Which Hyperliquid business are you measuring?
The first analytical choice is the perimeter. Perpetual trading, spot trading and other chain activity are related, but their volumes and fee flows are not interchangeable. A provider's combined protocol TVL can capture collateral beyond the specific perpetual markets in a volume table. Summing a parent protocol with its children can count the same activity twice.
For HYPE, a narrow perp-only analysis can be useful if it states what it excludes. A whole-platform thesis needs a reconciliation across products, not an unexplained larger number. That discipline also makes competitor comparisons more meaningful: a standalone exchange metric should not be ranked against an entire chain's economic activity without a visible scope adjustment.
23 September data review: include the builder books
Our original venue collector requested the default metaAndAssetCtxs response, which describes only the first perp dex. We expanded the 24-hour-volume collector to discover the current perpDexs registry and retrieve each book. This is a coverage correction, not organic growth in trading activity. Published open interest now uses the pinned DefiLlama perps overview instead of the venue aggregate because the latter followed a roughly double-sized measurement basis.
A dated venue capture at 08:22 UTC on 23 September contained 11 dex entries, including the native entry. Across 325 non-delisted markets, the collector found about $9.58 billion in rolling 24-hour volume. Its approximately $18.21 billion native open-interest aggregate is retained only as an audit observation; the comparable DefiLlama perps overview was around $9.06 billion later that day. Do not use the retired native OI in peer or FDV/OI comparisons. These are point captures, not a synchronized atomic chain state or daily closing observations.
All active books in this capture used the USDC collateral token index. We treat USDC units at face dollar value. Six registered deployments had no non-delisted markets in the captured response. The collector accepts a validated empty book, but rejects a failed response, duplicate market identity or unsupported active collateral rather than presenting a partial or mixed-currency total. Future non-USDC markets require an explicit conversion policy.
We also checked the DefiLlama perp adapter source. Its coverage includes HIP-3; its holder-revenue calculation allocates 99% of Hyperliquid revenue after builder/deployer shares, rather than multiplying all gross user fees by 99%. Spot is outside this row. Revenue and holder revenue reflect the same Assistance Fund allocation and must not be added. Priority-fee additions appear in descriptive labels but are commented out in the reviewed implementation.
Changing source scope starts a new comparable series. Historical native-only volume and retired native OI observations remain intact; weekly calculations refuse to compare them with the new basis. HIP-3 markets are not all RWA markets; an asset classification is a separate review from identifying the book that executes it.
Sources: Hyperliquid: perp dex discovery and asset contexts ↗ · Hyperliquid: collateral token index in spot metadata ↗ · DefiLlama: Hyperliquid perp adapter implementation ↗ · DefiLlama: open-interest overview, Hyperliquid Perps ID 5507 ↗
What the Assistance Fund does—and what it does not do
Hyperliquid's fee documentation assigns fees among HLP, the Assistance Fund and deployers. It describes automatic conversion of Assistance Fund fees into HYPE and burning the acquired tokens. This is a supply-reduction mechanism, not a direct cash dividend paid to every holder. Assigning all gross trading fees to HYPE purchases would therefore be incorrect.
Our interpretation is that fee-funded purchases can support a value-capture thesis, but their significance depends on execution and net supply. A model should count the funded purchase once and treat the resulting burn in the supply ledger. Adding purchase dollars to an independently monetized burn value would count the same mechanism twice.
28 September audit: holder-flow scope is clearer than the supply denominator
We resummed 90 complete UTC days ending 27 September 2026 from DefiLlama's Hyperliquid parent holder-revenue series. It recorded $142,293,990: $137,734,664 in the perps child and $4,559,326 in the spot-orderbook child, an exact parent-child reconciliation. The most recent 30 complete days recorded $55,163,589: $52,993,458 perps plus $2,170,131 spot. These are provider-classified token-flow observations, not audited profit or a cash dividend. The parent must not be added to its children, and revenue and holder-revenue labels must not be added as separate benefits when they cover the same flow.
At our 28 September capture, Hyperliquid's tokenDetails endpoint reported about 998.9 million total HYPE, 298.7 million circulating and 411.2 million future-emission units. Our contemporaneous CoinGecko feed reported about 955.3 million total and 222.4 million circulating. These are different supply conventions. The Assistance Fund balance is included in the official endpoint's accounting despite a documented burn mechanism. The two sources require a dated bridge before their circulating figures are compared; market cap from one source must not be paired with the other source's denominator.
Sources: DefiLlama Hyperliquid parent daily holder revenue ↗ · DefiLlama Hyperliquid Perps daily holder revenue ↗ · DefiLlama Hyperliquid Spot Orderbook daily holder revenue ↗ · Hyperliquid official tokenDetails API documentation ↗ · Hyperliquid: fees and Assistance Fund ↗ · CoinGecko HYPE market feed used for site price and FDV ↗
HIP-3 growth can change both market coverage and fee intensity
The documentation distinguishes fees and deployer arrangements across markets, including a lower-fee growth mode for eligible HIP-3 perps. That creates an important analytical possibility: activity can rise while revenue per dollar traded falls. The appropriate question becomes whether lower fees attract enough durable activity to expand the actual fee pool.
