We are rebuilding the global financial system. Not all at once. Bit by bit, piece by piece, step by step, until one day everyone wakes up to find the old infrastructure gone and the whole world onchain. If you blink you might miss it, but still it will catch you by surprise.

On-chain data suggests Trade[XYZ]/Unit is positioning to become the next HIP-4 deployer, and the window opens on 5 September at 09:41 UTC. The evidence sits in a single day of activity on the @tradexyz HIP-3 deployer wallet: At 02:06 UTC on 29 August it moved 230,430 $HYPE into staking, funded by roughly 226.4K $HYPE gathered from three addresses over the prior 48 hours, and delegated 230,420 of it to the Hyperliquid Strategies x Unit validator: That is the first delegation this wallet has made outside in thirteen months. Seven hours later, at 09:23, it undelegated 500,000 $HYPE, and at 09:41 queued the withdrawal: At 769,605 staked, releasing a clean 500,000 would have dropped the wallet under the 500K HIP-3 requirement and broken the existing xyz bond. The 230,430 deposit exists to make the exit possible without that happening. The wallet now holds 500,035 delegated, with 500,000 in queue. Staked $HYPE cannot move between accounts, and one deployer stake supports one perp DEX, so a second bond requires exactly this: a seven-day clock started in advance while the first bond stays intact. On 5 September the capital lands in spot, and spot-to-staking is instant. are we ready? @sershokunin

Perpetual futures were invented to price things that don't trade. Perpetual futures were proposed in 1993 for housing, where there's no continuous market to quote against. @entropyIO's argument is that pre-IPO companies sit in exactly that gap. A private company has no listing date and no live spot price, so the contract treats the eventual IPO as the delivery event and prices in market cap rather than shares, which sidesteps the constantly shifting share count. That leaves the real problem: what does the oracle anchor to? Last-round marks lag, and an unanchored oracle gets pushed around on a thin book. Their answer is to let the market's own price count for more when there are real bids and offers sitting on both sides, and lean on outside references when there aren't. Moving right along the bottom means more real size resting on the book, and the curve shows how much of the price the market gets to set for itself. A thin book gets around a third of the say, a deep one closer to nine-tenths, and the line never quite reaches the top, so outside references always keep a small hold. Funding comes out of whatever gap is left, which means a deep book on both sides brings funding close to zero, while a book that empties out hands control back to the outside references.

US spot HYPE ETFs have taken in $360M net across 76 sessions since $THYP listed on 12 May → launch month, 12 May onward: $THYP alone for three sessions, $BHYP joins on 15 May, and the pair pull $137.7M by month end with a $25.5M peak on 20 May → early June cools to a few million a day until Grayscale's first flow lands on 3 Jun, then 25 Jun prints $112.6M in one session and effectively closes the month at $171.0M → July reverses: $HYPE slides $71 to $54, flows go net negative at -$15.2M, the largest outflow day of the series hits on 29 Jul at -$8.8M, and most sessions print zero → early August stays flat through the $52 low, with the first ten sessions contributing $9.4M between them → the last four sessions turn: $5.7M, $7.5M, $14.7M, then $24.4M on 27 Aug into the $86.70 high - $52.3M, or 85% of August, arriving after price had already cleared $78 → across all 76 sessions: $BHYP $155.6M accumulated steadily, $HYPG $142.4M almost entirely from one day, $THYP $57.2M front-loaded into launch Track it live, thanks to @hl_eco:

The oracle for Outcome's markets is the exchange those markets trade on. On 29 August, @Outcomexyz went live as the first third-party HIP-4 deployer on @HyperliquidX mainnet, carrying a 500K $HYPE bond it had self-funded back in May: Twenty-three markets shipped: daily crypto binaries, a three-way on the September FOMC decision, and price contracts written on top of the HIP-3 equity, index and commodity perps. A $1M rewards programme opened alongside it at $200K a month. Until this week HIP-4 was one recurring bitcoin:native binary that validators deployed and settled themselves. It now has a supply side, which means it can finally be measured against @Polymarket, @Kalshi and @trylimitless. → Listing & supply: who is allowed to create a market. Permissionless with a bond (Outcome), curated (Polymarket), exchange-only (Kalshi), fully open with a creator fee (Limitless). → Collateral & risk: what backs a position and where it sits. All four are fully collateralised; the difference is $USDC on HyperCore versus $pUSD on Polygon versus USD held by the exchange versus $USDC on Base. → Resolution source: where the answer comes from. Hyperliquid's own perp mark, UMA, exchange determination, or Pyth with a manual fallback. → Dispute & recourseЮ what happens when the answer is wrong. A slashable bond, an economic dispute with a token vote, regulatory oversight, or nothing at all. → Cost to trade: taker and maker economics. Variance curves at Polymarket and Kalshi, an inverted curve at Limitless, and zero at Outcome because Hyperliquid waived the protocol fee. → Liquidity incentives: whether market-making is paid for out of a fixed pool or out of taker fees. → Margin & adjacency: whether outcomes share collateral with perps and spot, or sit in their own silo.

