Smart order routing: how DEX aggregators rebuilt an equities mechanism

An audit of 2.98 million WETH-USDC swaps on Ethereum, released in July 2026, analyzed the gap between the routes traders actually received and an optimal benchmark. The difference amounted to 2.02 basis points per trade, which totals approximately $24 million when all swaps are considered (Xi and Moallemi, arXiv:2607.20762). This paper is a preprint, not peer-reviewed, and highlights sandwich attacks and router logic, so it doesn't judge any specific aggregator. Even so, it provides a concrete number in an area often discussed qualitatively: routing quality, on one pair and one chain, can be measured, and it comes at a cost.

Equity markets reached similar conclusions two decades earlier, driven by legal rather than competitive forces. They developed a mechanism called smart order routing, and DeFi recreated it independently, without regulatory prompting.

What is smart order routing?

A smart order router takes an order and decides where it should execute. It evaluates prices at every location the asset trades, calculates what each would yield after its fees and the cost of accessing it, and then either sends the order to the best option or divides it across several, ensuring no single pool or book is overwhelmed.

This division is what distinguishes routing from simple comparison shopping. A quote involves a price for a certain size, and size can affect the price. Sending a large order to a single best price can often degrade that price until the average fill is worse than splitting it across three less-optimal sources. A router that doesn't account for size is merely a price ticker with added complexity.

On-chain, the same logic applies to automated market makers instead of order books, altering the arithmetic while maintaining the same task. Depth is represented as a curve rather than a list of resting orders, the cost of accessing another pool is a gas fee instead of a connectivity charge, and the entire process must be executed in one transaction that either completes or reverts.

Why did stock markets need order routing first?

Fragmentation occurred before the routers emerged. In the United States, the same stock can trade in numerous locations simultaneously: 16 exchanges traded NMS stocks in 2023 according to the SEC's own data (SEC Release No. 34-101070), and 34 NMS Stock ATSs, the alternative trading systems known as dark pools, had an effective Form ATS-N on file as of March 2026 (SEC). The volume reflects this fragmentation. About 45% of NMS stock share volume was traded away from exchanges in 2023, roughly 1,054 billion shares out of 2,324 billion, based on the SEC's data rather than a percentage they stated.

Regulation provided a baseline for this dispersion. The SEC adopted Regulation NMS on 9 June 2005, and its Rule 611, the Order Protection Rule, mandates trading centers to establish written policies reasonably designed to prevent trade-throughs, or executions at prices worse than a displayed automated quotation elsewhere (SEC Release No. 34-51808). Rule 611 is an obligation with specified exceptions, not an outright prohibition, and it only protects the top of the book. It made disregarding another market's displayed price an operational issue rather than a choice.

Europe framed the obligation as an outcome rather than a process. Best execution was introduced into EU law via Article 21 of MiFID, effective from 1 November 2007, which required firms to take "all reasonable steps" to achieve the best possible result for a client considering price, cost, speed, execution and settlement likelihood, size, and nature. MiFID II, effective from 3 January 2018, reiterated this in Article 27 and raised the standard to "all sufficient steps" (ESMA). That difference in wording separates firms that must document an effort from those that must document a result.

Why did DeFi rebuild order routing without a regulator?

There is no rule requiring a decentralized exchange to consider prices elsewhere. There is no on-chain order protection rule, no best-execution duty linked to a smart contract, and no supervisor asking a router to justify a fill. However, fragmentation is more severe than anything faced by equities: every token pair may have its own pool on every DEX across all chains, with pools created permissionlessly and liquidity constantly shifting between them.

So routing became more of a product than a compliance function. Aggregators exist because achieving a better fill is their main proposition, and the market quickly recognized this potential. DEX aggregators routed about $78.6 billion in the 30 days to 12 August 2026, an increase of 20.6% over the preceding 30 days, across the 159 aggregators tracked by DefiLlama with a volume adapter, on 174 chains (DefiLlama).

What does an aggregator of aggregators add?

A meta-aggregator queries aggregators for a price while simultaneously reaching out to the DEXes directly, then routes through whichever yields the highest payout. The aim is not simply more sources, but ensuring a quote assembled this way is at least as good as any single source it considered, because those sources are included in the comparison rather than excluded.

PulseSwap operates across 17 blockchains. On PulseChain, a quote request extends to 11 DEXes and aggregators, including PulseX, 9inch, 9mm, Phux.io, 0xTide, the routing partner Piteas, and PulseSwap's own router, accessing 18,000+ funded pools by the PulseX subgraph's count of pairs holding over $100. Compared to DefiLlama's 30-day volume by protocol, this setup covers 99% of PulseChain DEX trading. On other supported chains, quotes are obtained through commercial aggregator partners, providing access to 600+ liquidity sources with 1inch's own published reach included.

However, scale is not the core argument. PulseChain DEXes traded about $148.0 million in the 30 days to 12 August 2026, with an increase of 83.0% over the preceding 30 days (DefiLlama), demonstrating genuine activity on a chain most routing research has overlooked. The focus is on cost: the on-screen number's transparency matters, because no one is obligated to verify it for you.

What should a trader check before accepting a route?

Ask whether the quote is net. A payout figure that omits the router's fee is marketing, and the only way to verify is by comparing the promised interface amount with what the wallet actually receives in a swap.

Ask whether gas fees are included in the comparison. A route across four pools may offer a better payout than a single-pool route but may still result in a loss after transaction costs. A router that prioritizes payout alone might give you a worse trade without acknowledging it.

Check whether the losing routes stay visible. An interface that displays only the winner asks for trust, whereas one that shows rejected quotes allows you to choose a different route manually if desired. This transparency distinguishes between claims you can audit and those you cannot: how PulseSwap builds a quote is documented, the same route data is accessible through the quote API, and the swap itself ranks every source visibly before any decision is finalized.

Two basis points, spread across many trades, define the value of a routing decision. Equities required regulations to take this seriously. On-chain, the only enforcement is whether individuals care to examine it.

Sources

• Weiye Xi and Ciamac C. Moallemi, "Quantifying Sub-Optimality in Routing for Automated Market Makers", arXiv:2607.20762, submitted 22 July 2026. Preprint, not peer-reviewed.
• SEC, Regulation NMS adopting release, Release No. 34-51808, 9 June 2005 (Rule 611, the Order Protection Rule).
• SEC, Release No. 34-101070, Table 5: exchange and off-exchange share volume, 2023.
• SEC, Form ATS-N filings and information: NMS Stock ATSs on file, March 2026.
• ESMA, MiFID II Article 27, obligation to execute orders on terms most favourable to the client.
• DefiLlama, DEX aggregator volume, read 12 August 2026.
• DefiLlama, PulseChain DEX volume, read 12 August 2026.