The House Edge

Appearance

Chapter 9 — Research Note (Phase 1) ​

Status: RESEARCH (Phase 1 — written by agent, awaiting Nick's review) Date: 2026-05-14 Working chapter: 09 — Hyperliquid: The CLOB That Worked Part: III — The Architectures (Ch 8 Solana done; Ch 9 Hyperliquid; Ch 10 Ethereum + L2s follows)

How to read this note ​

Ch 9 is Part III's structural counterpoint. Its job is to develop the claim Ch 8's closing landed: Hyperliquid is the chain whose architecture eliminates the surface Solana's vertical integration could only compress — no mempool, no separable block constructor, matching engine inside HyperBFT consensus, and a permissionless-top-21-now-expanding-to-27 validator set. The chapter must then turn the question Ch 8 left open: does that architecture eliminate the extraction, or does it concentrate it differently?

Six things to know up front before drafting:

  1. The prompt's "Q1 2026 REV ~$1B" overstates the right number by ~7×. The accurate framing is the one already in Ch 8: Hyperliquid's Q1 2026 application revenue was ~$144.8M (#1 crypto app); Solana's Q1 2026 network REV was ~$89.5M (#2 chain, down 68% YoY). The chain-level claim "Hyperliquid passed Solana in Q1 2026 REV" is correct but the dollar figure is in the low hundreds of millions, not a billion. The "$1B" figure is cumulative Assistance Fund buybacks since HYPE launch (Nov 2024–Mar 2026), not quarterly revenue. (DefiLlama; Coinpedia/Coinpedia News on X; both accessed 2026-05-14)

  2. The validator set is no longer "top-21." It is mid-expansion to 27. The Hyperliquid Foundation announced on 18 May 2026 (four days after this chapter's writing date) that the active validator count is increasing from 24 to 27 over a 30-day period. The set crossed 21 in late 2025 / early 2026. Per Ch 5's existing framing of "permissionless-top-21," the chapter must update to "permissionless-top-N (24, expanding to 27 by mid-June 2026)." Slot #21 currently requires ~525,000 HYPE in stake. Self-delegation requirement remains 10,000 HYPE locked for one year. (Coin Edition; cp0x on X)

  3. The Arbitrum bridge is in active deprecation. Hyperliquid moved to natively minted USDC on HyperCore via Circle's CCTP (cross-chain transfer protocol); USDC was linked between HyperCore and HyperEVM in early 2026. The Arbitrum bridge — the chain's I/O surface and historic security surface (the December 2024 DPRK-tagged-wallet scare; record $502.71M single-day outflows on 23 Dec 2024) — is being retired. The chapter must update the "Arbitrum bridge as I/O surface" framing to "Arbitrum bridge being deprecated in favour of native CCTP USDC; the bridge that was the chain's load-bearing risk in 2024 is no longer." (Hyperliquid on X; CCN)

  4. HIP-3 is the chapter's most live mechanic. Builder-deployed perpetuals went live on Hyperliquid mainnet on 13 October 2025. The 500,000 HYPE (~$25M at recent prices) stake requirement gates deployment. Aggregate 30-day HIP-3 trading volume across builder-deployed DEXs is ~$2.82B as of early May 2026 with 20,348 unique traders; HIP-3 now represents >35% of all trading volume on Hyperliquid. Named builder-code operators: TradeXYZ (90%+ of HIP-3 open interest, including XYZ100 tokenized stocks generating $12.7B cumulative volume), Ventuals, HyENA, Markets by Kinetiq, Felix, Dreamcash, Paragon. The chapter's load-bearing finding: HIP-3 is the explicit programmability surface that gives builders a structural-deployment primitive while keeping execution inside HyperCore — the trade-off Ch 5 named ("the chain works extremely well for what it was designed to do; it does not generalise"). HIP-3 is how Hyperliquid generalises within its design without breaking it. (Loris HIP-3 dashboard; The Block; Hyperliquid HIP-3 docs)

  5. Bitwise BHYP launched on NYSE Arca yesterday (15 May 2026). The Bitwise Hyperliquid ETF (ticker BHYP) began trading on NYSE on 15 May 2026 — one day after this chapter's writing date. Custody at Anchorage Digital. Native staking component (~85% of staking rewards retained after fees, 0.67% management fee). Approved trading counterparties per the 10 April 2026 S-1/A: FalconX, Flowdesk, Nonco, Wintermute (replacing earlier A1/Anchorage Digital affiliate). The chapter must update the "S-1 amendment" cross-reference from Chs 5/7 to "BHYP now live on NYSE Arca." (Bitwise newsroom; SEC Form S-1/A; CoinDesk)

  6. Wintermute's BTC/ETH liquidity on Hyperliquid dropped sharply in mid-May 2026. Reports from mid-May 2026 indicate Wintermute's combined BTC + ETH resting liquidity dropped from ~$40M to ~$4M, a ~90% decline. The 76-market / ~$199M-notional figure cited in Ch 2 / Ch 6 / Ch 7 was a January 2026 snapshot; the May 2026 picture is materially compressed. The chapter must hedge the "76 markets" framing as point-in-time rather than current. (TechFlow)

1. Key claims ​

Each numbered claim is something the chapter is allowed to state. Sources cited inline. The chapter's load-bearing claim is 18: where the extraction goes when the surface is eliminated.

