Harvesting Retail
There is a trade in crypto that runs almost every hour of almost every day, and most of the people on the losing end of it think they are day trading.
A retail trader buys a perpetual future because they want leveraged exposure to a price going up. On the other side, a professional collects a funding fee for holding the short. The fee is small each time and persistent over time, and across the whole market it runs to a great deal of money moving in one direction, from the many to the few.
That flow was not designed. It was discovered. And to see why it exists you have to start with what the perpetual actually was, which is not a product so much as a way around a rule.
A way around the rule
The perpetual future is a regulatory arbitrage.
If you want to offer ordinary people fast, leveraged, directional exposure to a price, the regulated instrument for that is a future, and a future comes with a venue, a clearing house, margin rules, and a regulator who has opinions about who you can sell it to. The perpetual was the way to deliver the same economic experience without any of that. Strip off the on-chain wrapper if it is on-chain and it is close to a contract for difference: post a little margin, get amplified exposure, hold it as long as you can afford to. The one thing it adds over a contract for difference is that you sit on a public order book and join a bid or an offer, so it feels like trading on a venue rather than being a customer of a bookmaker. The feeling is the point.
But a perpetual has a problem a real future does not. A future has an expiry, and that expiry is what drags its price back to the underlying. A perpetual never expires, so something else has to do that job. That something is the funding rate, a periodic payment between longs and shorts engineered to tether the contract to spot. Almost everything that makes a perpetual strange, the funding mechanics, the eight-hourly settlements, the auto-deleveraging when the pot runs dry, is not a feature anyone set out to build. It is the running repair bill for not being a future.
And it worked, commercially. Perpetuals are now the overwhelming majority of all crypto trading activity, more than ninety percent of it, with over ninety trillion dollars traded across the major venues in 2025 (https://www.coingecko.com/research/publications/state-of-crypto-perpetuals-report-2026). No regulated instrument got near that crowd, because the regulated instruments were not allowed to.
The respectable label
It is worth being honest about the demand the workaround served.
A large part of the retail crowd wants to gamble. That is a description, not a judgement. People want leveraged, fast, directional exposure with the chance of a large pay-out, and they have wanted it for as long as markets have existed. What changes is the label that makes it acceptable. Today the label is crypto, and increasingly event speculation. Both let someone take a leveraged punt while telling themselves, and their friends, that they are trading or investing rather than betting. The perpetual is the instrument that fits the label. It dresses an old appetite in the clothes of a market.
Who the crowd does not see
Three things sit underneath that volume that the average participant does not price.
The first is direction. Retail is, in aggregate and especially when prices are rising, almost entirely on the long side. You can see it in the account ratios the venues publish: far more individual accounts are long than short at almost any given moment (https://www.coinglass.com/LongShortRatio). The natural read is that the many small accounts are long and the few large accounts are short. This is a one-directional crowd treating a leveraged derivative as if it were a holding in a security.
The second is who stands opposite them. Market makers are structurally short the perp. As retail leaned in over time, professional traders worked out what that persistent one-sided demand was worth, and the basis trade became one of the most reliable sources of yield in the asset class. The clearest example at scale is the kind of delta-neutral structure that protocols like Ethena run, built specifically to harvest perpetual funding, now operating at several billion dollars of size (https://docs.ethena.fi/solution-overview/usde-overview). The important point is what this does to the market’s behaviour. The party who would normally compete a distortion away is here to feed on it, not to fight it. They have no reason to push funding back toward fair, because the gap is the yield.
The third is the funding rate itself, which most holders treat as a small fee and almost none treat as the main event. Over time, crypto funding is paid by longs to shorts far more often than the reverse, and the average carry runs well into double digits annualised, with episodes far higher (https://www.bis.org/publ/work1087.pdf). A retail trader holding a long perpetual for weeks is not paying a rounding error. They are paying a structural toll to the short on the other side. That is the harvest, and it needs no manipulation at all. It is the ordinary mechanics of the instrument meeting a one-sided crowd, in a market where the natural arbitrageur is collecting rather than correcting.
When the toll is not enough
If the ordinary funding flow is the patient version of the harvest, there is a faster version on thinner contracts, and it is worth working through precisely, because the detail is where the argument is won or lost.
