The real question is not whether MEV exists. It is who you are handing the pricing power over ordering to.
In this essay, Jascha Samadi makes a blunt argument: MEV cannot be designed away. Any ledger that executes transactions against valuable shared state will create economic value in execution order. That value may be auctioned, hidden, burned, or renamed. It does not disappear.
Samadi frames the point like a law of physics. Energy cannot be destroyed; it only changes form. Ordering value, he argues, behaves the same way. Ethereum turned it into an open market. Permissioned networks such as Canton moved it behind contracts and confidentiality arrangements. Traditional finance spent decades wrapping it in labels such as payment for order flow and last look. Hyperliquid has recently started burning it. For investors and operators, the key issue is not the presence or absence of MEV, but whether ordering power is visible, who captures the rent, and whether any of that value flows back to the market that created it.
The piece extends the authors’ broader block-building work, focused on the path a transaction takes from a wallet click to its final place in a block, and on the market structure that forms around that path. Samadi tracks the conservation of ordering value across four regimes: open markets, private ledgers, regulated intermediaries, and public auctions. His bottom line is simple. The meaningful divide is not “MEV versus no MEV.” It is whether the price of ordering is public or hidden.
Ethereum: an open market makes ordering value visible
An open, permissionless blockchain rests on a few core design choices. Anyone can run a node and independently verify the network’s history. Anyone can submit a transaction. Participation in consensus, meaning the right to produce blocks and extend the chain, is open and gained through capital or work rather than granted by a gatekeeper. Underneath all of it sits the decisive choice: every participant replicates the same ledger.
That shared global state is what lets public chains become more than ordinary distributed databases. Because every application lives inside the same state machine, any contract can read from and synchronously compose with any other contract. A lending protocol can consume an oracle it never coordinated with, accept collateral tokens issued by a third party, and be wrapped by an aggregator that no one had to approve, all inside one atomic transaction. Samadi says this permissionless composability over commonly verifiable state is the source of much of what is genuinely new in onchain finance.
The same design also creates a less celebrated consequence. If everyone shares the same state, transactions must be applied to that state in some order, and because many transactions touch the same state, order changes outcomes. Two swaps against the same liquidity pool are not interchangeable. The one ordered first gets the better price. A liquidation is available to only one caller, the one that arrives first. An arbitrage window exists for a single transaction, the one that gets ahead of the rest. Position in sequence is access to an economic opportunity, and scarce access with economic value does not stay unpriced for long.
Ethereum is Samadi’s clearest example of what happens once that value becomes legible. An industrial transaction supply chain has assembled itself around the path from a user wallet to the final block: searchers, builders, relays, and validators. Searchers scan the public mempool and chain state for extractable opportunities, including cross-pool arbitrage, liquidations, and at the more predatory end, sandwich trades around visible user orders. Builders assemble candidate blocks, package and order transactions and bundles, and maximize total block value. Under proposer-builder separation, or PBS, validators no longer build blocks themselves. They auction that right through relays to the highest-paying builder. Further upstream, order flow itself has become an asset that can be sold, with wallets and applications selling the right to interact with user transactions.
This machinery is often described as parasitic, and parts of it may be. Samadi’s more precise description is structural: PBS is what a market for a scarce resource looks like once the resource is acknowledged. Ordering was always valuable. Ethereum’s architecture made that value visible, contestable, and measurable.
The research agenda that followed, order-flow auctions, MEV rebates that return extraction to the users whose transactions create it, encrypted mempools, batch auctions, is not an attempt to deny the phenomenon. It is an attempt to govern it. The live question is who captures ordering value and how much of it returns to users.
Canton: what disappears is the market, not the discretion
Samadi then turns to permissioned networks, focusing on Canton. In his telling, Canton reverses nearly every major design choice of public chains, deliberately and coherently. There is no global state and no global transparency. The system is built as a network of networks. Each participant runs a validator and holds only the contracts to which it is a party. Data is distributed strictly on a need-to-know basis rather than broadcast network-wide.
This privacy model extends inside a single transaction. With sub-transaction privacy, even a counterparty to one branch of an atomic transaction cannot see the others. In the example given in the article, a bank handling the bond leg of a delivery-versus-payment settlement would not learn the commission terms in another branch of the same atomic submission. There is no public mempool. Pending transactions are not broadcast, only delivered to relevant stakeholders. The ordering layer is blind: synchronizer operators order encrypted confirmations without seeing the underlying transaction content.