The chapter's central architectural claim

  1. Hyperliquid is the chain on which the matching engine runs inside consensus. HyperBFT — a pipelined HotStuff-derivative BFT protocol with up to 1/3-malicious tolerance — produces approximately 70-millisecond block intervals and one-block (~200ms end-to-end median, 900ms p99, per Ch 2) finality. HyperCore (perpetuals, spot, order books, liquidations) and HyperEVM (general-purpose programmability layer) share the same consensus and the same blocks while maintaining independent state. No proposer-builder separation. No mempool. No relay layer. No block-construction outsourcing. The chain's value proposition — and its fundamental trade-off — is built on this single architectural fact. (Hyperliquid Docs — HyperCore Overview; Zealynx Security — Understanding HyperBFT; accessed 2026-05-14)

  2. The structural reason no Solana CLOB worked at spot, and Hyperliquid's CLOB works at perpetuals: visibility plus latency. Solana's gas regime is cheap; the problem was not gas. Phoenix and OpenBook ran on a chain where every order placement, cancellation, and modification was a transaction visible (via Gulf Stream forwarding to current and next slot leaders) before inclusion. Market makers exposed to that visibility could not quote tightly without being adversely selected. Hyperliquid's design closes this surface twice: (a) by matching inside consensus, there is no pending state to observe; (b) by running at HyperBFT's ~70ms block intervals, a market maker's cancel reaches the matching engine before an opportunistic taker can see and lift the stale quote. Both conditions hold simultaneously only on Hyperliquid. (Ch 2 DRAFT Ch 3 DRAFT; Hyperliquid Docs; accessed 2026-05-14)

HLP — the chain's house

  1. The Hyperliquid Liquidity Provider vault (HLP) is the chain's protocol-owned market-maker pool. Any depositor can fund it; deposits are subject to a 4-day withdrawal lock. HLP runs multiple market-making strategies inside HyperCore, quotes against incoming flow, and absorbs liquidator surplus and shortfall on the venue's perpetuals. Current TVL ~$383M as of May 2026 (recovered from a JELLY-triggered post-March-2025 trough). Lifetime cumulative PnL was ~$121.8M as of October 2025 (~450% total return); recent reports cite continued positive PnL through Q1 2026. HLP earns ~3% of trading fees; the remaining 97% routes to the Assistance Fund for HYPE buybacks. (DefiLlama HLP; Hyperliquid Docs — Protocol Vaults; 0xian Substack — HLP Vault explainer; accessed 2026-05-14)

  2. The JELLY incident (26 March 2025) is the published stress test of HLP's exposure to targeted adversarial flow. A whale opened ~$8M of JELLY shorts, then pumped the underlying spot price by ~400% in under an hour, attempting to force HLP to inherit the squeeze. Initial unrealised loss to HLP: ~$12M. The Hyper Foundation delisted JELLY and closed all positions at $0.0095 (a $700K gain to HLP). HLP TVL fell from ~$500M (pre-12 March 2025) to ~$177M by 1 April 2025 (a 65% drawdown), then recovered. The episode produced tighter leverage rules and is the named precedent for HIP-3's slashing framework (deployer faces up to 100% stake slash for invalid state transitions; "the slashed tokens are burned rather than distributed, which prevents perverse incentives between users and validators"). (Hyperliquid HLP docs; Monolith analysis; OAK Research; The Block; accessed 2026-05-14)

The validator set

  1. The active validator set crossed 21 in late 2025 / early 2026 and is mid-expansion from 24 to 27. The Hyper Foundation announced on 18 May 2026 (four days after this chapter's writing date) that the active validator count is increasing from 24 to 27 over a 30-day period. Slot #21 currently requires approximately 525,000 HYPE in delegated stake (~$26M at $50/HYPE). Self-delegation requirement: 10,000 HYPE locked for one year (a validator entering "undelegate-only mode" if self-delegation drops below the threshold). Validator commission cap rules: commissions cannot be increased except in steps where the new commission ≤1% (an anti-bait-and-switch safeguard). (Coin Edition; Hyperliquid Docs — Staking; Imperator validator guide; accessed 2026-05-14)

  2. Slashing is not yet automatically implemented on Hyperliquid for validator misbehaviour. Jailing is. Per Hyperliquid's published staking documentation: "Slashing is reserved for provably malicious behavior such as double-signing blocks at the same round. There is currently no automatic slashing implemented." Validators may vote to jail peers that do not respond with adequate latency or frequency to consensus messages; a jailed validator stops producing rewards for delegators until unjailed by quorum vote. This is structurally different from Ethereum's automatic slashing (Ch 5) and Solana's no-slashing-but-stake-loss model (Ch 5). HIP-3 does implement slashing for deployer misbehaviour (up to 100% of staked HYPE for invalid state transitions, 50% for brief downtime, 20% for network degradation) — slashing exists on Hyperliquid, but at the application-deployer layer rather than at the consensus layer. (Hyperliquid Docs — Staking; Hyperliquid Docs — HIP-3; accessed 2026-05-14)

  3. HYPE staking reward formula is inversely proportional to the square root of total HYPE staked; APR at ~400M HYPE staked is approximately 2.37%. Validator commission rates typically range 1–5%. Foundation Nodes (operated directly by the Hyper Foundation) run five validators with low commission. Named named institutional / professional validators currently in or near the active set: Imperator (~50 protocols); Bharvest (~13 protocols, 16,000+ delegators); P2P.org; the Hyperliquid Strategies x Unit Labs validator (NASDAQ: PURR; live 11 May 2026; HYPE custodied at Anchorage Digital Bank, N.A.); Figment (Ch 5 anchor for Solana also operates a Hyperliquid validator). The set is permissionless-top-N in the sense that any operator can join if they accumulate enough stake — but the practical concentration is "top-N with most stake operated by a handful of large delegators." (Hyperliquid Docs — Staking; Imperator validator guide; PRNewswire — Hyperliquid Strategies / Unit Labs; accessed 2026-05-14)

HIP-3 — the programmability primitive

  1. HIP-3 builder-deployed perpetuals launched on mainnet on 13 October 2025, with a 500,000 HYPE (~$25M) staking requirement to gate deployment. From the user perspective, fees on HIP-3 markets are 2× the standard validator-operated perp fees; the deployer receives 50% of the fee, the protocol the other 50%. The net effect: the protocol collects the same fee regardless of whether the trade is on an HIP-3 or a validator-operated perp. Deployers control oracle selection, contract specifications, maximum leverage, margin ratios, and open-interest caps, and must provide their own liquidity and front-end interfaces. The deployer's stake must be maintained for 30 days after all perps are halted (clawback against malicious behaviour). (Hyperliquid Docs — HIP-3; OAK Research; accessed 2026-05-14)

  2. HIP-3 represents >35% of all Hyperliquid trading volume as of early May 2026; aggregate 30-day HIP-3 volume ~$2.82B across builder-deployed DEXs with 20,348 unique traders. Named operators and volumes:

    • TradeXYZ (XYZ100): >90% of HIP-3 open interest; tokenized stocks and commodities; $12.7B cumulative volume; flagship example of the "TradFi-asset-on-on-chain-perp-CLOB" use case.
    • Ventuals: 13 listed markets; 30-day perp volume ~$702K; venture-stage product surface.
    • Felix Protocol: Felix has ~$300M TVL on HyperEVM and operates HIP-3 markets paired with feUSD (a Hyperliquid-native stablecoin from a Liquity V2 fork).
    • Markets by Kinetiq: HIP-3 perpetuals built on Kinetiq's kmHYPE LST; Kinetiq is the chain's largest liquid-staking protocol (kHYPE).
    • HyENA, Dreamcash, Paragon: live HIP-3 operators, smaller volumes.