for the first time you can short bonds on @HyperliquidX, $TLT and $IWM went live on markets today under the @Markets_xyz/@Kinetiq_xyz deployer. two ETFs, but they carry more macro meaning. TLT TLT holds US treasuries with 20+ years to maturity. it's effectively a leveraged bet on where rates go, without any leverage at all. when the market thinks the fed is cutting, TLT rallies. when inflation surprises to the upside, it gets destroyed. IWM IWM tracks the russell 2000 - two thousand small US companies. small caps live and die on domestic credit conditions, but when the cycle turns up, it moves before the S&P does. Now avaliable 24/7/365 on @Markets_xyz:

Loracle is $17.5M underwater on his $HYPE short and just queued the rest of his stake 1) · $HYPE short 658,404 → 605,580, trimmed 52,824 for the first time in this whole sequence. · Entry $53.97 against an $83.06 mark: −$17,535,055, plus $392,630 paid in funding. Break-even $54.58, liquidation $114.68. · Position value $50.22M · Spot now holds 163,840 $HYPE ($13.61M, +1.87%) alongside $25.16M $USDC. 2) · On 27 August at 20:07 UTC he undelegated 274,925.27 $HYPE from Node - Alphaticks and initiated the withdrawal a minute later. · Staked is down to 96,861.28 with 275,025.59 pending across two withdrawals, unlocking around 3 September. · Wallet total $31.33M, of which $30.89M is still staked and $442,219 sits on HyperEVM. 3) The book turned over almost entirely. Gold, $PAXG, bitcoin:native, $ETH, binancecoin:native, $TRX, $PUMP, $ZEC and the second $HYPE short are all closed. What remains is a $52.05M net short across eight positions at 0.81x leverage: · xyz:SNDK short 15,567, $22.56M, +$398,322. · New xyz:NVDA short 60,880, $13.73M, +$69,131. · xyz:MU short 10,470, −$461,358. xyz:PLTR short 23,045, −$258,306. · xyz:CRWV short 42,250, +$849,447. xyz:AMZN short 14,090, +$206,766. · xyz:COPPER long 739,790, −$6,526 with −$21,691 in funding. $SOL long 3,270, −$9,252. 4) The fourth wallet sent its last 144,584.91 $HYPE to the trading wallet on 27 August at 20:44, 37 minutes after the unstaking was queued. $HYPE has run from $51.51 when the first short went on to $83.06, up 61%. He has responded by trimming 8% of the position, buying 163,840 coins in spot, and queuing the last of his delegation. Netting everything he holds 536,727 long against 605,580 short, so still net short by 68,853 coins.

HPC and trade[XYZ] filed a joint comment letter with the CFTC yesterday asking the Commission to open a regulated path for energy perpetual contracts. The setup: → A hedger holding WTI exposure has to close the expiring contract and open the next one, every single month → The roll spread is not a fixed cost, it moves hour to hour → CME closes over the weekend, so the window to actually do it is narrower than the calendar suggests → Large participants have to telegraph the trade, which moves the price against them inside that window What the chart shows: → Rolling a $10m position on Monday April 13 at 02:00 UTC cost roughly $945,000 → The identical roll on Friday April 17 at 20:00 UTC cost roughly $109,000 → Same mandatory trade, four days apart, $835,912 apart in cost → Crossing the spread either way cost about $2,500 So the timing of the roll was worth roughly three hundred times the execution of it. That cost never shows up in a fee schedule, never gets quoted as a spread, and is not something a hedger can plan around, because the number is only knowable after the fact. It is the cleanest empirical answer to the question of why anyone would want a contract that never expires: a perpetual position has no roll window to mistime.

The biggest AI narrative of the decade or a generational short? Reuters, citing The Wall Street Journal, reports that @AnthropicAI is likely to tell investors it sees a total addressable market above $30T And the valuation headline is getting bigger. The New York Times reports that Anthropic could seek to raise more than $100B in an IPO that could value it at $2T. If completed on those terms, it would exceed @SpaceX’s $1.77T June debut. From Anthropic’s confirmed $965B post-money Series H valuation in May to a reported potential $2T IPO: the valuation is growing fast. Want to express a view on that narrative?

