Hyperliquid spent 2026 setting activity records while the earnings stream that funds its token buyback moved the other way. The crossover moment came in July, when tokenized real-world assets — not bitcoin — became the platform's largest market.
Did RWA volume really overtake bitcoin on Hyperliquid?
Yes, on both measures that matter. In July 2026, real-world-asset (RWA) perpetual volume on Hyperliquid reached 99.2% of bitcoin perp volume, and RWA open interest set a record $3.6 billion — passing bitcoin to become the platform's largest market by open interest .
Quick Answer: Yes. Between July 13–19, 2026, RWA perps were roughly 52% of Hyperliquid's weekly volume (~$25 billion), the first week tokenized assets led all categories. Yet gross protocol revenue fell from about $357 million in Q3 2025 to $201.83 million in Q2 2026 — a ~43% decline.
The weekly data confirms it was not a single-day artifact: between July 13 and July 19, RWAs accounted for about 52% of weekly volume, roughly $25 billion, the first time tokenized assets were Hyperliquid's largest trading category . Quarterly RWA share climbed from 1.8% in Q4 2025 to 20.7% in Q1 2026 and 32.2% in Q2 2026 .
The complication sits on the revenue line. Across the same stretch, gross protocol revenue declined for four consecutive quarters — from roughly $357 million in Q3 2025 to $295 million, then $217.47 million, then $201.83 million in Q2 2026 . Record activity, shrinking take. The rest of this analysis explains why the two lines diverged, and what it means for HYPE.
HIP-3 in plain terms: how builder-deployed perps work
HIP-3 is Hyperliquid's permissionless framework for builder-deployed perpetual futures: any qualified deployer can launch its own perp market, defining the asset, the oracle, the contract specifications, the leverage limits and the settlement logic, while inheriting HyperCore's existing order books and margining engine . That division of labor is why tokenized equities, commodities and pre-IPO contracts appeared so quickly — the hard part, matching and risk, already existed. According to Hyperliquid's HIP-3 documentation, deployers take on responsibilities that a centralized venue would normally keep in-house.
The gate is capital, not permission. Mainnet deployers must maintain 500,000 staked HYPE — roughly $28 million at recent prices — and that stake stays locked for 30 days after all of a deployer's markets are halted, with validator slashing available as a penalty for misbehavior . For HYPE, this is a genuine demand sink: every serious deployer removes eight figures of tokens from liquid supply for as long as it operates, plus a month.
The economics run the other way. Hyperliquid's official fee page states that spot and HIP-3 perp deployers may keep up to 50% of the trading fees generated by their assets . On top of that, when HIP-3 growth mode is active, protocol fees, rebates, volume contributions and L1 user rate-limit contributions are all reduced by 90% .
Stacked together, HIP-3 creates several legitimate channels through which a traded dollar can be diverted before it reaches the Assistance Fund that funds the HYPE buyback:
- Deployer share — up to 50% of fees on that deployer's markets .
- Growth mode — a 90% reduction in protocol fees and related contributions while active .
- Builder codes — front-end and routing partners taking a cut, essentially a pass-through line.
- Maker rebates and aligned-collateral discounts — paid out to liquidity providers and to traders using preferred collateral.
None of this is leakage in a pejorative sense; it is the incentive budget that bought the volume. But it explains the mechanical point behind the divergence: HIP-3 raises volume and locks HYPE, while simultaneously lowering the share of each traded dollar that converts into tokenholder buyback. Coinbase Institutional described the same pattern on March 5, 2026, noting that take-rates were rising while buyback-per-fee conversion fell .
The RWA growth curve, quarter by quarter
Tokenized-asset perps went from a rounding error to the largest category on Hyperliquid in three quarters. Real-world assets accounted for 1.8% of platform volume in Q4 2025, 20.7% in Q1 2026 and 32.2% in Q2 2026 — the latter worth $213 billion in a single quarter . By the week of July 13–19, 2026, RWAs were roughly 52% of weekly volume, about $25 billion, the first week tokenized assets outranked every crypto pair on the venue .