    The volume migration is the chapter's most striking concentration-shift data point: a year after HIP-3 launch, the original validator-operated-perps surface is now the minority of HyperCore's trading volume. (Loris HIP-3 dashboard; Dune — HIP-3 dashboard; Metaverse Post; The Defiant — tokenized equity on Hyperliquid; accessed 2026-05-14)

HyperEVM and the dual-chain architecture

  1. HyperEVM launched on Hyperliquid mainnet on 18 February 2025 as the chain's general-purpose programmability layer. EVM-compatible smart contracts run inside the same HyperBFT consensus as HyperCore but with independent state. WHYPE (canonical wrapped HYPE) connects the two layers; USDC was natively linked between HyperCore and HyperEVM in 2026. HyperEVM TVL ~$1.9B as of mid-2026 with 60,000+ weekly active addresses and 100+ dApps. Named ecosystem firms: Kinetiq (~$639M TVL liquid-staking; peaked ~$2.28B); Felix Protocol (~$300M TVL, lending + feUSD stablecoin); LayerZero / Stargate (bridge); Bitget Wallet integration. (The Block — HyperEVM mainnet; The Defiant — HyperEVM TVL; Kinetiq Hyperliquid; Hyperliquid Docs — HyperEVM; accessed 2026-05-14)

HYPE tokenomics and the buyback machine

  1. HYPE has a 1B maximum supply; ~333M circulating in March 2026. 99% of trading fees on validator-operated perp markets route to the Assistance Fund, which executes autonomous on-chain HYPE buybacks. Weekly buyback rate as of January 2026: ~$1.7M (accelerating in high-volume periods). Cumulative buybacks crossed $1B in 15 months from HYPE launch (Nov 2024 → Feb 2026). The community proposed and executed a December 2025 burn of ~37.5M HYPE (~4.17% of total supply, ~$912M at execution prices) held by the Assistance Fund; an additional ~13% of circulating supply was proposed for burning in January 2026 per the same governance pattern. The 7%/93% protocol-revenue-to-HLP split routes the remaining fee economics back to the depositor base. (DL News — buyback machine; The Defiant — burn proposal; Phemex — Hyperliquid Fund Surpasses $1B; Tokenomics.com — HYPE tokenomics; accessed 2026-05-14)

Volume, revenue, and the Solana ranking

  1. Hyperliquid's Q1 2026 application revenue was approximately $144.8M, the #1 crypto application revenue ranking that quarter. Hyperliquid's 2025 annual revenue: ~$843M (perp fees ~$808M; spot fees ~$35M). Q1 2026 perpetual trading volume on HyperCore: ~$619.46B (the highest quarterly figure of any decentralised perp venue). 30-day perp volume in April 2026: ~$180B+. Hyperliquid's share of all global perpetual futures volume (including centralised exchanges) climbed to nearly 6% by March 2026 (up from ~3.5% one year prior); within on-chain perp DEXs, Hyperliquid commands ~70% market share with the #2 protocol (dYdX) running at ~10–12% of Hyperliquid's volume. (DefiLlama — Hyperliquid; Coinpedia on X — Q1 2026 app revenue; Crowdfund Insider — annual revenue; The Block — 6% global perp share; Yellow — 13% / 70% on-chain; accessed 2026-05-14)

  2. Hyperliquid passed Solana in Q1 2026 chain-level REV for the first time. Hyperliquid #1, Solana #2 (~$89.5M, down 68% YoY). The reversal is the chapter's structural anchor: a chain optimised for a single product can outearn a general-purpose L1 in REV terms when the general-purpose L1's memecoin-trading flow cools and the single-product L1's perpetual flow continues to grow. (PANews; accessed 2026-05-14; cross-referenced with Ch 8 Claim 17)

Market makers

  1. The named on-Hyperliquid market makers (per regulatory disclosure rather than rumour): Wintermute, Flowdesk, Nonco, FalconX — all approved trading counterparties for Bitwise's BHYP ETF per the 10 April 2026 S-1/A. Wintermute's January 2026 published positioning: ~$199M total resting notional across 76 markets, ~1,700 active orders. Mid-May 2026 reports suggest a sharp compression in Wintermute's BTC + ETH positioning (from ~$40M to ~$4M combined, a 90% reduction), the cause cited as broader market-maker repositioning. The chapter must frame the 76-market figure as a January 2026 anchor, not as a current state. (CoinShares — Hyperliquid as Xetra; Crowdfund Insider — Bitwise S-1 ETF; TechFlow — May 2026 liquidity exit; accessed 2026-05-14)

The Arbitrum bridge and CCTP migration

  1. The Arbitrum bridge is being deprecated in favour of natively-minted USDC via Circle's CCTP. USDC was linked between HyperCore and HyperEVM in late 2025 / early 2026. Bridge TVL (USDC + other assets) sat around ~$2.22B in late 2024 at the time of the DPRK-tagged-wallet scare; record single-day outflow on 23 December 2024 was approximately $502.71M. The chapter must frame the Arbitrum bridge as: (a) the chain's I/O surface in 2024–2025; (b) the locus of the chain's largest single security event (a non-exploit but a market-confidence event); (c) being retired as the chain matures into native USDC, removing the "four-of-four validators" attack surface that drove the December 2024 outflows. (Cointelegraph — record outflows; Hyperliquid on X — bridge deprecation; CCN — bridge retire; accessed 2026-05-14)