Samadi says Canton’s position on MEV follows almost mechanically from those choices and rests on four mechanisms:
- There is no public mempool to snipe, because pending order flow is never openly broadcast.
- Ordering is blind, so operators cannot front-run what they cannot read.
- There is no market for ordering, no builder or searcher auctions, no priority-fee bidding for position, and no PBS supply chain. Ordering is treated as a utility function carried out by identified institutions under contract and legal liability.
- Ordering rules live at the application layer, where exchange-style fairness rules such as price-time priority can be enforced by construction rather than left to validator behavior.
By the adoption numbers in the article, that design is not a niche experiment. Broadridge’s distributed-ledger repo platform settles about $8 trillion to $9 trillion in monthly repo volume on Canton infrastructure, roughly $354 billion a day, which the piece describes as the largest tokenized real-world asset settlement platform currently in operation. Public tracking data cited in the article shows around $345 billion in represented asset value on the network, with claimed monthly throughput close to $9 trillion.
Samadi argues that the fit with its flagship use cases is real. Interbank repo and collateral markets value privacy, permissioning, and recourse-backed finality more than censorship resistance. No bank is going to broadcast its funding position into a public mempool, and no regulator is likely to license critical market infrastructure where a software failure could mean irreversible loss.
But he is careful about what this does and does not prove. The structural observation is that someone is still ordering transactions. Ordering discretion has not vanished. It has moved. On the public layer, it sits with synchronizer operators. On private synchronizers, it can sit with a single company, and Samadi says the large majority of today’s volume is actually settled in those environments.
That means the discretion still exists. It is simply non-competitive and difficult to observe, constrained by confidentiality agreements and terms of service rather than removed by cryptography. He draws a line between two statements that are often blurred together: “there is no permissionless MEV market,” and “there is no public evidence that anyone is extracting MEV.” Those are not the same claim, and much of market-structure history sits in the gap between them.
He offers a second explanation for why observed MEV on Canton appears close to zero. The answer may be the use case as much as the design. Canton’s flagship workflows, bilateral repos and pre-arranged delivery-versus-payment settlement, involve little contested shared state. In a repo where price, size, and counterparty are fixed before the transaction reaches the ledger, there is not much to sandwich. Ordering a queue of pre-negotiated settlements does not create economic value on any chain under any design.
That is why Samadi says the current record is better described as “not falsified” than “proven.” The real test comes if a network like this begins carrying genuinely competitive market structure: shared central limit order books, liquidation engines, oracle-triggered margin calls, or tokenized Treasury markets traded at scale. At that point, arrival order at the synchronizer becomes economically decisive again, and those who can capture it will be the participants with latency advantages, physical proximity, or relationships with operators.
The rent will not disappear. It will migrate into privileged and opaque channels: preferred connections that never appear on a fee schedule, or information adjacency that is never formally listed anywhere. In Samadi’s reading, that architecture reintroduces trusted intermediaries with ordering discretion at the ledger level and rebuilds the trust model of traditional finance.
Traditional finance: the same rent has been hidden for a century
To show how that trust model behaves in practice, Samadi moves to traditional finance. His argument is that regulated markets have run the trusted-intermediary-with-ordering-discretion model for decades, and the enforcement record does not show that ordering value was eliminated. It shows that opacity made it profitable to capture.
Structurally, any venue that combines valuable shared state with sequential execution will generate MEV-like incentives. Traditional finance did not remove that property. It prohibited some forms, regulated others imperfectly, and pushed extraction into channels that victims could not observe in real time.
The article runs through several enforcement examples. Payment for order flow turned retail orders themselves into a monetizable asset. In 2021 alone, US brokers collected about $3.8 billion in payment for order flow, the value of seeing and internalizing that order flow, captured through bilateral arrangements between brokers and wholesalers that were largely invisible to the customers whose orders were being sold. Robinhood paid a $65 million SEC settlement in 2020 over misstatements tied to execution quality in that arrangement.
Dark pools were marketed as protection for institutions against predatory flow, yet the record described here runs in the opposite direction. Barclays paid $70 million in 2016, and Credit Suisse paid $84.5 million, over misstatements about how their dark pools operated, including cases where high-frequency flow that clients were told had been excluded was in fact participating. ITG paid $20.3 million in 2015 because it operated an internal desk inside its own dark pool and traded using confidential customer order information.