The regulatory filing gives the cleanest cumulative figure. Hyperliquid Strategies Inc's SEC S-1/A, filed June 16, 2026, states that HIP-3 launched in November 2025 and had driven more than $319 billion of cumulative volume since launch, representing approximately 37.5% of total Hyperliquid trading volume as of June 15, 2026 . That figure predates the July surge, so the mid-July weekly reading of ~52% is the leading edge of the same curve rather than a separate phenomenon.
| Period | RWA share of Hyperliquid volume | Notable marker |
|---|---|---|
| Q4 2025 | 1.8% | HIP-3 launched November 2025 |
| Q1 2026 | 20.7% | Equities and commodities markets scale up |
| Q2 2026 | 32.2% | $213B RWA volume; HIP-3 ~37.5% of all volume as of June 15 |
| Week of Jul 13–19, 2026 | ~52% | ~$25B weekly; RWAs become largest category |
Participation widened alongside the volume. RWA holders rose 56% in a month to 1.6 million, and RWA open interest set a record $3.6 billion in July, overtaking bitcoin as the platform's largest market by open interest while RWA perp volume reached 99.2% of bitcoin perp volume . Holder growth outpacing volume growth is the more informative of the two: it points to new accounts arriving for the product rather than existing desks rotating size between tickers.
The product set explains the pull. Live markets span crude oil, gold and silver, S&P 500 and Nasdaq-100 trackers, single names including Nvidia and Tesla, and pre-IPO exposure such as SpaceX . Three structural properties do the work:
- Stablecoin settlement. Contracts settle in stablecoins, so a trader never touches a brokerage account or a fiat rail to take equity-index exposure .
- No expiry. Perpetual structure removes roll mechanics that deter retail users in traditional futures.
- Weekend and overnight access. Markets trade continuously while equity venues are closed; leveraged Nvidia exposure at 2 a.m. on a Sunday has few alternatives .
That third property is the one competitors will find hardest to copy quickly, and it is also why the growth curve is unlikely to reverse on its own. The open question is not whether the volume is real — the filings and third-party trackers agree it is — but which quarter's mix carries forward, given that the categories growing fastest are the ones subject to deployer fee shares and growth-mode reductions.
Where the fees actually go: the buyback breakdown
Hyperliquid's fee stream is no longer a single pipe into the HYPE buyback — it is a splitter with five outlets, and the newest outlet is growing fastest. In Q2 2026, DeFiLlama's decomposition of protocol revenue shows $148.64 million routed to token buyback/tokenholder net income, $16.38 million to builder code distribution, $8.80 million to maker rebates, $7.49 million to HIP-3 deployer distribution and $1.09 million to Unit revenue . The buyback still takes the largest share by a wide margin, but the direction of travel across categories is what matters for anyone modelling HYPE.
Compare that with the prior quarter. Q1 2026 produced $165.34 million of buyback, $17.92 million of builder codes, $10.40 million of maker rebates — and just $1.10 million of HIP-3 deployer distribution . The deployer line rose almost sevenfold quarter over quarter, from roughly $1.1 million to roughly $7.5 million, while the buyback line fell about $16.7 million over the same span. Neither number is large relative to total revenue in isolation. The relationship between them is the signal: the category tied to the fastest-growing volume is the category taking a bigger cut before the Assistance Fund sees a dollar.
| Revenue category | Q1 2026 | Q2 2026 | Change |
|---|---|---|---|
| Token buyback (tokenholder net income) | $165.34M | $148.64M | −10.1% |
| Builder code distribution | $17.92M | $16.38M | −8.6% |
| Maker rebates | $10.40M | $8.80M | −15.4% |
| HIP-3 deployer distribution | $1.10M | $7.49M | +581% |
| Unit revenue | — | $1.09M | — |
| Gross protocol revenue | $217.47M | $201.83M | −7.2% |
The broader payout trend confirms it. Cost of revenue — the fees returned to builders, market makers and the liquidity vault — rose from under 6% of gross revenue in Q2 2025 to 18% a year later, with builder code fees functioning as essentially a full pass-through . A protocol keeping 94 cents of every gross revenue dollar and a protocol keeping 82 cents are different businesses from a tokenholder's seat, even if headline volume triples in between.