The team — concentrated, pseudonymous, no outside capital

  1. Hyperliquid was co-founded by Jeff Yan (Harvard 2017, Hudson River Trading, founder of Chameleon Trading) and a pseudonymous developer "iliensinc" (Yan's Harvard classmate). Hyperliquid Labs is approximately 11 people. The team took zero outside venture capital; Yan bootstrapped the protocol with profits from his prior trading firm. The structural framing the chapter must develop: the most economically successful single-application chain of 2025–2026 was built by ~11 people with no outside capital, against a competitive surface (Solana, Ethereum L2s) that absorbed hundreds of millions in VC. Whether that team concentration is a strength or a structural risk is the chapter's secondary architectural question. (Fortune — Harvard grad helped make Hyperliquid; Datawallet — Who is Jeff Yan; accessed 2026-05-14)

The BHYP ETF — institutional access surface

  1. The Bitwise Hyperliquid ETF (BHYP) began trading on NYSE Arca on 15 May 2026. Custody: Anchorage Digital. Approximately 85% of staking rewards retained for ETF investors after fees. 0.67% annual management fee. Approved trading counterparties: FalconX, Flowdesk, Nonco, Wintermute. The ETF launching on the day after this chapter's writing date is a publication-timing fact the draft will address with present-tense framing. The institutional implication: Hyperliquid is the first single-application L1 to acquire a regulated US ETF wrapper with native staking yield. The structural significance is that the chain's permissionless-top-N validator set now has Anchorage-custodied HYPE among its delegated stake — institutional flow is being introduced to the validator base, not just to the trading surface. (Bitwise newsroom; PR Newswire — BHYP launch; accessed 2026-05-14)

The chapter's verdict frame

  1. The extraction surface Solana's vertical integration could only compress, Hyperliquid's architecture eliminates — but the value capture re-concentrates into four named loci. The chapter must land this as the verdict:
  • The HLP: the chain's house earns ~3% of trading fees and absorbs all liquidation surplus, but is also exposed to targeted adversarial flow (JELLY). HLP's depositors are the chain's passive LP class equivalent — paid by the surface they fund, but tail-exposed to the venue's worst flows.
  • The validator set: 24-expanding-to-27 large stakers (each requiring >525K HYPE delegated; ~$26M at recent prices). 99% of fees flow to the Assistance Fund which buys back HYPE, which the validator set is largely the holders of. The validator set is the structural beneficiary of every trade on the chain, in proportion to its HYPE stake. The named professional operators (Imperator, Bharvest, Foundation Nodes, P2P.org, Figment, Hyperliquid Strategies / Unit) are the chain's incumbent stake holders.
  • The HIP-3 deployers: TradeXYZ, Ventuals, Felix, Kinetiq Markets, HyENA. They earn 50% of the doubled-fee on their deployed markets. The 500K HYPE deployment stake (~$25M) is a structural moat against permissionless competition.
  • HYPE holders, via the buyback machine: $1B+ of cumulative buybacks since launch routes protocol revenue back to the token's market price, which the validator set, the HIP-3 deployers, and the early Foundation distribution disproportionately hold.

The structural finding: Hyperliquid does not have searchers, sandwich-attackers, builders, or relays in its extraction stack. It has the house, the validator set, the deployers, and the token holders. The names are different; the concentration is comparable; the visibility is in some ways higher (the HLP P&L is on-chain; the buyback is on-chain) and in some ways lower (the deployer-economics surface is opaque; the bilateral arrangements between HIP-3 deployers and their underlying market makers are not publicly disclosed). This is the chapter's "concentrated differently" answer to the question Ch 8 left open.

  1. What Hyperliquid loses: the smart-contract surface, the dev ecosystem, the cross-application composability. HyperEVM exists as an explicit answer to part of this trade-off, but ~$1.9B TVL on HyperEVM is roughly 1/50th of Solana's DeFi TVL and ~1/30th of Ethereum's. The chain is not optimising to be a general-purpose application platform. It is optimising to be a derivatives venue with sufficient programmability to allow secondary protocols (Kinetiq for liquid staking; Felix for borrowing; LayerZero for bridging) without compromising the matching engine's performance. Ch 12 picks up the "appchain" thesis broadly; Ch 9's role is to land the trade-off as deliberate, not accidental, and quantify what was sacrificed and what was gained.

