Perps Just Turned Tokenized Stocks Into Working Collateral

RWA perps out-traded every crypto category on Hyperliquid for the first time, and Ondo shelved its blockchain plans to launch a TEE execution network where tokenized stocks serve as margin. The Open Money read: perps are the machinery converting tokenization into live capital.

Summary: In the week of July 13, perpetual futures on real-world assets did $25.1 billion in volume on Hyperliquid, 52 percent of the platform's $48.2 billion total and the first time RWA markets out-traded every crypto category on the venue combined. Single-stock contracts drove 61 percent of that flow, and Hyperliquid's RWA book by itself out-traded the combined crypto-perp volume of every other DEX.

Then on July 28, Ondo Finance shelved its long-planned tokenized-asset blockchain and launched the Ondo Network instead: an offchain matching engine running in trusted execution environments (TEEs), checked by decentralized attestors, settling to public chains.

Its anchor tenant is Ondo Perps, the first venue that accepts tokenized US equities and ETFs directly as margin for perpetual futures on stocks, indices, and commodities at up to 20x leverage.

The category behind both moves has been compounding all year: monthly RWA perp volume grew from roughly $85 billion in January to more than $470 billion in June, and RWA flow is now approaching Bitcoin's own perp volume on the largest venues.

Thesis: Two weeks ago this newsletter watched the rail deepen. Last week it watched a coordination bet die and noted that what grew on Base without subsidy was prediction markets, perps, and stablecoin payments.

This week completes the thought. The industry is consolidating around the two products with proven organic demand, perps and prediction markets, and the perp side just produced the clearest structural signal of the year.

Perpetual futures are the Intermediation machinery that converts tokenization from inventory into activity. A tokenized stock sitting in a wallet is a receipt: Settlement-layer progress, real but idle. The same token posted as margin on an open perp venue is working capital, priced around the clock, borrowed against, composable with everything else on the rail, and still in the holder's custody.

The week's second lesson is about how that openness arrives. Hyperliquid got there with maximum permissionlessness, builder-listed markets and public order books.

Ondo got there by gating exactly one thing, the matching engine, while keeping custody sovereign and settlement public.

The Open Money Lens

The framework reads infrastructure across three layers, Settlement, Intermediation, and Coordination, and five dimensions of openness: Permissionless Access, Transparent Verification, Programmable Logic, Composable Infrastructure, and Sovereign Custody.

July 12 put Robinhood and DTCC side by side and scored the gates each kept. July 19 read Lean Ethereum as a depth event, the Settlement row thickening in place. July 26 watched Base retreat from Coordination and noticed that every cell that advanced was a finance cell.

Read in sequence, the past two months have been a story about the Settlement row getting stronger and the Coordination row refusing to be built. This week the action moves to the row between them. Intermediation is where tokenized assets either become financial services or sit as decorative inventory, and perps are currently the service they're becoming.

Map the two events and they bracket the row. The Hyperliquid flip lands at Intermediation x Permissionless Access and Composable Infrastructure: HIP-3 markets get listed by stake rather than by committee, order books settle in public view, and the RWA share of the book got there with no incentive program attached (Transparent Verification rides along, since every position and liquidation is auditable in real time).

Ondo's launch spans two rows at once: Settlement x Sovereign Custody and Transparent Verification, since users keep their assets and state settles to public chains, joined to a deliberate constraint at Intermediation x Permissionless Access, since matching happens inside private hardware and US users stay geofenced out.

The flip on Hyperliquid

The numbers first. In the week of July 13 to 19, HIP-3 markets on Hyperliquid, the builder-deployed perps covering equities, indices, and commodities, did $25.1 billion against $48.2 billion for the platform overall, per figures an ARK Invest analyst circulated.

RWA perps have gone from a rounding error to the majority of the book on crypto's largest open derivatives venue in under a year, and the composition is as telling as the total: single-equity contracts alone were 61 percent of RWA flow. Traders on an open venue, given a full menu, are choosing Nvidia and SpaceX exposure over the coins the venue was built for.

The listing mechanism deserves the attention the volume is getting. Listing on Hyperliquid is a staking transaction: a builder posts HYPE and deploys the market, and the market lives or dies on flow.

So nobody decided tokenized equities should become the platform's biggest category. Traders decided, one funding interval at a time. This is what the framework means by composability doing the pulling.

Once a high-performance open rail exists, whatever people actually want to trade shows up and gets traded, and it turns out a large share of what they want to trade is the stuff whose home market closes at 4 pm Eastern.

The category context says this generalizes past one venue. The Block's data puts monthly RWA perp volume at about $85 billion in January and above $470 billion by June, a fivefold climb in six months, led by single stocks and pre-IPO names. By late July, RWA perp turnover was closing in on Bitcoin perp turnover on Hyperliquid and Binance.

Bitcoin perps have been crypto's deepest, most liquid market for a decade. A category drawing even with them is a change in what crypto's market structure is for.