What has not changed is the mechanism itself. Hyperliquid Strategies Inc's SEC S-1/A states that 99% of all fees are allocated to the Assistance Fund, up from 97% after an August 26, 2025 announcement, and that 45.23 million HYPE had been acquired and permanently removed from circulation as of June 15, 2026 — equal to 4.5% of initial supply . The routing rule is intact and was tightened, not loosened. The pool it draws from is what has been diluted, because deployer shares, builder codes, rebates, aligned-collateral discounts and growth-mode reductions all apply upstream of the 99% split. Hyperliquid's own fee documentation confirms the size of those upstream claims: spot and HIP-3 perp deployers may keep up to 50% of trading fees generated by their assets, and when growth mode is active, protocol fees, rebates, volume contributions and L1 rate-limit contributions are reduced by 90% .
Coinbase Institutional described this gap before it showed up cleanly in the quarterly splits. In a March 5, 2026 research note, the desk observed that Hyperliquid's economics look equity-like because eligible fees route to buybacks, but that take-rates were rising while buyback-per-fee conversion was falling, with rebates, builder and deployer economics and incentive structures intercepting the stream ahead of the Assistance Fund .
One caveat belongs on the record before any of this is treated as settled attribution. No public decomposition exists that splits the buyback shortfall by individual market, growth-mode status, deployer share or maker/taker mix, and a softer overall crypto tape during the period is a concurrent contributor rather than a controlled-for variable . What the data supports is a directional claim — fee capture per unit of volume is falling for tokenholders — not a precise allocation of blame across the venues driving it.
Base case: what happens if the current mix holds
The base case for HYPE is not falling revenue — it is flat-to-rising gross fees paired with a shrinking tokenholder share of each traded dollar. If HIP-3 keeps compounding at its current pace, gross fees can grow while buyback dollars stagnate, because the fastest-growing segment is the one where deployer shares of up to 50%, growth-mode reductions of 90% on protocol fees and rebates, and builder codes intercept flow before it reaches the Assistance Fund . That is a mix problem, not a demand problem.
Coinbase Institutional described the mechanism before the RWA surge was obvious. In a March 5, 2026 note, its analysts observed that Hyperliquid's take-rates were rising while buyback-per-fee conversion was falling, with rebates, builder and deployer economics and incentive structures diverting the stream ahead of the Assistance Fund .
"Hyperliquid's economics look equity-like because eligible fees route to buybacks — but the conversion from fees to buyback is weakening as rebates, builder and deployer economics sit in between," — Coinbase Institutional research team (source: Coinbase Institutional, 2026-03).
The supply side runs the other way, and that is what makes the base case defensive rather than neutral. Monthly buybacks have been running at roughly $53–83 million, while about 9.92 million HYPE vests monthly to core contributors and a further ~10 million HYPE — near $550 million at recent prices — unlocked on August 6, against circulating supply of about 222 million . Buyback and emission move in opposite directions: the sink shrinks with earnings while the tap runs on a schedule that ignores them.
Priced against that, HYPE near $55 — about 28% below its June 16 record of roughly $77 — trades at approximately 16x circulating market cap and 70x fully diluted on around $785 million of annualized earnings . The base case is therefore a token whose fundamental story keeps improving on volume, market share and staking demand, while the specific line item most retail holders underwrite — net supply removed per quarter — keeps thinning.
Bull and bear case for HYPE from here
The bull and bear cases for HYPE both start from the same fact set and diverge on whether fee dilution is a transition cost or a permanent feature. The bull case rests on three measurable supports: HIP-3's staking sink, a structurally larger addressable market, and gross fees that are still expanding. Each mainnet HIP-3 deployer must maintain 500,000 staked HYPE — roughly $28 million at recent prices — locked for 30 days after all of that deployer's markets halt, and exposed to validator slashing . That is hard, non-speculative demand that scales with the number of venues, not with sentiment. Meanwhile DeFiLlama listed annualized fees of $1.056 billion against annualized revenue of $808.46 million at crawl time — the top line is not collapsing, it is being shared differently. Equities, commodities and pre-IPO exposure are a far deeper pool than crypto-native perps, and a venue that captures even a diluted share of that pool can out-earn one that captures a pure share of a smaller one.