The current 24-hour-volume collector discovers native and builder-deployed perp books. A research claim about RWA expansion still needs an instrument-by-instrument classification and fee-economics check; HIP-3 is a market deployment mechanism, not a synonym for RWA. Open interest is published on the separate DefiLlama perps basis, so venue volume and published OI must not be combined into a valuation multiple without reconciling their market coverage.
Sources: Hyperliquid: fees and Assistance Fund ↗ · Hyperliquid: public information API ↗
Use open interest and volume to ask different questions
Open interest is outstanding exposure at a point in time; volume is activity over a period. A high-turnover venue can process large volumes without proportionate persistent exposure. Conversely, sticky open positions do not guarantee that those positions generate a high rate of fees. Neither observation alone establishes organic demand.
A useful research series compares consistent market sets over several weeks, alongside the effective fee rate and concentration. Our metrics preserve source time when supplied and otherwise show fetch time. One API response is not an audited time series. We also avoid treating futures open interest on the HYPE token as open interest across Hyperliquid's own trading venue.
Sources: Hyperliquid: public information API ↗ · DefiLlama: Hyperliquid metrics and product scope ↗
Strong contributors are evidence about execution, not valuation
Hyperliquid's own contributor page names Jeff and iliensinc as leaders of Hyperliquid Labs and says the organization is self-funded. Those disclosures help establish who is building the system and how the project describes its financing. They do not prove that the token is attractively priced or eliminate operational dependencies.
Our interpretation separates the team's ability to ship from the resilience of the market infrastructure. A useful review would examine behavior during congestion, liquidation stress, oracle disruption and sudden withdrawal demand. Past scale does not remove the need for those checks, and a biography score cannot substitute for them.
Sources: Hyperliquid: core contributors ↗
What a fair comparison with Aster or newer venues requires
Comparing market share is useful only with matching products and time windows. Comparing token economics additionally requires understanding who is eligible for the benefits. HYPE purchases and burns are different from an arrangement that distributes acquired tokens to time-locked stakers. A common label such as holder revenue does not erase that difference.
Our constructive thesis would require sustained, reconciled economic capture across an explicitly defined market set. It would weaken if apparent expansion mainly reflected perimeter changes, temporary fee incentives or an aggressive annualization window. The live profile and comparative articles keep these components separate so readers can challenge the thesis using the same inputs.
Sources: Hyperliquid: fees and Assistance Fund ↗ · DefiLlama: Hyperliquid metrics and product scope ↗
What we monitor next
- Reconcile fees and purchases across perps, spot and builder-deployed markets without parent/child double counting.
- Track fee intensity alongside volume under changing fee schedules.
- Reconcile the official tokenDetails and CoinGecko supply conventions with a dated bridge, without combining incompatible denominators.
Frequently asked questions
Are Assistance Fund burns a dividend?
No. The documented mechanism buys and burns HYPE; it is distinct from cash paid directly to holders.
Does the default API request cover every HIP-3 venue?
No. The default request covers the native perp dex. Our current 24-hour-volume collector explicitly discovers the additional builder books; published OI uses a pinned DefiLlama perps overview rather than the venue-native aggregate.
Does a large volume lead imply HYPE is cheap?
No. Price, dilution, normalized value capture and the valuation assumptions must be assessed separately.
Compare the evidence
- Hyperliquid metrics and sources →
- Aster metrics and sources →
- Extended metrics and sources →
- Paradex metrics and sources →
Source register
- Hyperliquid: fees and Assistance Fund · checked 2026-09-22
- Hyperliquid: public information API · checked 2026-09-22
- Hyperliquid: core contributors · checked 2026-09-22
- DefiLlama: Hyperliquid metrics and product scope · checked 2026-09-22
- Hyperliquid: perp dex discovery and asset contexts · checked 2026-09-23
- Hyperliquid: collateral token index in spot metadata · checked 2026-09-23
- DefiLlama: Hyperliquid perp adapter implementation · checked 2026-09-23
- DefiLlama: open-interest overview, Hyperliquid Perps ID 5507 · checked 2026-09-23
- DefiLlama Hyperliquid parent daily holder revenue · checked 2026-09-28
- DefiLlama Hyperliquid Perps daily holder revenue · checked 2026-09-28
- DefiLlama Hyperliquid Spot Orderbook daily holder revenue · checked 2026-09-28
- Hyperliquid official tokenDetails API documentation · checked 2026-09-28
- CoinGecko HYPE market feed used for site price and FDV · checked 2026-09-28
Research revisions
- 2026-09-22 — Initial evidence-based research edition.
- 2026-09-23 — Corrected native API coverage to include discovered HIP-3 books; documented captured totals, quote-asset checks and fee-adapter scope. Prior native-only data is retained and not compared as growth.
- 2026-09-23 — Reconciled venue-native open interest to the roughly half-sized DefiLlama comparison basis. Published OI now uses the pinned DefiLlama perps overview; the earlier native OI figure is retained only for audit.
- 2026-09-28 — Reconciled 30/90-day parent holder-flow rows to perps and spot children; identified the official tokenDetails versus CoinGecko supply-convention gap.
AI-assisted research checked against cited sources. Facts, assumptions and interpretation are distinguished; this is not a financial audit or a recommendation tailored to you. Editorial standards.
View current metrics →