There's a structural reason perps are the instrument doing this absorbing, and it's worth stating plainly: perps let open rails trade things they can't yet hold. A spot market in Apple requires someone to custody Apple shares and make the wrapper legally portable, which is slow, jurisdictional work.

A perp requires a price feed and margin. Synthetic exposure travels lighter than the underlying, so the trading came onchain years before the assets could, and the funding rate quietly became an open, continuous price for instruments whose official markets sleep 128 hours a week.

Ondo moves the gate

Ondo's announcement is the more architecturally interesting one because it's a public change of mind. Ondo Chain, announced in early 2025, was going to be a purpose-built blockchain for tokenized assets.

On July 28 the company called the Ondo Network its evolution: an offchain execution layer where matching runs inside trusted execution environments, a decentralized attestor set verifies that the engine ran honestly, and settlement lands on public chains. The company was explicit about the reasoning.

Institutional flow needs CEX-grade speed and pre-trade privacy, and public mempools give away both. So the binding constraint on institutional RWA trading turned out to live in execution, and Ondo redesigned exactly that layer while leaving custody with users and verification in public.

The anchor application is the point. Ondo Perps accepts tokenized US equities and ETFs, the instruments Ondo's own issuance business has been minting, directly as collateral for perpetual futures on stocks, indices, and commodities at up to 20x.

That closes the capital-efficiency gap that kept tokenized stocks decorative. Until now, a holder who wanted leveraged exposure had to sell the token and post stablecoins, which made the token a detour rather than an asset. Now the token itself is margin. The market noticed quickly: the venue crossed $50 million in open interest and $4 billion cumulative volume by July 24, then printed roughly $300 million in a single day by month's end with open interest near $70 million, inside its first month of public operation.

Read against the July 12 taxonomy, this is a new gate placement. Robinhood opened the rail and controlled the asset. DTCC kept the vault and opened the receipt. Ondo keeps custody sovereign, keeps settlement and state publicly verifiable, keeps the resulting positions composable, and gates the matching engine.

The gate sits at the narrowest point in the stack, and it's the one place where a credible engineering argument, rather than an incumbency argument, defends it.

Pre-trade privacy is a real market-structure requirement; every dark pool and block desk in TradFi exists because of it. The Open Money question is what meeting that requirement costs: whether users give up custody and whether the public gives up the ability to verify. Ondo's design answers no to both, which makes it the most surgical gate this newsletter has scored all year.

The precedent matters more than the venue. The framework has been arguing since spring that openness advances dimension by dimension, and that hybrid cells, open where it counts and constrained where regulation or market structure demands, are how regulated capital actually arrives.

Ondo just published the reference architecture for the Intermediation version of that trade, and did it by abandoning a chain, the maximal crypto-native ambition, for a component. Purpose-built execution that settles to shared public rails is a bet that the rails have won, and that the remaining competition happens in the layer above them.

The honest case against this read

The first objection is the obvious one: TEEs and geofences are gatekeeping, and calling a gate surgical is still calling it a gate. Private matching means trusting hardware vendors and an attestor set that is decentralized on a roadmap rather than in fact, and a US-excluding perimeter reproduces exactly the access hierarchy open money was supposed to dissolve.

The framework's answer is about trajectory and residue: custody, verification, and composability stay open through the gate, the gate is auditable at the settlement boundary, and the history of these constraints (geofenced DeFi frontends, permissioned pilot networks) is that they widen under competitive pressure rather than narrow. That's a probabilistic defense, and the watch item that would falsify it is written below.

The second objection is scale and fragility. Four hundred seventy billion a month is still a fraction of global futures turnover, single-stock flow is concentrated in a handful of hot names, and a hostile regulatory turn on synthetic equity exposure could cut the category off at the feed. All true.

What the objection has to explain away is velocity and breadth at once: a fivefold climb in six months, a category flip on the largest open venue with no incentive program, and the same demand showing up simultaneously on a permissionless book and a geofenced institutional one. Fragile things usually grow in one place.

The third objection comes from this newsletter's own archive, and it deserves the most care. Last week I discounted Base's perp growth as speculation with better rails, and two weeks before that I ran the autopsy on Robinhood Chain's memecoin churn.

Consistency demands the same discount here, and some of it applies: a 20x levered SpaceX perp is a casino product by any honest description. The difference is what the speculation is doing structurally. Memecoin churn prices nothing and leaves nothing behind. RWA perp flow builds continuous price discovery for assets that had none outside market hours, pulls tokenized instruments into use as margin, and gives issuers a reason to mint more of them.

Stablecoins started as casino chips for exchange arbitrage and became payment infrastructure; DEXs started as venues for token gambling and became the liquidity layer everything else composes with. Speculation is how open financial primitives bootstrap their liquidity, and the test of whether it hardens into infrastructure is whether non-speculative uses accumulate on top. Collateralized tokenized equities are that accumulation starting.