The bear case is about concentration and velocity. Trade.xyz accounts for more than 90% of all HIP-3 open interest, which means the headline growth curve is effectively one deployer's risk parameters . That fragility is not theoretical: a single trade on a thin Korean pre-market venue moved Trade.xyz's SK Hynix contract 19%, triggering liquidations the firm agreed to reimburse . Near-term momentum also reversed hard: perp volume fell from $84 billion in early July to $43 billion within three weeks, total open interest slipped from $75 billion to $65 billion, and weekly revenue dropped from a $23 million average to $7.5 million . Spot HYPE ETFs logged their first weekly outflow, roughly $7 million in the week to July 17, ending a nine-week inflow streak .
| Factor | Bull reading | Bear reading |
|---|---|---|
| HIP-3 deployer stake (500K HYPE, ~$28M) | Hard demand sink that scales with venue count | Only two-plus meaningful venues today; Ventuals already sunset |
| Gross fees ($1.056B annualized) | Top line still expanding | Buyback fell to $148.64M in Q2 2026 from ~$290M in Q3 2025 |
| RWA addressable market | Equities, commodities, pre-IPO dwarf crypto perps | >90% of HIP-3 open interest sits with Trade.xyz |
| Near-term flow | Record 6.63% share of monthly perp volume vs CEXs | Volume $84B → $43B in three weeks; first ETF outflow (~$7M) |
Price already reflects part of the tension. HYPE traded near $55, about 28% below its June 16 record of roughly $77, on approximately 16x circulating market cap and 70x fully diluted against around $785 million of annualized earnings . The bull case does not require new highs in volume; it requires the deployer-share and growth-mode lines to stabilize while gross fees keep climbing. The bear case does not require the RWA thesis to fail; it only requires the fastest-growing markets to stay the least tokenholder-accretive ones while roughly 9.92 million HYPE vests monthly to core contributors . Broader DeFi flows are also rotating between venues rather than expanding uniformly (video: 3.0 TV).
Portfolio implication: how to read this if you hold or trade HYPE
The correct framing for position sizing is dilution of tokenholder capture per unit of volume, not a collapse in fee generation. Hyperliquid still produced $201.83 million of gross protocol revenue in Q2 2026, and DeFiLlama listed annualized fees of $1.056 billion against $808.46 million of annualized revenue. What changed is the split: token buyback net income fell to $148.64 million in Q2 2026 from $165.34 million in Q1 while the HIP-3 deployer line rose from $1.10 million to $7.49 million. That distinction changes how much you own, not whether the asset is investable.
The practical monitoring list is short:
- Quarterly category breakdown, not headline volume. DeFiLlama separates buyback, builder codes, maker rebates, Unit revenue and HIP-3 deployer distribution . If deployer distribution keeps growing faster than buyback, volume records are not the metric that matters.
- Deployer concentration. Trade.xyz holds more than 90% of all HIP-3 open interest, and HIP-3 assigns oracle, leverage and settlement design to the deployer . One venue's parameter choice is a platform-level event.
- Supply calendar. Roughly 10 million HYPE (~$550 million) unlocked on August 6 against monthly buybacks of $53–83 million — the arithmetic of that ratio is the near-term risk, not sentiment.
Finally, treat attribution as unfinished work. No official venue-by-venue decomposition of the buyback shortfall by growth-mode status, deployer share or maker/taker mix has been published, and broader crypto-volume weakness is a concurrent contributor . The concrete takeaway: size HYPE against buyback dollars per quarter, not volume records, and re-underwrite the position each time DeFiLlama's deployer line grows faster than the buyback line.
Frequently asked questions
Why did Hyperliquid's revenue fall while trading volume hit records?