On a venue like Hyperliquid, the funding rate is built from a premium: how far the contract’s own order book sits from an external oracle price, where that oracle is a median of spot prices from several large exchanges (https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding). The premium is not measured off the last trade. It is measured off the average price to trade a fixed, and surprisingly small, amount of the book. For everything other than Bitcoin and Ether, that amount is six thousand dollars (https://hyperliquid.gitbook.io/hyperliquid-docs/trading/contract-specifications). Six thousand dollars is the entire slice of the order book whose average price sets the premium that sets the funding rate.
Now the structure becomes visible. A participant accumulates a large position over time on the side that will receive funding, built slowly so as not to move the market. Then, around the funding window, they lean on that six thousand dollar slice using separate wallets, one pushing the price away from the oracle, another taking the other side, so the trading wallets stay near flat and only slippage is spent. They do not need to move the whole market. They need to hold the average price of a small slice of book away from the oracle for the period over which the premium is sampled. The funding is then paid on the whole accumulated position, which can be many multiples of the capital it took to lean on the book. That asymmetry, a large position earning on a small and cheap distortion, is the entire game.
The rate is capped, and the cap is instructive rather than protective. Hyperliquid caps funding at four percent per hour (https://hyperliquid.gitbook.io/hyperliquid-docs/trading/funding). On a contract priced near two dollars, four percent is about eight cents. The cap does not stop the squeeze. It tells the attacker exactly how far to push and no further, because anything beyond it is wasted slippage. And you rarely need anywhere near the cap. A small premium, held every hour, on a position far larger than the cost of holding it, compounds into a real number.
You can see the milder version of this in live data. Across the thin end of one venue’s contracts over the past week, most mid-sized markets behaved well, with the premium staying inside a fraction of a percent. But on the genuinely thin names the book detaches from the oracle and stays detached. One sub two-hundred-thousand-dollar contract ran its premium roughly ten times above its own baseline for two consecutive funding hours before reverting (Hyperliquid funding history, observed 9 June 2026). I would not call that proof of a deliberate squeeze, and I am not going to. Funding data alone cannot separate a push from a genuine one-sided sell-off on a thin venue. But it shows the thing the structure permits: on a thin book, the contract price comes away from the wider market, and the funding moves with it, for hours. And it is not expensive of difficult to do.
The oracle is supposed to be the defence, and on liquid contracts it can be, except when volatility hits. But an oracle is only as robust as the markets it reads. The two most expensive incidents in this space did not game the funding snapshot at all. They moved the price the oracle was reading. Mango Markets in 2022 moved thin spot venues to inflate a collateral mark and borrow against it, taking around one hundred and ten million dollars (https://www.cftc.gov/PressRoom/PressReleases/8647-23). The Hyperliquid JELLY episode in March 2025 forced a loss onto the venue’s own market-making vault through a thinly traded mark, and the venue resolved it only by overriding its own oracle by validator vote (https://oakresearch.io/en/analyses/investigations/hyperliquid-jelly-attack-context-vulnerability-team-solution). The lesson is not that any specific contract is rigged. It is that price integrity on these venues rests on the liquidity of the underlying spot, and where that liquidity is thin, the mark can be moved.
What it actually shows
None of this means retail should be locked out. It means close to the opposite.
The appetite is real and it is not going away. What perpetuals reveal is a large population that wants accessible, leveraged exposure to prices and is willing to pay for it. The precautionary instinct, that you protect people by denying them the product, does not work, because they find a route. They always do. The product simply moves to wherever the rules are loosest, and the profit moves with it.
There is a better destination for that appetite than the one we have built. A regulated capital market is a fairer place to take a leveraged position than a gambling product, for one plain reason: the underlying usually has value. A share is a claim on a business. A bond is a claim on cash flows. A future on either converges to something real on a date you can see coming. A perpetual on a memecoin converges to nothing, on no schedule, while charging you to hold it. If the goal is genuinely to protect people, steering the appetite toward instruments with real underlying value does more than any ban.
The access we actually block
Which makes our current settlement strange.
A retail trader in the United Kingdom or Europe can open an offshore account in minutes and run a hundred times leverage on a perpetual by tonight. The same trader has a complex process to invest their speculative capital in securities in their own country. The same trader cannot easily buy a plain American index fund, because it does not carry the right European disclosure document (https://www.justetf.com/en/news/etf/us-domiciled-etfs.html). They can trade thirty-to-one leverage on a currency pair through a regulated onshore broker, where most retail accounts lose money (https://www.fca.org.uk/news/press-releases/fca-confirms-permanent-restrictions-sale-cfds-and-cfd-options-retail-consumers), but they cannot hold the US-listed fund that simply tracks the S&P 500. It is easier to reach an offshore prediction market or a perpetual venue than to open an onshore stock account let alone a futures account. Let us not pretend the prediction market is anything more than a VPN away (https://docs.polymarket.com/polymarket-learn/FAQ/geoblocking).