2. Numbers to verify ​

#NumberSourceDateFlag
N1Hyperliquid 2025 annual revenue ~$843M (perp ~$808M; spot ~$35M)Crowdfund Insider; Bitget News2025Strong
N2Hyperliquid Q1 2026 application revenue ~$144.8M (#1 crypto app); Solana Q1 2026 chain REV ~$89.5M (#2 chain)Coinpedia on X; PANewsQ1 2026The structural anchor; corrects prompt's "~$1B"
N3Hyperliquid Q1 2026 perp volume ~$619.46B; April 2026 30-day perp volume ~$180B+DefiLlama; Yellow researchQ1–Q2 2026Strong
N4Global perp futures share ~6% (March 2026); on-chain perp DEX share ~70%; dYdX at ~10–12% of Hyperliquid's monthly volumeThe Block; Atomic WalletMarch 2026Strong
N5HLP TVL ~$383M (May 2026 DefiLlama snapshot); lifetime cumulative PnL ~$121.8M as of October 2025 (~450% total return)DefiLlama HLP; 0xian SubstackMay 2026; Oct 2025Strong; PnL flag — newer cumulative figure not in any single primary source
N6JELLY incident (26 Mar 2025): HLP initial unrealised loss ~$12M; would have lost ~$230M entire reserve at $0.15374; closed at $0.0095 = $700K HLP gain; TVL drawdown $500M → $177M (65%); recovered to ~$195M by May 2025Monolith; The Block; OAK ResearchMarch–May 2025Strong
N7HIP-3 mainnet launch 13 Oct 2025; 500K HYPE staking requirement; 50/50 deployer/protocol fee split; user fee 2× standardHyperliquid HIP-3 docs; The Block13 Oct 2025Primary source
N8Aggregate 30-day HIP-3 volume ~$2.82B; 20,348 unique traders; HIP-3 >35% of all Hyperliquid trading volume; TradeXYZ XYZ100 cumulative volume $12.7B (>90% of HIP-3 OI)Loris HIP-3 dashboard; Dune dashboard; Metaverse PostMay 2026New chapter-level finding; verify with Loris.tools live on draft date
N9HyperEVM launched 18 Feb 2025; TVL ~$1.9B; 60,000+ weekly active addresses; 100+ dApps; Kinetiq ~$639M TVL (peak ~$2.28B); Felix ~$300M TVLThe Block; The Defiant; Kinetiq HyperliquidFeb 2025–May 2026Strong
N10Validator set crossed 21 in late 2025; currently 24, expanding to 27 by mid-June 2026; slot #21 ~525,000 HYPE; self-delegation 10K HYPE for 1 year; commission cannot rise by more than 1%Coin Edition; Hyperliquid Docs — StakingMay 2026New / corrects Ch 5 framing
N11HYPE staking APR ~2.37% at ~400M HYPE staked (inverse-square-root formula); commissions typically 1–5%Hyperliquid Docs — Staking; StakingRewards2026Primary source
N12Cumulative HYPE buybacks crossed $1B in 15 months (Nov 2024–Feb 2026); weekly buyback ~$1.7M (Jan 2026); 37.5M HYPE (~4.17% of total supply, ~$912M) burned 25 Dec 2025; additional ~13% of circulating supply proposed for burn Jan 2026DL News; Phemex; The Defiant — burn proposal2025–2026Strong
N13HYPE max supply 1B; circulating ~333M (March 2026); 99% of validator-perp fees to Assistance Fund (buybacks); 3% of trading fees to HLP, 97% to AF (per Ch 5 footnote 19)Tokenomics.com; Buildix; cross-ref Ch 52026Strong; flag the 3%/97% vs 99%/1% framing distinction below in §3
N14Bitwise BHYP launched on NYSE Arca 15 May 2026; 0.67% management fee; ~85% of staking rewards retained for ETF; custody Anchorage Digital; counterparties FalconX, Flowdesk, Nonco, WintermuteBitwise newsroom; PR Newswire; Bitwise S-1/A 10 Apr 202615 May 2026Same-day publication; verify with NYSE listing on draft date
N15Hyperliquid bridge Dec 2024 single-day outflow record ~$502.71M; net outflow ~$250M; TVL remaining ~$2.22B; DPRK-tagged wallets accrued ~$701K loss; "no DPRK exploit"Cointelegraph; CoinDesk; Blockworks; Hyperliquid Labs Discord — denied exploitDec 2024Strong
N16Arbitrum bridge being deprecated; USDC natively linked HyperCore ↔ HyperEVM via Circle CCTP; final state: all USDC natively minted, bridge retiredHyperliquid on X; CCN — bridge retire; Bitget — native USDC2026Update from Ch 8 framing
N17Wintermute Hyperliquid positioning January 2026: ~$199M total notional across 76 markets, ~1,700 active orders; May 2026 BTC+ETH liquidity dropped from ~$40M to ~$4M (~90% compression)CoinShares; TechFlowJan–May 2026Time-of-snapshot flag
N18HyperBFT block intervals ~70ms; end-to-end finality median ~200ms / p99 ~900ms (Ch 2); throughput ~200K orders/sec on HyperCoreHyperliquid Docs — Overview; Zealynx Security; Ch 2 anchor2025–2026Primary source
N19HIP-3 slashing framework: up to 100% deployer stake for invalid state transitions; 50% for brief downtime; 20% for network degradation. Validator-layer slashing not yet automaticHyperliquid HIP-3 docs; Hyperliquid Docs — Staking2025–2026Primary source
N20Bitwise S-1/A trading counterparties (10 Apr 2026): FalconX, Flowdesk, Nonco, Wintermute (replacing A1/Anchorage Digital affiliate)Bitwise S-1/A 10 Apr 2026; The BlockApril 2026Already cited Chs 5, 7
N21Hyperliquid Labs ~11 people; no outside venture capital; Yan + iliensinc (Harvard '17 classmate) co-foundersFortune; Datawallet; Medium — 10-person team2025–2026Soft (team-size figure varies 10–11 across sources)
N22Hyperliquid Strategies (NASDAQ: PURR) + Unit Labs joint validator: announced 7 May 2026; live 11 May 2026; HSI HYPE custodied at Anchorage Digital BankPR Newswire; StockTitanMay 2026Same-week event; Ch 5 anchor

Deliberately not pinned (flag for Nick / Phase 2):

  • Per-validator stake distribution and named identities of all 24 active validators (no canonical published list as of writing date; app.hyperliquid.xyz/staking is the live source)
  • Specific HIP-3 deployer P&L (deployer revenue split with their underlying market makers is not publicly disclosed)
  • BHYP first-day flows (publication-timing — ETF launched same day as chapter writing)
  • HLP cumulative PnL in May 2026 (DefiLlama displays "Earnings $0" because all gross revenue is classified as cost-of-revenue against the buyback program; the cumulative trading-PnL figure requires reading the on-chain vault performance separately)
  • The "3% of trading fees to HLP / 97% to AF" (Ch 5) vs the "99% of fees to Assistance Fund for HYPE buybacks" (DefiLlama / Tokenomics.com) framings are not necessarily inconsistent (the 3%/97% may describe an internal allocation after the 99%/1% protocol-vs-deployer split), but the chapter should clarify with a single primary-source citation

3. Contested or evolving claims ​

  • Validator set size. Ch 5 framed Hyperliquid as "top-21" — accurate through late 2025 / early 2026. The set is now 24, expanding to 27 by mid-June 2026. Ch 9 must update the framing and Ch 5 should be cross-checked in REVIEW_NOTES. The structural argument ("small permissionless-but-elected validator set") is durable; the specific number is moving.

  • Wintermute's positioning. January 2026 published figures (~$199M / 76 markets) are the regulatory-anchor data point. Mid-May 2026 reports of a sharp BTC + ETH liquidity reduction (~90%) suggest the picture is dynamic. The chapter must cite the January figure (regulatory record) and flag the May compression.