Markets @Markets_xyz just opened a new F1 Singapore competition on the web terminal, and it ranks on ROI rather than raw size. That changes who can realistically win it: → The leaderboard is scored on return, not notional, so a smaller account trading well can finish ahead of size → $100k minimum volume to qualify → Runs from today until September 15 → Toggle Leaderboard → F1 tickets to see where you sit What's on the table: 1st: 2x F1 weekend tickets, around $2.5k in value 2nd: $500 3rd: $250 4th to 7th: $100 each 8th to 10th: $50 each Why I'm paying attention to this one: → I'm going to be in Singapore for that weekend anyway → So I might as well lock the fuck in and try to cover the tickets by trading

HYPE/USDC now sits at nearly $2B in open interest, second only to $BTC, and @HyperliquidX genuinely doesn't care what you trade, because it takes the commission either way. Hyperliquid's business model is decoupling from the crypto cycle. When crypto vol dies, traders don't withdraw, they rotate into gold, S&P500, SK Hynix, whatever is moving that week, and the collateral never leaves the account. Four of the top ten markets by OI are already non-crypto. A crypto winter used to mean a revenue winter for an exchange token. Here it mostly means a different ticker at the top of the OI table.

Builder revenue on @HyperliquidX just printed its best month Aggregate revenue: → 30-day total: $7.17M → Most of the month ran $100K–$300K a day; 21.08. had above $800K Top earners, last 7 days: → MetaMask: $507.8K on $537.30M volume, +135% → Phantom: $418.4K on $836.76M: more volume, less revenue, because it charges ~0.05% of notional vs MetaMask's ~0.095% → Trust Wallet: $286.9K on $489.52M, +224%; all-time of $696.8K means nearly all of it was earned this week → Non-wallet builders growing fastest: Invo $197.6K (copytrading), fomo $181.5K (+379%), BasedApp $85.1K (+2,606%) None of these apps built an exchange. They routed orders into one that already had the liquidity and turned an existing userbase into a revenue line.

Loracle sold 336k spot $HYPE on OTC? 1) The bag that sat untouched on HyperEVM since 14 August was bridged back to HyperCore on 19 August at 16:49, but not into any wallet we had been watching. It went to a fourth address, Two minutes later, at 16:51, 91,702 $HYPE went to and between 17:00 that evening and 09:11 the next morning that address wired $18,112,504 $USDC back to the main wallet across six sends. An OTC settlement, never touched the order book. At 22:11 another 100,000 $HYPE went straight to the trading wallet. 144,584.91 $HYPE still sits in the new address, and 371,767.54 stays delegated to Alphaticks with the lock long expired and nothing queued. 2) · $HYPE short 637,768 → 658,404. Entry averaged up to $53.8258 against a $77.71 mark: −$15,626,029, versus −$2,513,717 on 16 August. · Leverage raised 2x → 3x, liquidation moved from $78.42 to $106. · Account value $17.06M against $51.07M notional. · Spot $24.77M $USDC and 64,912 $HYPE. 3) · Spot swept to zero on 21 August, the $4,670,807 $USDC moved out. · Core book expanded from two positions to eight: $BTC long 21.27, $ETH short 662, $BNB long 2,381, $TRX long 4.6M, $PUMP short 53.5M, $ZEC short 2,676 (−$393,289), $PAXG long 209.63 (+$103,532). · A second $HYPE short of 21,369 opened here at entry $77.996, currently green. · xyz:GOLD long 12,802 oz now +$5,419,040. xyz:CRWV short expanded 16,997 → 42,249 (+$819,746). xyz:SNDK short grown to 15,567 (−$1,890,077), xyz:MU (−$904,855), xyz:COPPER long (−$77,767). $HYPE has run from $57.38 to $77.71, up 35% in a week and 50.9% since the first short went on. Loracle answered by adding another 20,636 coins and raising leverage instead of cutting. Netting all four wallets he now holds 580,864 long against 679,773 short, which puts him 98,909 coins net short for the first time in this whole sequence, at the highs, with the $18.1M of OTC proceeds funding the margin.

Hyperliquid Policy Center's new report tests whether perpetual futures help or hurt the dated futures markets they reference, using CME, Binance, BitMEX, and HIP-3 oil data. The answer across every test is that perps expand hedging access without measurably degrading the incumbent, though the oil evidence rests on just 19 weekends. → Weekend perp moves pass through to the CME reopen almost exactly one-for-one: Binance 1.0019 and BitMEX 0.9989 over 205 weekends, xyz:CL 1.0729 over 19. → The perp gave a better read on the reopen than a stale Friday close in 74% of weekends. → On the March 6, 2026 weekend, WTI gapped 15.8% at the reopen. A perp hedge would have cut $1.58M of exposure on $10M down to $62k after funding and fees. → The hedge benefit holds across the full sample and across every hedge ratio and cost assumption tested, from 1bp to 25bps. → Post-launch CME reopens were no worse than WTI's own history predicts. → A cross-market check against corn and wheat, neither of which has ever traded on HIP-3, shows no consistent sign of harm and flat pre-trends. → The funding rate anchors properly during closures: elevated premia mean-revert, and the final off-hours premium matches the reopen direction in 74% of weekends and near 100% of high-premium ones. → Median off-hours fill on xyz:CL runs near $1,300, roughly two orders of magnitude below a median CME WTI trade. → Rolling dated contracts is the hidden cost: the same April 2026 roll cost roughly $950k on Monday and $110k on Friday, on $10M notional. Perps never face that decision.

