From public AMM to PropAMM, professional market makers take over on-chain transactions, with bidding shifting to private.

Written by: Andjela Radmilac

Translated by: Saoirse, Foresight News

Open Jupiter, choose to exchange SOL for USDC, view the quotes and click to exchange. To the user, this is all there is to the operation.

Therefore, it is natural for people to assume that the trading is completed on the Jupiter platform. But in reality, Jupiter functions more like a liquidity search engine. The software traverses various channels willing to take on counterparty trades, compares quotes from each side, and routes orders to the optimal path. Its existing system can choose from ordinary decentralized exchanges, proprietary automated market makers (PropAMM), and a quote request network - where professional market makers compete to provide trading quotes.

The user sees a unified interface and a single quote, but underneath, there are various types of markets vying for this order.

The counterparty selling you USDC may not be an ordinary user, nor a public liquidity pool made up of many average people's funds. Instead, it could be a professional trading firm using its own capital and proprietary pricing software. Such firms do not have exchanges aimed at ordinary users, nor are there liquidity pools for ordinary people to deposit tokens into. The entire complex mechanism is completely invisible to the user; you can only see the final quote given.

A new analysis from DWF Ventures estimates that PropAMM accounts for approximately 15% to 27% of the total daily on-chain DEX trading volume. In a niche market - the SOL to stablecoin transactions routed through Jupiter, DWF estimates the market share of PropAMM exceeds 90%.

The share of PropAMMs has climbed, reaching as high as 27% of weekly on-chain spot trading volume (source: DWF Ventures)

DWF Ventures' data comes with an apparent limitation: it belongs to DWF Labs, which is itself a crypto market maker and investment institution. However, compared to an exact percentage, the phenomenon reflected by the numbers is more noteworthy: a form of trading that does not exist in the perception of most people has developed to a scale that can change the operation of some high liquidity markets.

This has brought the crypto market to a very delicate situation. The original intention behind the design of the first generation of decentralized exchanges was to replace professional traders with open code, making the market more open. Anyone can view the liquidity pool, deposit funds, and complete transactions according to the rules preset by the code. However, now professional traders are returning, and in certain high liquidity markets, their market-making capabilities are extremely competitive.

This itself represents a significant shift for DeFi. And as Nasdaq, the London Stock Exchange, Robinhood, and Kraken are all pushing to bring traditional stocks onto blockchain infrastructure, the implications of this shift will be even more profound.

For years, the financial sector has debated whether Wall Street will integrate with crypto infrastructure. But now, a more challenging question has arisen: if Wall Street does make the move, what kind of market awaits them?

Traders Return to the Market

The initial automated market maker (AMM) was a uniquely innovative feature within DeFi.

Imagine a currency exchange counter, with no traders behind it, placing SOL on one side and USDC on the other, and the software uses a formula to relate the two. You deposit SOL and receive USDC in return. As SOL accumulates and USDC diminishes, the formula automatically adjusts the exchange price. Anyone can inject funds into this counter and share a portion of the trading fees.

This solved a major problem in early DeFi: new markets do not need to laboriously persuade professional institutions to buy and sell incessantly. As long as there are tokens, smart contracts, and willing depositors, the market can operate automatically.

However, being able to operate automatically does not equate to high trading efficiency.

Assuming the price for SOL in the liquidity pool is $100, an unexpected news item comes out that sends the price at centralized exchanges to immediately change, with a professional trader valuing SOL at $101. The liquidity pool itself cannot perceive the price change, but arbitrageurs can spot the opportunity. They continue to buy the undervalued SOL until the price in the pool catches up with the external market. The liquidity pool effectively sells SOL at a low price, only to “learn” that SOL has already risen to $101.

Academics refer to this issue as rebalancing loss. The terminology is not important; the core idea is: a passive liquidity pool will repeatedly transact at outdated prices, trading with market participants who have the latest information. This flaw is acceptable when the market is stable, and trading fees can compensate liquidity providers. However, once prices fluctuate rapidly, outdated quotes will invite large-scale arbitrage, leading to substantial losses.