  • "Hyperliquid passed Solana in Q1 2026 REV." Per Ch 8 (the structural anchor): Hyperliquid #1, Solana #2 (~$89.5M) for the quarter. Q2 2026 in-progress; whether the ranking persists is an open question. The chapter must frame as Q1 2026 fact, not as durable shape.

  • HLP cumulative PnL. Different sources cite different snapshots ($61.2M as of May 2025; $121.8M as of October 2025; positive but not explicitly stated through Q1 2026). The chapter should pin a specific dated snapshot rather than aggregate.

  • The Arbitrum bridge. Two simultaneously true facts: (a) the bridge was the chain's load-bearing security surface in 2024 (DPRK scare; $502.71M single-day outflow); (b) the bridge is being deprecated in favour of CCTP-native USDC. The chapter should treat the bridge as historical I/O surface while noting the current migration.

  • The 99%/1% vs 3%/97% fee framings. Per Ch 5 (HLP earns 3% of trading fees, 97% routes to HYPE buybacks). Per multiple secondary sources (99% of fees to Assistance Fund for HYPE buybacks). The two are not necessarily contradictory but require a primary-source citation to align — likely the difference between "97% of HLP-routed flow" and "99% of validator-operated perp fee flow." Phase 2 should resolve via Hyperliquid Docs.

  • The "Solana #2 by REV behind Hyperliquid" durability. Solana's REV depends substantially on memecoin trading activity (the 2025 peak month was January, $551.7M of REV with Jito tips). The Q1 2026 reversal coincides with the memecoin-cycle cooling. If Solana's volume recovers in H2 2026 (Alpenglow mainnet, Q3/Q4 2026 — Ch 8), the rank could flip back. The chapter should frame the Q1 2026 fact rather than projecting durability.

  • HIP-3 share. ">35% of all Hyperliquid trading volume" is a striking finding but is partly TradeXYZ-driven (>90% of HIP-3 OI). The chapter should frame as "HIP-3 reached >35% in mid-May 2026 with one dominant operator" rather than as a diversified migration.

  • HLP "earnings $0". DefiLlama displays Hyperliquid earnings as $0 across all periods because all gross revenue is classified as cost-of-revenue (the buyback program offsets earnings). This is an accounting framing, not a P&L claim. The chapter should resolve as "the chain runs at structural breakeven by design — revenue and buybacks net out."

  • The "no outside venture capital" framing. Genuinely zero outside VC for the protocol; HYPE airdrop distributed to users (no investor pre-sale). But the secondary ecosystem (Kinetiq, Felix, Hypurr, etc.) does have VC backing. The chapter should be precise: the chain itself was bootstrapped; the ecosystem around it took normal venture rounds.

4. Characters introduced ​

New "Meet the Actor" sidebar candidate: The Chain's House. Ch 2 introduced HLP as the protocol-owned liquidity vault. Ch 9 introduces HLP-as-actor: depositors fund it, it quotes against incoming flow, it inherits liquidation surplus, it gets attacked by targeted adversarial flows (JELLY). The actor parallel is structurally different from "Meet the Market Maker" (Ch 2) — it is the venue itself as a counterparty. Recommend a "Meet the House" sidebar developing this distinction: how HLP is structurally different from a centralised exchange's market-maker desk (Citadel Securities) and different from a passive AMM LP (Uniswap V3) — closer to a clearinghouse's residual position than to either.

Returning institutions / actors, with Hyperliquid-specific roles:

  • Jeff Yan + iliensinc + Hyperliquid Labs (Chs 2, 3, 5): the founders and the ~11-person team; the structural character of "concentrated team, no outside capital." First named in Ch 9 as the chapter's institutional protagonist.
  • HLP (Chs 2, 5): the chain's house; structurally separate from the validator set; JELLY-tested.
  • Wintermute (Chs 2, 5, 6, 7): the 76-market market maker per January 2026; approved Bitwise BHYP counterparty; mid-May 2026 compression flag.
  • Flowdesk, Nonco, FalconX (Chs 5, 7): Bitwise BHYP approved counterparties; institutional market-maker roster.
  • Bitwise + BHYP (Chs 5, 7): the ETF launched 15 May 2026 on NYSE Arca; the chapter's institutional-access anchor.
  • TradeXYZ / XYZ100 (new in Ch 9): HIP-3 flagship operator; >90% of HIP-3 OI; $12.7B cumulative volume; tokenized stocks and commodities on-chain.
  • Ventuals (new in Ch 9): venture-stage HIP-3 operator; 13 markets; the "long-tail builder" example.
  • Felix Protocol (new in Ch 9): HyperEVM lending + feUSD stablecoin; ~$300M TVL; the Liquity-V2-on-HyperEVM example.
  • Kinetiq / kHYPE (new in Ch 9): the chain's largest liquid-staking protocol; ~$639M TVL (peak ~$2.28B); kmHYPE underlying for Markets by Kinetiq HIP-3 perp deployment.
  • Hyperliquid Strategies (NASDAQ: PURR) + Unit Labs: 11 May 2026 validator launch; HYPE treasury custodied at Anchorage Digital Bank.
  • Hyper Foundation / Foundation Nodes: governance + 5 protocol-operated validators.
  • Imperator, Bharvest, P2P.org, Figment: named institutional validators (Figment as Ch 5 anchor for Solana also operates Hyperliquid).
  • Circle (CCTP): enables native USDC on HyperCore and HyperEVM; the structural replacement for the Arbitrum bridge.
  • Anchorage Digital: custodian for BHYP and for the Hyperliquid Strategies / Unit Labs validator.

Returning characters in narrowly different roles: the chapter does not re-introduce Phantom, Helius, Jupiter, Jito Labs, Anza, etc. — they belong to Solana and don't operate on Hyperliquid. The chapter explicitly notes the absence of a third-party infrastructure layer on Hyperliquid as a structural finding (per Ch 6 anchor).

5. Worked example candidates ​

Per the SPEC, the chapter benefits from a worked example that illustrates the chain's structural finding rather than dissecting any single mechanic.