In foreign exchange, Barclays paid $150 million in 2015 over its use of last look, a dealer’s option to reject a trade after seeing which way prices moved, and over coordinated sharing of customer order information ahead of benchmark fixes. Samadi writes that industry fines for this kind of conduct have exceeded $10 billion in aggregate.
The pattern is consistent across the cases. Ordering value is real. Extraction is carried out by trusted intermediaries sitting in the sequence. The affected party cannot see it in real time. The facts emerge only after the rent has been collected for years, often through whistleblowers and enforcement. In that sense, when “MEV-free by design” is implemented through this kind of operating model, what it often means in practice is that ordering rents are adjudicated by compliance teams and discovered by regulators after the fact, not priced openly by the market.
Hyperliquid: public auction, then burn
If traditional finance represents hidden allocation, Samadi presents Hyperliquid as a far more transparent alternative. The same structural rent is moved into a public auction, and the proceeds are directed into a public pool.
Hyperliquid’s architecture has no public mempool and handles ordering at the consensus layer, which is why it has long been described as resistant to MEV. By mempool-extraction standards, Samadi says, that description is fair. But from the structural perspective, execution advantage still exists because it always exists. Priority goes to participants who engineer lower latency. Mature market makers win queue position through infrastructure optimization and network proximity, which is functionally close to colocation. That advantage has existed for a long time, carries real economic weight, and is only available to firms with the capital and access needed to build it.
In April of this year, Hyperliquid formalized that dynamic. It introduced priority fees, denominated in $HYPE, through a public, continuously repeated Dutch auction. What gets sold is exactly what low-latency firms were already buying in practice: earlier access to incoming transaction data and earlier position in the execution queue.
The effect can be measured with unusual precision. According to the article, every basis point spent on priority fees improves end-to-end latency by about 45 milliseconds. An advantage that used to be implicit, private, and capped by engineering capability became explicit, public, continuously repriced, and open to all participants.
Samadi separates two properties of this mechanism. The first is that it formalizes a paid priority tier, and that can plainly be criticized. A visible tax on queue position now exists where it did not before. But the relevant comparison is not a market without priority tiers. It is the same priority tier when it was being allocated invisibly through capital expenditure and connectivity, while appearing to cost nothing. In his view, an advantage that is public and priced is more honest, and easier to measure, debate, cap, and redesign, than one that is hidden and free.
The second property, and the one he sees as more novel in market-structure terms, is where the revenue goes. These fees are burned. The clearing price for execution priority is not paid to the venue operator, not shared with a favored counterparty, and not embedded inside a bilateral contract. It is removed from supply. Economically, Samadi says, that is equivalent to distributing the value pro rata to every holder of the network asset, $HYPE.
The contrast with traditional markets is direct. Colocation fees accrue to exchange shareholders. Payment for order flow is split between wholesalers and brokers through contracts end clients never see. The information rent in dark pools and last look goes to whoever has the privileged seat until enforcement claws part of it back. Hyperliquid puts the same structural rent, the value of position in sequence, out in the open, lets a permissionless auction discover the price, and sends the proceeds to a public pool.
The queue remains. What changes is that the queue has a public price, and the economic benefit is shared across holders rather than settled inside closed bilateral structures. Samadi is not claiming that ordering value has been designed away. He is claiming the opposite: it cannot be removed, so the honest move is to price it openly and redistribute the proceeds.
The distinction is whether the price is public
Samadi closes on a single organizing point. The meaningful comparison across chain designs is not “MEV or no MEV.” Every venue that orders valuable state transitions will, as a matter of structure, produce ordering value.
The sharper questions are these: who holds the ordering power, is that power contestable or assigned, is it observable or hidden, and are the rents priced and redistributed in an open market or quietly captured by whoever sits closest to the sequencer.
Permissioned networks answer those questions through governance, legal accountability, and institutional guarantees. For pre-negotiated settlement flows, that answer may be sufficient, and the adoption data reflects that. For genuinely competitive markets, Samadi says the evidence from crypto’s short history and traditional finance’s long history points the same way. Ordering value that is denied tends to be captured in the dark. Ordering value that is acknowledged can be priced, audited, and returned to the market that creates it.
In that sense, ordering value is conserved across designs. The difference is whether its price is public.