PropAMM disrupts this model. It is not reliant on large numbers of external users depositing tokens but rather on professional trading institutions using their own inventory assets; it does not wait for trading to push formulas to slowly adjust prices but instead relies on proprietary software to monitor external markets in real time, continuously updating buy and sell quotes. Trading is still settled on-chain, but the pricing logic is determined by the institution's internal system.

Rather than inheriting the early DeFi vending machine-like model, it is better to understand PropAMM as an electronic trader deployed within the blockchain.

This is also where the term "proprietary" comes from. According to official explanations from Solana, today’s PropAMMs are mostly closed-source; their liquidity pools are not open for ordinary users to deposit funds; to be integrated by an aggregator, they also need to obtain permission. Users can verify that trades occur on-chain, but cannot see the underlying systems that determine the quotes, nor can they see why this order is assigned to a specific market maker.

From the perspective of early DeFi ideals, this appears to be a regression. However, for ordinary users who simply want to quickly and cost-effectively exchange SOL, this matter does not need to be judged on an ideological level.

Jump Crypto analyzed approximately 20 million PropAMM transactions from March, comparing them with quotes from Binance, Coinbase, OKX, and Bybit. The study found that the trading prices for SOL-USDC transactions had a median difference from the centralized exchange midpoint of only 0.72 basis points; 91.9% of the transactions had costs less than what Jump estimated as the lowest fee tier at centralized exchanges for institutions.

In Jump Crypto's Solana sample, over 90% of PropAMM transactions had actual costs lower than the estimated costs of the lowest institutional rates at centralized exchanges (data source: DWF Ventures, based on Jump Crypto data)

Jump itself is also a trading institution, which introduces a commercial interest in this market structure, so this study carries some bias. Nevertheless, the results explain why such channels can secure orders: for certain trades, they can provide better execution prices, leading routing tools to direct traffic to them.

This not only changes the operational mechanics and economic logic of decentralized trading but also breaks the underlying concepts of early DeFi. The first generation of DeFi regarded openness itself as the core of the product: public liquidity pools, auditable code, and liquidity coming from ordinary users, meaning anyone proficient in technology could generally trace the link from funds to trading.

The new generation model raises a question closer to the ordinary consumer: if another system offers better quotes, do you need to fully understand all of its operational mechanisms? Once the answer is “not necessary,” the exchange itself will be hidden behind a simplified user interface.

DeFi is Transforming Exchanges into Routing Tools

The Solana ecosystem is particularly suited to this model because professional market makers are extremely wary of outdated quotes. When a firm quotes SOL at $100, it absolutely does not want its old quote to linger on-chain when the external market has already risen to $101. The longer it stays up, the more likely it is to be exploited for arbitrage based on outdated quotes.

The current block interval for Ethereum is 12 seconds. In contrast, Solana achieved a block time target of 300 milliseconds on August 28. The two networks build and process transactions differently, so they cannot be directly equated, but the allure is easy to understand: professional market makers want to update their quotes as frequently and cost-effectively as possible. Solana's research on shortening block intervals clearly indicates that this can reduce the risks of outdated quotes faced by liquidity providers and market makers.

However, there is another pathway to achieve this. It is not necessary to continuously upload all quotes to the blockchain; the system can obtain off-chain private quotes from professional institutions, only recording the final executed trades on-chain.

The request for quotes (RFQ) system inherently works this way. In the 0x RFQ network, applications request off-chain private quotes from professional market makers, compare these quotes to ordinary public AMM liquidity, and select the path that yields the best execution. 0x states that for high-volume trading pairs like USDC-WETH and WBTC-WETH, RFQ liquidity is superior to AMM in about 52% of the cases.

Jupiter has integrated various similar solutions on Solana. Public DEX liquidity, PropAMM, and professional quote systems are all encapsulated within the same interface. Traders no longer have to manually select which type of market to use before placing an order.