Candidate A — "$10,000 USDC long-BTC perp through Hyperliquid" ​

The Alice / Bob / Carlos / institutional-trader convention from prior chapters opens a $10,000 USDC long BTC perpetual position on Hyperliquid's validator-operated BTC market.

StepWhat happensTake from $10KCumulative
Deposit$10,000 USDC bridged via CCTP (native, free)$0$0
Order placementLimit/market order submitted to validator; matched inside HyperBFT consensus in ~70ms; no mempool visibility$0 (no sandwich surface)$0
Maker/taker fee~0.025% taker / 0.015% maker (varies); say 2.5 bps taker on entry$2.50$2.50
SpreadBTC perp spread ~$1 vs Binance ~$5.50; ~0.5 bps embedded$0.50$3.00
Liquidation risk premiumIf position liquidated, surplus/shortfall accrues to HLP (the depositor base, not to a specialist)N/A on entry$3.00
ExitSimilar 2.5 bps taker$2.50$5.50
Net$5.50 round-trip$5.50 vs ~$73 on Solana 2024; ~$8–15 on Solana 2026 (Ch 8)

The dollar arithmetic shows the chain's claimed structural advantage: the same $10,000 trade pays roughly half what it pays on Solana 2026, with no sandwich surface and no formal exclusive-flow extraction. Where the dollar still flows: fees to the chain (99% to Assistance Fund / HYPE buybacks; the validator set and HYPE holders are the structural beneficiary); HLP earnings on the inventory side; tighter spread because of the design, but spread still earned by the market makers (now Wintermute, Flowdesk, Nonco, FalconX; the institutional-disclosure roster).

The chapter's verdict: the dollar Alice paid is smaller — but it still flows to a concentrated set of actors.

  • Pros: integrates HLP, fee mechanics, the validator set, and HYPE tokenomics in one trace; mirrors Ch 8's structural worked example.
  • Cons: per-step fees are illustrative (taker/maker rates vary by tier); the HLP-as-counterparty mechanic is subtler than direct fees.

Candidate B — "An institutional desk runs $1M XYZ100 (tokenized S&P 500) via TradeXYZ on HIP-3" ​

An institutional trading desk takes a $1M long-S&P-500 position via TradeXYZ's XYZ100 HIP-3 perp. The trade routes to TradeXYZ's deployer infrastructure, executes inside HyperCore's matching engine, and pays a 2× standard fee — 50% of which routes to TradeXYZ, 50% to the protocol (which routes to the Assistance Fund / HYPE buybacks). The dollar trace shows the HIP-3 economic surface: the chain pulls TradFi flow on-chain by giving a single firm the right to deploy a market against $25M of staked HYPE.

  • Pros: most differentiating new material for the chapter; integrates HIP-3, the 500K HYPE stake, and the "concentrate the extraction differently" verdict; clear contrast with Solana's prop-AMM displacement.
  • Cons: institutional desk operating tokenized-stock perps is narratively further from the book's retail-trader spine; harder for the Goldman MD to read on first pass.

Candidate C — "Alice deposits $5,000 into HLP and trades alongside it" ​

Alice routes part of her exposure through HLP instead of trading directly. She deposits $5,000 to HLP (4-day withdrawal lock), earns ~3% of trading fees and her share of HLP's market-making P&L. During Q1 2026, HLP returned ~X% (pin the number at draft time). During the JELLY incident, HLP's TVL drew down 65% before recovering. Alice is now economically the chain's house — earning on every trade that crosses her capital, exposed to the worst flows.

  • Pros: the cleanest illustration of HLP-as-actor; positions Alice as a depositor in the chain's structural extraction beneficiary.
  • Cons: doesn't develop the matching-engine mechanic or HIP-3.

Recommendation ​

Candidate A as the primary worked example, with HLP's depositor economics threaded as a sidebar or short subsection rather than a separate worked example. A delivers the dollar arithmetic the book's spine requires; the HLP-as-depositor story is best handled in the "Meet the House" sidebar rather than as a competing example. Candidate B's HIP-3 case is best deployed as a one-paragraph passage with the XYZ100 cumulative volume figure rather than as the chapter's anchor — it's structurally the most differentiating new mechanic but narratively the furthest from the book's retail-trader frame.

6. Open questions for Nick ​

Q1 — Word budget. Ch 5/6/7/8 ran ~5,500–7,500 words inc. footnotes. Ch 9 has a lot of new mechanics (HIP-3, HLP, HyperEVM, the validator set, the BHYP ETF). Recommend 6,000–7,000 words. Acceptable, or tighter?

Q2 — The prompt's "Q1 2026 REV ~$1B" framing. The accurate Q1 2026 numbers are Hyperliquid app revenue ~$144.8M (#1 app) and Solana network REV ~$89.5M (#2 chain). The "$1B" figure is cumulative Assistance Fund buybacks since HYPE launch. Recommend Ch 9 frames the rank reversal correctly without overstating the dollar figure; the rank reversal is the structural point. Confirm.

Q3 — Validator set update. Ch 5 framed Hyperliquid as "top-21." The set is currently 24, expanding to 27 by mid-June 2026. Recommend Ch 9 updates the framing and we flag Ch 5 for a Phase-3 retroactive edit. Acceptable, or keep the "top-21" historical framing and let Ch 9 land the update?

Q4 — BHYP ETF launch timing. BHYP went live on NYSE Arca on 15 May 2026 (one day after this chapter's writing date). Recommend Ch 9 frames as present-tense ("BHYP began trading on NYSE Arca yesterday") since publication will be well after the launch. Confirm.

Q5 — "Meet the House" sidebar. HLP is the structurally most novel actor in the book — the venue itself as counterparty. Recommend a dedicated sidebar developing the distinction from a Citadel Securities desk and a Uniswap V3 LP. Or skip and keep HLP inline?

Q6 — HIP-3 treatment depth. HIP-3 is the chapter's most live mechanic and produces the chapter's most striking concentration shift (>35% of trading volume in seven months from launch; >90% of HIP-3 OI on one operator). Recommend full subsection inside mechanics. Alternative: dedicated worked example (Candidate B). Confirm preference.