Because volume and tokenholder revenue are no longer the same trade. Hyperliquid's fee page states that when HIP-3 growth mode is active, protocol fees, rebates, volume contributions and L1 user rate-limit contributions are cut by 90%, and that spot and HIP-3 perp deployers may keep up to 50% of the trading fees generated by their assets . The result is dilution of capture per traded dollar rather than a collapse in activity: gross protocol revenue fell from roughly $357 million in Q3 2025 to $295 million in Q4 2025, $217.47 million in Q1 2026 and $201.83 million in Q2 2026 — about a 43% decline — while trade count and open interest climbed . Cost of revenue paid out to builders, market makers and the liquidity vault rose from under 6% of gross revenue in Q2 2025 to 18% a year later .
What is HIP-3 and why does it matter for HYPE holders?
HIP-3 is Hyperliquid's permissionless framework for builder-deployed perpetuals: deployers define the market, oracle, contract specifications, leverage limits and settlement while inheriting HyperCore's order books and margining . It cuts both ways for tokenholders. Each mainnet deployer must maintain 500,000 staked HYPE — roughly $28 million at recent prices — held for 30 days after all of that deployer's markets are halted and exposed to validator slashing, which is a real staking sink . At the same time it opens legitimate channels — deployer shares, builder codes, maker rebates, aligned-collateral discounts and growth mode — through which fees can be intercepted before reaching the Assistance Fund that funds the HYPE buyback .
Is the HYPE buyback shrinking permanently or is this temporary?
The buyback is a fixed share of earnings, so it tracks revenue directly rather than decaying on its own schedule. Hyperliquid Strategies Inc's SEC S-1/A, filed June 16, 2026, states that 99% of all fees are allocated to the Assistance Fund — up from 97% after an August 26, 2025 announcement — and that 45.23 million HYPE, equal to 4.5% of initial supply, had been acquired and permanently removed from circulation as of June 15, 2026 . Because the percentage is fixed, the dollar amount moves with earnings: the fund purchased nearly $290 million of HYPE in Q3 2025 versus roughly $149 million in Q2 2026 . Whether that reverses depends on whether growth keeps skewing toward high-deployer-share, growth-mode venues and on broader crypto volume, which weakened sharply — perp volume fell from $84 billion in early July to $43 billion within three weeks .
How concentrated is HIP-3 RWA volume among deployers?
Extremely concentrated. Trade.xyz accounts for more than 90% of all HIP-3 open interest, which means the platform's headline RWA numbers depend on one deployer's oracle definitions, leverage limits and settlement decisions — precisely the responsibilities HIP-3 assigns to deployers rather than to the protocol . That is not a theoretical exposure: a single trade on a thin Korean pre-market venue moved Trade.xyz's SK Hynix contract 19% and triggered liquidations the firm agreed to reimburse . Deployer exit is also a live scenario — Ventuals, a HIP-3 venue for pre-IPO and thematic index perps, announced on June 15, 2026 that it was sunsetting after raising over 500,000 HYPE and trading over $650 million of volume, halting and settling all its markets .
What should I watch next for Hyperliquid's fee mix?
Four measurable items, in order of signal quality. First, DeFiLlama's quarterly revenue-category breakdown, which separates buyback/tokenholder net income from builder codes, maker rebates, Unit revenue and HIP-3 deployer distributions — in Q2 2026 those lines read $148.64 million, $16.38 million, $8.80 million, $1.09 million and $7.49 million respectively . Second, the deployer-distribution line specifically, which rose from $1.10 million in Q1 2026 to $7.49 million in Q2 — almost sevenfold — while the buyback line fell from $165.34 million . Third, supply: roughly 9.92 million HYPE vests monthly to core contributors, with about 10 million (~$550 million) unlocking August 6 against circulating supply near 222 million and monthly buybacks of $53–83 million . Fourth, any official venue-by-venue attribution of the buyback shortfall, which does not yet exist publicly — Coinbase Institutional flagged the same conversion gap on March 5, 2026 without a full decomposition .
Enjoyed this article? Subscribe to get new stories by email whenever they're published.