This could represent a deeper and more crucial shift: the old DeFi essentially told users, there is a liquidity pool here, you trade directly with it; while the new DeFi logic has become, tell me what trade you want to do, and I will help you find a counterparty for the order.

Traditional finance has long adopted this logic. When users submit stock buy orders, they typically do not specify the exact execution venue or broker, but rather leave the order routing to the broker. The basic explanation of the execution mechanism of US stocks by the SEC mentions that brokers will choose among exchanges, electronic networks, market makers, and internal matching to find the optimal execution channel.

PropAMM does not directly transfer Wall Street's dark pools to Solana; if understood this way, it overlooks the unique aspects of the crypto solution. The blockchain settlement process can be observed, assets transferred on public infrastructure, anyone can verify the final trade records. However, the mechanisms behind the trade can remain private: pricing models, inventory management logic, some routing information, and even the identities of institutions participating in the bidding can remain undisclosed before the execution.

This gives rise to an entirely new definition of transparency: financial markets can make asset ownership and settlement processes public for verification by all, but do not need to disclose all the details of pricing. As the assets flowing within the system are no longer limited to native crypto assets, this differentiation will become increasingly important.

As Wall Street Enters, DeFi is Shifting

Nasdaq announced on Thursday that its venture capital division agreed to invest $100 million in Payward, the parent company of Kraken, as the two continue to advance the Nasdaq equity token project. Nasdaq expects this framework to go live in the second quarter of 2027 and plans to integrate with Payward's xStocks network.

The London Stock Exchange is also laying out plans in the same direction. On September 1, the London Stock Exchange Group announced a collaboration with Payward to develop tokenized stocks; pending regulatory approval, they plan to launch xStocks trading on its planned LSE 24 platform in 2027.

This is no longer a mere pilot project. CryptoSlate has observed tokenized stocks reaching the trading volumes of the crypto market scale. The asset ownership issue, which is even more complex, has also evolved into a question of market structure - tokenized stocks appear similar to ordinary stocks, but the legal rights obtained by the buyer differ, as well as the underlying infrastructure.

Robinhood is an instance of this infrastructure. Robinhood Chain can already enable tokenized stocks to circulate as assets in an unlicensed crypto market, but the professional liquidity layer and open AMM experiments are quite different. 0x states that its stock token plan connects applications to professional market makers through RFQ, using USDG as the main trading pair, with Tokka Labs being the core market maker in this system.

This is precisely where the speculation about real-world assets in the DWF report is more important than the report itself: mature tokenized assets will not simply carry over the early DeFi market design just because they settle on a blockchain. Tokenized stocks of Apple already have mature reference markets elsewhere, with professional institutions capable of managing inventory, and traders prioritize execution quality. Moving stocks onto the blockchain will not automatically make passive public liquidity pools the preferred trading venues.

For years, discussions about tokenization have revolved around whether Wall Street assets will migrate to crypto infrastructure. But a more important question follows: once assets are on-chain, what market structure will they adopt?

For a long time, the positioning of these markets was to eliminate financial intermediaries, yet they still heavily rely on traditional financial infrastructure, including brokers and custodians who hold the underlying stocks. Currently, there is a brokerage that undertakes a significant part of the backing work for the underlying assets of tokenized stocks. The transformation brought by PropAMM corroborates this major trend from the transaction level: blockchain can change which parts of the financial system are open and transparent, but it will not make all intermediaries disappear into thin air.

Transaction Records Are Public, but the Trading Counter is Not

The simplest conclusion is: DeFi has spent years trying to break away from Wall Street, only to end up reconstructing the Wall Street model, but this statement is too one-sided. What the crypto market is genuinely doing is to separate two things that were once bound together: the transparency of settlement and the transparency of trade execution.

Blockchains can publicly prove that a transaction has occurred, completing asset transfers, and assets can circulate in programmable infrastructures without requiring the pricing algorithms of market makers to be disclosed. The settlement layer remains open, while the execution layer is contested among multiple private institutions.