"I understand that [CFTC Chairman] Mike [Selig] is also working to bring Hyperliquid into the United States in a fully compliant and legal fashion, working very hard on that." Trump said that, but what it actually means? What happened At a White House meeting with crypto and traditional finance executives on August 19, Trump said CFTC Chair @ChairmanSelig is working to bring @HyperliquidX into the United States in a fully compliant and legal fashion. Markets read it as an endorsement, $HYPE ran to roughly $73.63, up 26% in a day, and $PURR rose as much as 32.4%. What exists on paper What does exist is precedent set for other firms. The load-bearing document is the May 29 order approving @Kalshi's BTCPERP, treating a perpetual as a futures contract, the first time a US regulator called perps futures rather than swaps. The companion policy statement routes every other asset class through case-by-case review, but it is a general statement of policy that imposes no obligation on anyone and modifies nothing in the statute. The July 9 letter from the @HyperliquidPC and @phantom asks the CFTC to confirm that a designated contract market can use an onchain protocol as its matching layer, that a clearinghouse can margin and settle onchain, and that a broker can accept customer orders and funds onchain. Under HIP-3 that means the licence wraps the deployer. The protocol stays unlicensed on the argument that publishing software is not operating an exchange. What Americans would touch is a gated venue settling on HyperCore, not which keeps geoblocking. How can Hyperliquid US look like? @shaundadevens outlined it perfectly in his article A permissioned HIP-3 deployment runs its own order book with its own allowlist and its own relisted tickers, but it clears on the same HyperCore substrate as the open market. Whitelisted market makers sit across both books, quote both sides, and hedge positions taken in the permissioned pool against the open one. The spread they capture is the incentive, and the effect of them working it is that the two books converge on one price. So the permissioned instance does not have to bootstrap its own liquidity. It borrows the open market's, and the wall stops being a commercial problem. Risks The mechanism delivering the price is a market maker hedging into an offshore book that no US regulator supervises. Citadel Securities and the traditional exchange lobby are already arguing that platforms should be judged by function rather than technology, and this is the diagram that argument was built for. Underneath all of it sits the order everything rests on. CME sued the CFTC and Selig personally in June, alleging he overrode congressional direction and circumvented the regime required for that type of derivative. If perps are ruled swaps, margin and registration requirements push the product back offshore and the path closes before anyone gets to argue about market structure.

The President named @HyperliquidX from the White House podium, and the last structural argument against $HYPE disappeared in a single sentence. Trump confirmed on August 19 that CFTC Chair Mike Selig is working to bring Hyperliquid into the US in a fully compliant fashion. $HYPE rallied 20%, rest is history. · US access was the missing piece of the thesis. Hyperliquid was going to keep compounding globally regardless, but the "house of all finance" case needs the largest capital market on earth, and that door is now visibly opening. · The reflexive bid is switching back on. $PURR trades at a 1.14x mNAV premium again, which historically triggers equity issuance funding further $HYPE accumulation; the trailing year already carries roughly $1.7B across AF buybacks, ETF inflows and DAT purchases. · Volume has likely bottomed with the broader cycle. Bitcoin accounts for around 36% of Hyperliquid's lifetime volume, so the past year's compression tracked market-wide risk appetite rather than anything venue-specific. · The mix shifted before the monetization did. HIP-3 and Trade[XYZ] now carry 40-50% of total volume at deliberately thin take rates. · Clarity changes the calculus for builders, not only traders. The risk premium deployers priced into launching on Hyperliquid collapses once the largest jurisdiction offers a compliant path, which should expand the universe of listed markets rather than just the flow through existing ones. · AQAv2 adds a revenue line that doesn't depend on volume. Roughly $170M annualized from $USDC reserve yield translates to about 6.5K $HYPE bought daily, diversifying the base away from trading seasonality. · Valuation looks undemanding on any forward view. $HYPE trades at 22.6x circulating earnings on $946M of trailing operating net income, and the market still prices these protocols off trailing 30-day revenue rather than what the next two years actually set up. Read full report from @GLC_Research:

This should be some kind of social experiment In December 2025 I redeemed a @jumperapp perk for a SafePal wallet. It never arrived!!! Today I got the email telling me my order data was part of a breach. Lmao