What to watch

The first read is persistence on Hyperliquid. One flipped week is a headline; a full quarter of RWA majority, with single-stock concentration spreading past the current handful of names into indices, commodities, and more of the equity list, would confirm the structural version of the story. A retreat to crypto-majority flow when the hot names cool would support the fragility objection.

The second is Ondo's attestor set. The whole defense of the surgical gate rests on verification staying real: who runs the attestors, how many exist, whether they're independent of Ondo, and whether the TEE design gets a public third-party audit. An attestor set that decentralizes on schedule validates the hybrid architecture. One that stays in-house turns the execution layer back into a trusted exchange with extra steps.

The third is the collateral menu race. Tokenized equities as margin is a one-venue feature this month. The signal that it's becoming market structure is other venues accepting tokenized Treasuries and equities as collateral, and issuers competing on whose token is accepted at the most venues. Margin eligibility is about to become the metric that decides which tokenized assets matter, the way index inclusion decides it in TradFi.

The fourth is the other half of the consolidation. Prediction markets are the second product crypto's venues are reorganizing around, they grew on Base without permission alongside perps, and they're converging on the same architecture of open settlement plus performant matching. How the two products end up sharing margin systems and venues deserves its own issue, and it will get one.

Strategic implications

For builders. Treat perps as a horizontal layer rather than a vertical product. Anything tokenized should be designed for margin eligibility from day one, which concretely means clean price feeds, deep enough spot or synthetic liquidity to liquidate against, and standard interfaces so a venue can accept the token as collateral without bespoke integration. On the venue side, Ondo just published the architectural split worth copying: put privacy and speed in the execution layer where institutions require them, and keep custody, settlement, and verification on public rails where users require them. Teams building either half of that sandwich are building for the flow that actually showed up this year.

For capital allocators. The metrics that matter are rotating. TVL and crypto-perp volume measured the last cycle; the venues capturing this one are the ones hosting both crypto-native and RWA flow, and the diligence questions are about category mix, collateral eligibility, and open interest durability rather than headline volume. The Hyperliquid flip and Ondo's first-month curve are the two data points to generalize into a screen. The archive's standing discount applies here too: perp revenue is cyclical, and the durable claim is on the collateral infrastructure underneath it, the issuance, custody, and margin plumbing that survives whichever direction the levered flow points.

For policymakers. The fastest-growing equity trading venues in the world now run around the clock, settle on public chains, and sit outside US jurisdiction because that's where the rules put them. The demand is measurable, the geofences are working as designed, and the result is that American assets trade continuously on infrastructure Americans are excluded from. The Ondo design shows what a supervisable version looks like: private execution with attested integrity, public settlement, user custody, auditable state. Clear rules for tokenized-collateral derivatives that accept that architecture would bring the flow inside the perimeter with better transparency than the offshore alternative offers. The current path exports the market and keeps the risk.

Sources

[1] Cointelegraph. "RWAs become Hyperliquid's largest trading category." July 2026. https://cointelegraph.com/news/hyperliquid-rwa-volume-crypto-trading-first-time

[2] ChainCatcher. "ARK Invest analyst: Hyperliquid enters the RWA era, HIP-3 trading volume surpasses crypto assets for the first time." July 2026. https://www.chaincatcher.com/en/article/2278044

[3] Yahoo Finance. "Hyperliquid RWA Volume Hits $25.1B, Overtakes Crypto Trading." July 2026. https://finance.yahoo.com/markets/crypto/articles/hyperliquid-rwa-volume-hits-25-142800915.html

[4] CoinDesk. "Ondo drops tokenized asset blockchain plans for private, high-speed trading network." July 28, 2026. https://www.coindesk.com/business/2026/07/28/ondo-drops-tokenized-asset-blockchain-plans-for-private-high-speed-trading-network

[5] The Block. "Ondo launches new execution network, calling it 'evolution' of Ondo Chain." July 2026. https://www.theblock.co/post/409792/ondo-launches-new-execution-network-calling-it-evolution-of-ondo-chain

[6] Yahoo Finance. "Ondo Perps Launches First Equity Perpetuals Platform With Tokenized Stock Collateral." July 2026. https://finance.yahoo.com/markets/options/articles/ondo-perps-launches-first-equity-161200015.html

[7] DefiLlama. "Ondo Perps Volume Stats & Charts." Accessed August 1, 2026. https://defillama.com/protocol/ondo-perps

[8] The Block. "Tokenized equity perps drive RWA trading boom to $470 billion monthly volume." July 2026. https://www.theblock.co/post/408961/tokenized-equity-perps-drive-rwa-trading-boom-to-470-billion-monthly-volume

[9] Cointelegraph (via TradingView). "RWA perpetual futures volume nears Bitcoin on Hyperliquid, Binance." July 2026. https://www.tradingview.com/news/cointelegraph:b14edd199094b:0-rwa-perpetual-futures-volume-nears-bitcoin-on-hyperliquid-binance/