The participation numbers tell the same story. In the United States, around three in five households hold equities in some form (https://www.federalreserve.gov/publications/files/scf23.pdf). In the United Kingdom, roughly a third of adults hold any investment at all (https://www.fca.org.uk/publication/financial-lives), and direct individual ownership of the domestic market has fallen for decades (https://www.ons.gov.uk/economy/investmentspensionsandtrusts/bulletins/ownershipofukquotedshares/2024). Meanwhile the London market has shrunk to around fifteen hundred listed companies, with 2024 the worst year for new listings on record (https://www.investmentweek.co.uk/news/4392463/london-stock-exchange-suffers-lowest-ipo-volume-record-2024). We have built a system where the easy, frictionless, encouraged path for a curious saver runs through the most extractive products available, and the hard, gated, paperwork-heavy path runs through the instruments with real underlying value. When a neobroker will sell fractional American shares in a tap but the domestic exchange feels out of reach, the incentives point exactly the wrong way.
John Doerr’s warning, that you should be careful what you measure because you will get it, applies to policy as much as to companies (https://www.supersummary.com/measure-what-matters/summary/). We measured whether we had banned the dangerous product onshore. We achieved that. We did not measure whether the activity stopped, and it did not. It moved offshore, and the profit and the tax base went with it. We did not reduce the gambling. We exported it.
Why the wholesale market never adopted them
There is a deeper reason perpetuals stayed a retail instrument and never became serious infrastructure, and it is the same reason they extract so efficiently.
A perpetual is a poor hedge, and again it comes back to not being a future. A genuine hedger needs a known cost and a known endpoint. A dated future gives both: a spread you can see, and convergence to the underlying on a fixed date. A perpetual gives neither. The funding cost is uncertain, it can be moved against you, and there is no convergence date to anchor it. So the institutions that actually need to hedge size do not reach for perpetuals. They reach for options, most of which trade bilaterally rather than on a central venue, and for dated futures where listed ones exist. The depth of the options market and the open interest on listed futures is the tell: that is where hedging lives, precisely because perpetuals cannot do the job.
The hedge also fails to track at size. A position large enough to matter cannot be carried in a perpetual without the funding uncertainty swamping the thing it was meant to protect. That breaks the one use case that would have brought diverse, two-sided, professional participation into the market, the participation that would have tightened it. Without real hedgers you are left with leveraged longs on one side and funding harvesters on the other, which is exactly the one-directional structure we started with.
Then there is margin. Because a perpetual cannot be netted cleanly into a portfolio the way a properly cleared instrument can, it ends up individually collateralised at a high margin rate, and the market has normalised this. It treats heavy, position-by-position margin as simply the cost of trading these things. That is compensating for a poor design with a high price, and retail pays it because it is hungry enough for the access to accept the terms. Perpetuals also behave badly inside a portfolio: the funding adds a volatile, path-dependent cost, and the squeezability we walked through earlier means they can demand more margin precisely when you can least afford it. An instrument that needs more capital per unit of risk than the alternatives is not an advance. It is a step backwards in new clothes.
The pivot
So perpetuals were a workaround, not a breakthrough.
They were a way to sell leveraged directional exposure to people the regulated system had decided to keep at arm’s length, and everything clever that came afterwards was repair work on the consequences of dodging the rule rather than following it. The technology around them is genuinely new. The instrument itself is old wine, and not especially good wine.
The honest response is not another ban. It is to broaden access to the better instruments and put the protection where it belongs. Retail traders in the United States can already trade options on individual shares, including the meme-stock favourites, and they can use dated futures. None of that ended the world. The precautionary case for locking British and European retail out of their own markets does not survive contact with the fact that the same people are one click from a hundred-times offshore perpetual. Give them access to their own stocks as easily as they can reach American ones. Make leverage products fairer and more transparent, with real in-app caution rather than a blanket prohibition that just relocates the harm.
The technology is real, the use case is narrower than the hype
It is worth being precise about what the underlying technology actually solves, because the perpetual story tends to oversell it.