Q7 — HyperEVM treatment depth. HyperEVM is structurally important (the chain's programmability answer) but not where the chapter's value-extraction story lives. Recommend ~300-word subsection landing the dual-chain shape, with Kinetiq + Felix as the named ecosystem anchors. Or tighter?

Q8 — The Arbitrum bridge story. December 2024 DPRK scare + ~$502.71M single-day outflows + active CCTP migration. Recommend: brief subsection in mechanics (the bridge was the chain's biggest 2024 security event; the migration to native USDC retires the surface). Or save the bridge story for Ch 10 (Ethereum + L2s) and Ch 11 (Losers)?

Q9 — Worked example. Candidate A (full-stack dollar trace of a $10K BTC perp) is my recommendation. Confirm, or prefer B (institutional XYZ100 trade) or C (Alice as HLP depositor)?

Q10 — Jeff Yan / iliensinc / 11-person team treatment. Genuinely a 2026 anomaly — biggest crypto application by revenue, smallest team, zero outside VC. Recommend a one-paragraph cameo plus a footnote citing Fortune's profile, not a full sidebar. Or develop more?

Q11 — Q1 2026 ranking durability. The "Hyperliquid passed Solana in Q1 2026 REV" anchor is Ch 8's reveal and Ch 9's load-bearing chapter-frame. If Q2 2026 reverses it (Alpenglow mainnet + memecoin volume recovery), the chapter's structural claim survives but the headline narrows. Recommend framing as "Q1 2026 fact" rather than "durable shape." Confirm.

Sources cited ​

All URLs accessed 2026-05-14. URLs cited inline in claims/numbers above; this is the consolidated bibliography.

Primary research and protocol documentation:

  • Hyperliquid Docs — HyperCore Overview, HyperBFT, Staking, Vaults, HIP-3, Bridge: hyperliquid.gitbook.io
  • Hyperliquid X / official posts — bridge deprecation, USDC native: x.com/HyperliquidX
  • DefiLlama — Hyperliquid (chain + protocol), Hyperliquid HLP, Hyperliquid Bridge, Kinetiq Markets, Felix: defillama.com
  • Loris HIP-3 dashboard — HIP-3 builder DEX stats: loris.tools/hip3
  • Dune — Hyperliquid HIP-3 dashboard (yandhii): dune.com/yandhii/hip3
  • Bitwise SEC S-1/A (10 April 2026) and BHYP launch newsroom: bitwiseinvestments.com, streetinsider.com, sec.gov
  • PR Newswire — BHYP launch, Hyperliquid Strategies / Unit Labs validator: prnewswire.com

News and analytical coverage:

  • Fortune — "How a Harvard grad helped make Hyperliquid": fortune.com
  • CoinDesk — BHYP filings, Bitwise updates, JELLY coverage: coindesk.com
  • The Block — HIP-3 activation, HyperEVM mainnet, 6% global perp share, BHYP filings, JELLY: theblock.co
  • The Defiant — HyperEVM TVL, burn proposal, tokenized equity on Hyperliquid: thedefiant.io
  • Crowdfund Insider — Bitwise S-1/A; $820M annual revenue: crowdfundinsider.com
  • Cointelegraph — Dec 2024 outflows + DPRK scare: cointelegraph.com
  • Blockworks — Hyperliquid analytics, security scrutiny: blockworks.com (with both .com and .co coverage)
  • DL News — Hyperliquid buyback machine $1B, dark AMMs (cross-ref Ch 8): dlnews.com
  • Phemex — Hyperliquid fund $1B + 4.17% burn: phemex.com
  • CoinShares — "Hyperliquid: the Xetra of digital finance": coinshares.com
  • TechFlow — Wintermute liquidity drop May 2026: techflowpost.com
  • Yellow — perp dominance research: yellow.com
  • Crypto Times — $5.23M revenue day April 2026: cryptotimes.io
  • Atomic Wallet, Datawallet, OAK Research, Metaverse Post — HIP-3, HyperEVM, JELLY: various
  • Monolith — JELLY post-mortem: medium.com/@monolith.vc
  • 0xian Substack — HLP vault explainer: 0xian.substack.com
  • Zealynx Security — HyperBFT/HyperCore/HyperEVM technical deep dive: zealynx.io
  • Imperator — validator guide: imperator.co
  • StakingRewards — HYPE staking: stakingrewards.com
  • Coin Edition — validator set expansion 24→27: coinedition.com
  • Hodl Group, OneKey, Mefai (Medium), Tokenomics.com — additional analytical coverage

Pulled through from prior chapters' RESEARCH.md (URLs in those files):

  • Ch 2: HyperBFT consensus ~70ms / ~200ms finality; CCN coverage of Jeff Yan's January 2026 spread comparison; Yellow on Hyperliquid 13% / 70% on-chain perp share; CoinShares "Hyperliquid as Xetra"; Geronimo HLP risk-return analysis
  • Ch 3: Hyperliquid HyperCore Overview docs (order book semantics inside consensus)
  • Ch 5: Hyperliquid Staking docs (delegated PoS, top-21 mechanics, 10K HYPE self-delegation, ~2.37% APR); HLP 3% / 97% routing per DefiLlama (footnote 19); Hyperliquid Strategies/Unit Labs 7 May 2026 PR
  • Ch 6: Hyperliquid as the chain with no third-party infrastructure layer
  • Ch 7: Bitwise April 2026 S-1/A naming Wintermute, Flowdesk, Nonco, FalconX as approved-not-exclusive counterparties; HIP-3 builder codes as a different structural primitive from Ethereum-style EOF; Dwellir analysis (~$40M cumulative builder revenue; ~40% of DAUs on third-party frontends — figures pre-date Ch 9's $2.82B HIP-3 30-day volume)
  • Ch 8: Q1 2026 Solana REV $89.5M; Solana #2 by REV behind Hyperliquid for the first time (PANews)

Phase 1 is complete. Per the user's compressed-review pattern, Phase 2 (OUTLINE.md) follows immediately after Nick's review of this RESEARCH.md.