial liquidation is logged all vary. To compare perpetuals activity across venues, the same event has to resolve to the same fields: venue, market, side, position size, entry and exit price, funding paid or received, USD notional and event type. Those fields sit in different places, under different names, in each protocol's event logs. Allium's perpetuals datasets normalize those records into a shared schema, including venue-level detail such as funding rates, and a live Hyperliquid view built on the same tables. Allium standardizes this activity; it does not operate a venue or provide trading advice.
What is still unsettled
The regulatory status of leveraged perpetuals offered to retail users is genuinely open in several jurisdictions, and settling onchain does not resolve it. How derivatives rules apply to a protocol with no single operator is contested, as is how index and settlement prices for onchain perps should be sourced and audited. None of this is settled law, and none of it should be read as investment advice.
Data and methodology last reviewed: September 23, 2026.
Interested in learning more about Allium's onchain data infrastructure? Speak to someone on the team.
Frequently asked questions
Is onchain perp DEX volume the same as the amount of money traded?
No. The headline figure is notional volume: contract size times price, summed per matched trade. It is the face value of leveraged positions, not the cash traders deposited. A trader using 20x leverage generates 20 times their capital in notional, and a round trip counts on both entry and exit, so reported volume can be many multiples of the money actually at risk.
Does perp DEX volume double-count maker and taker?
In a well-structured dataset, no. In Allium's Hyperliquid table, one row equals one matched trade carrying both the buyer and seller address, so each trade is counted once. Over the seven days to 2026-09-23, 44,010,086 rows carried 44,010,086 distinct trade_ids. Naive summing of separately logged maker and taker legs is a common source of inflated numbers elsewhere.
What is the difference between volume and open interest on a perp DEX?
Volume is a flow: total notional traded over a period. Open interest is a stock: total notional of positions still open at a moment in time. Volume can spike from churning the same positions, while open interest shows how much exposure is genuinely committed. Reading both together is more reliable than volume alone.
How is USD notional derived for a perp trade?
From the notional value of each matched trade: contract size multiplied by price. In Allium's Hyperliquid data this sits in the usd_amount column on the same row as the raw amount and price, so USD value is read directly rather than reconstructed after the fact.
How are opens and closes distinguished in the data?
Through explicit direction fields per side. In Allium's Hyperliquid table, buyer_dir and seller_dir record whether each side is opening or closing (for example Open Long or Close Short), so entries and exits are distinguishable rather than inferred. Over the seven days to 2026-09-23, opens and closes were near-symmetric, which is why notional measures activity rather than unique capital.
Does this data cover all perp DEXs?
No. The figures here are Hyperliquid and its HIP-3 builder-deployed venues only. They do not cover dYdX, GMX, Drift or the wider perp DEX market, and they should not be read as a cross-protocol total. They also cover a single seven-day window and should not be annualized.
What is the funding rate and does it affect volume figures?
The funding rate is a recurring payment between long and short holders that keeps a perpetual's price near spot. It does not add to trade volume directly, but sustained funding costs push traders to close and reopen positions, which does generate volume. It is a separate metric worth tracking alongside volume and open interest.
Allium provides onchain data infrastructure. Companies named in this article may be Allium customers, prospects or commercial counterparties. This article is informational only and is not investment, legal or tax advice. Data and information last reviewed: September 23, 2026.