Fast collateral movement, whether through stablecoins or tokenised real-world assets, and near-instant settlement, are genuinely useful. But across most of the payments world they remain a niche. The biggest friction in payments is not settlement speed. It is identity: know-your-customer and onboarding. Stablecoins add real value in cross-border corridors where correspondent banking is slow and expensive and where modern identity rails are not yet in place, which is why the global average cost of sending a remittance still sits well above the target the official sector set for it (https://www.worldbank.org/en/news/press-release/2024/06/26/remittances-slowed-in-2023-expected-to-grow-faster-in-2024). They add far less in domestic markets that already settle in seconds at near-zero cost through instant rails. And the headline volumes mislead: most of the trillions in reported stablecoin transfers are not organic payments at all (https://www.mckinsey.com/industries/financial-services/our-insights/stablecoins-in-payments-what-the-raw-transaction-numbers-miss). The technology is real. The claim that it has already remade payments is not.
Where it genuinely matters is when you combine the pieces: a real need, flexible product structures, faster settlement, and better risk management, brought together in the one place they would compound. That place is not retail speculation. It is the plumbing of the capital markets themselves.
What we should be building
Which brings me to the part that actually matters, and to a deadline most people are not watching.
Since 2008, the system’s answer to almost every kind of uncertainty has been the same: lock up more capital. Post more margin, hold more collateral, raise more buffers. The instinct is understandable and in many places it was right. But it has a cost that rarely gets counted. The industry now holds record amounts of margin against its derivatives, well over a trillion dollars of it sitting at the clearing houses alone, and the totals keep climbing (https://www.isda.org/2026/04/29/isda-margin-survey-shows-leading-derivatives-firms-collected-record-1-6-trillion-of-margin-in-2025). When stress hits, the demands spike violently: in March 2020 the daily variation margin calls across the major clearing houses rose more than fivefold in a fortnight (https://www.bis.org/press/p220929.htm). Capital locked as collateral is capital doing nothing else. It does not fund a business, build a project, or take a risk that grows the economy.
And it concentrates who wins. When the answer to every uncertainty is more trapped capital, the people who can post it, and the investors who can get out early, capture most of the upside. Growth funnels toward whoever can demonstrate the retail appetite to absorb the risk and hold the balance sheet to fund the margin. That is a narrow set of winners.
The answer is not an instrument that locks up even more capital per unit of risk, which is precisely what a perpetual does. It is to rebuild the plumbing so that less capital is trapped for the same amount of safety. Bigger netting sets, so offsetting risks cancel instead of being collateralised twice over, the way multilateral netting at the large clearing houses already saves tens of billions against what gross margining would demand (https://www.greenwich.com/market-structure-technology/portfolio-margining-imperative-interest-rate-derivatives). The qualifying clearing house treatment, which drops the capital cost of a cleared trade from a punitive risk weight to a small one (https://www.bis.org/basel_framework/chapter/CRE/54.htm), and which the consultation closing on 18 June will help settle for cleared trades. Real-time risk management that frees capital as fast as conditions change, rather than holding it hostage to a once-a-day calculation.
Alongside the wholesale rebuild, the retail fix follows the same logic: fairer, more accessible markets, with leverage available under sensible caution rather than driven offshore, and an identity layer that is portable and codified, so that every firm stops re-running the same checks from scratch.
The biggest lever we have on growth is not a new way to let retail bet. It is unlocking the capital trapped across the entire system. Perpetuals are a symptom of the problem, an instrument that found a profitable way to extract from a crowd the regulated market pushed away. The work worth doing is the opposite of extraction. It is rebuilding the infrastructure so that capital is free to do something useful again.
I guess the question is what do I know. Here is my 2013 Patent on “Constant Maturity Futures” Perps, https://patents.google.com/patent/US20150026028 and here is the Eurex announcement for their listing: https://www.eurex.com/ec-en/find/news/Eurex-and-GMEX-cooperate-in-swap-futures-trading-and-clearing-160824 . They didn’t work, because the issues laid out above were seen by the buyside traders. These issues have not changed.
Having said all of that, as we build DCN we have a clear everything philosophy, we will support perps. Our view though is that better access to more instruments for the market is a superior solution.
Thanks for reading. James.
This first appeared in The Clearing Gap, my LinkedIn newsletter, on 11 June 2026. If you want to comment, discuss, or subscribe, that’s the place.