There are economic considerations for doing so. Professional market makers can avoid outdated quotes and prevent arbitrage, thus not needing to factor potential losses into the quotes given to ordinary users. More complete information and rapid repricing capabilities often enable private traders to offer cheaper prices compared to completely public liquidity pools.

However, the cost is transferred elsewhere. If users no longer choose exchanges but instead use routing tools, the routing platforms gain greater market power. They decide which institutions can compete, which liquidity sources receive order assignments, what information is attached, how quotes are compared, and how trades are recorded on-chain.

Solana's official PropAMM documentation highlights a series of related issues: the code can be closed-source, aggregator access requires permission; when multiple institutions have similar quotes, users cannot independently verify why a specific market maker secured an order. The blockchain can only display the final executed trade, failing to reconstruct the complete bidding process prior to execution.

As market making, order routing, trade delivery, and block building begin to intertwine, this issue will become more pronounced. Blockworks Research studied HumidiFi, the trading delivery system Nozomi, and block builder Harmonic, all part of the Temporal ecosystem, finding that during certain time frames, HumidiFi's price update messages reached Harmonic's verification nodes earlier than those of other solutions, though subsequent advantages disappeared.

Compared to qualitative accusations, this matter resembles a prelude to the next round of disputes in the DeFi market structure: when multi-layer systems can all influence transaction outcomes, what truly counts as "optimal execution" on the blockchain?

Traditional finance has spent decades formulating relevant rules. Optimal execution, order routing, market access, quoting precedence, conflicts of interest between brokers and traders - these rules arose because when an investor clicks "buy," it is merely the start of the transaction. DeFi is simultaneously using new components to rebuild systems while re-experiencing this complex set of traditional financial rules.

The difference is that functions that used to concentrate within a single financial institution can now be split and deployed on public networks. This renders parts of the market more auditable while reducing transparency in others. This outcome cannot simply be judged as the market becoming more transparent or less transparent.

This does not imply that traditional AMMs will disappear. Their shortcomings are exactly the reason they can continue to exist.

The premise for professional market makers to leverage their advantages is that assets have reliable fair valuations. SOL has ample liquidity in other markets, Apple stocks have pricing from securities exchanges, and government bonds have reference market prices. Trading institutions can refer to external markets, estimate value, maintain inventory, and provide quotes.

Imagine a new token, Token X, birthed at three in the morning. It has no mature external market, no stable reference price, and no professional institutions willing to hold inventory. Before a mature professional market emerges, this asset first needs a tradable market. Public AMMs do not need to know the reasonable value of Token X; as long as someone deposits Token X and USDC, trading can commence.

Therefore, DWF's data is better seen as evidence of market stratification rather than proof that one model completely replaces another. Professional liquidity is most suitable for mainstream assets with established external markets; traditional AMMs still hold an advantage in the long-tail asset domain, where the ability to create markets without permission is far more crucial than reducing trading costs by minute fractions.

The tokenized financial market will also diverge along this line: high-volume assets like Apple stocks, government bonds, SOL, and ETH naturally suit competition among professional traders behind routing tools; assets without mature external markets still rely on public liquidity pools where anyone can inject funds.

The crypto industry initially promised an ambitious vision: software will eliminate financial intermediaries. But today’s trading market provides a more complex answer. Some intermediaries are products of old infrastructure; while others exist because there always need to be individuals assessing risks, holding asset inventories, and ready to undertake trades.

Blockchain technology does not necessarily need to eliminate intermediaries to reshape the market. The underlying infrastructure can be public, settlement records verifiable, assets able to transfer across applications rather than being locked in a single institution; professional traders can compete for the same orders without needing to possess the underlying infrastructure. Users clicking the exchange button will never know which institution ultimately completes the transaction.

The future exchange may just be a blockchain that anyone can verify, encasing a trading counter that is nearly invisible to all.