This is adapted from an unrecorded talk I gave at MEV Day before Breakpoint in 2025. All data in this article is from the week of December 1–8, 2025. It was collected by querying historical data from Allium and manually categorizing tip and fee addresses.
Application developers who own the user hold a privileged role in financial markets. They get to decide where trades are routed and how they are filled. While much is known about how this works in traditional markets, there has been nearly nothing written publicly about how orderflow is routed and monetized on Solana, the largest decentralized spot market in the world.
This report is an attempt to pull back the curtain on how apps monetize user orders and provide some data on what the landscape looks like today.
In traditional finance, users place market or limit orders that are sold to trading firms for execution. Regulation ensures these orders receive reasonable execution relative to current market prices. Trading firms are willing to pay to fill these orders because retail flow is extremely unlikely to be toxic.
In crypto, applications control how user swaps get filled and subsequently routed. There are several key decision points:
All of these decisions monetize differently and pit the interests of the user against the interests of the application.
The simplest way to monetize orderflow is to charge a fee for each swap. This is the approach taken by most applications on Solana and is already legible to the ecosystem. But it is not the only way apps monetize trades.
Similar to traditional PFOF, applications can charge market makers for access to fill user orders. This manifests differently if the order is filled through RFQ or DEX routing. At this point it is not clear which approach is more profitable for applications.
Request-for-quote is the closest system to traditional PFOF. The obvious angles to monetize RFQ order flow are:
There are limited reports of this happening in practice but we should expect it to develop as spreads get tighter.
Prop AMMs and DEXs make meaningful profit from being connected to aggregators. Aggregators should be able to charge market makers for flow and pass it back to apps. There are also no reports of this happening yet although many have speculated that it would be a natural strategy for aggregators and frontends and may begin to happen quietly.
Due to path dependency, apps and wallets often end up with already-routed trades in the form of signed transactions. You can think of these as market orders with slippage parameters.
The user-aligned approach is to sell the backrun to a trading firm. The extractive approach is to facilitate sandwich attacks. This creates an incentive to slow down transaction landing to increase the value of the backrun. Another perverse incentive is that apps can route users through pools with lower liquidity, creating a larger backrun opportunity. There are reports of all of this happening with prominent frontends on Solana.
Users have an expectation that they must pay a high transaction fee to land, and that a higher fee will get them speed and priority. This creates an incentive for frontends to overcharge their users and keep the excess fee for themselves.
Almost all transaction fees on Solana are currently paid in SOL. There are two different ways that apps can set fees on a user transaction. Many apps will set both fees as we will see later.
Apps can route user transactions through either the vanilla RPC services or through transaction landing services that give them more flexible options to monetize high user fees. The existence of transaction landing services is mostly because of the fragmented builder market that makes transaction routing complex.
Data collected December 1–8, 2025
There were 450 million transactions sent on Solana in the past week
Most of the transactions are related to trading
Priority fees vary dramatically by wallet activity level. Low-activity wallets pay significantly higher fees at the tail. Given that blocks are not full, these users are getting overcharged. These tips are not as easy to monetize, but the builders prefer higher fees because they can market higher yield to validators.
| Wallet Tx Count | Total Transactions | p50 | p90 | p99 |
|---|---|---|---|---|
| 1-9 | 16,000,000 | 5,000 | 36,650 | 1,005,000 |
| 10-99 | 22,000,000 | 6,700 | 127,421 | 3,006,790 |
| 100-999 | 32,000,000 | 8,007 | 505,000 | 5,005,000 |
| 1k-9k | 66,000,000 | 6,000 | 35,350 | 1,005,000 |
| 10k-99k | 62,000,000 | 6,700 | 12,855 | 202,011 |
| 100k-999k | 59,000,000 | 5,401 | 11,704 | 109,424 |
| 1M-9M | 136,000,000 | 5,150 | 11,000 | 48,803 |
| 10M-99M | 38,000,000 | 5,013 | 5,106 | 7,300 |
Priority fee units are in lamports
Overtipping on transaction landing services is the most obvious way to improve transaction monetization. If there was actual congestion on transaction inclusion it may be rational to set extremely high tips in certain situations. But given these numbers during a relatively calm week, it is clear that users are just being overcharged and the excess tip is split between apps and service providers.
| Swap Tips | Swap Count | Other Tips | Other Count | |
|---|---|---|---|---|
| Jito | $300,000 | 24,000,000 | $1,100,000 | 55,000,000 |
| Nozomi | $1,250,000 | 2,500,000 | $330,000 | 1,000,000* |
| Bloxroute | $150,000 | 250,000 | $35,000 | 50,000 |
| Astralane | $150,000 | 1,600,000 | $60,000 | 1,000,000 |
| Zero Slot | $320,000 | 2,250,000 | $60,000 | 400,000 |
| Helius | $20,000 | 630,000 | $5,000 | 280,000 |
* HumidiFi does not pay per tx so these numbers undershoot substantially for Nozomi
Data collected December 1–8, 2025
Axiom generates more transaction fees than any other app on Solana. This is because they dramatically overcharge their users.
Axiom users are so sensitive to perceived speed that the app is able to double-charge the users without any blowback. This generates a substantial kickback to Axiom above the swap fees they collect.
| Wallet Tx Count | Total Transactions | p50 | p90 | p99 |
|---|---|---|---|---|
| 1-9 | 113,029 | 1,005,000 | 3,005,000 | 15,044,032 |
| 10-99 | 928,373 | 1,005,000 | 2,510,681 | 10,950,314 |
| 100-999 | 3,358,861 | 505,000 | 2,005,000 | 10,005,000 |
| 1k-9k | 1,783,981 | 105,000 | 1,010,024 | 7,116,000 |
| 10k-99k | 297,197 | 6,000 | 15,000 | 1,005,000 |
Priority fee units are in lamports
| Landing Service | Total Transactions | p50 | p90 | p99 |
|---|---|---|---|---|
| Nozomi | 2,567,189 | $0.13 | $1.30 | $3.25 |
| Zero Slot | 1,204,008 | $0.01 | $0.18 | $1.95 |
| Jito | 916,178 | $0.001 | $0.04 | $0.12 |
| Astralane | 755,794 | $0.001 | $0.008 | $0.32 |
I would guess that the vast majority of the transaction fees paid by Axiom users end up back in the pocket of the Axiom team.
In the days after I gave this talk, I heard from folks in the industry that the backrunning profit on flow from Axiom and other trading apps is possibly quite a bit higher than the profit generated from overtipping. I have not done the work to verify this but may investigate it in the future.
Regardless, there are numerous ways to monetize transaction orderflow on Solana besides the swap fee, and I hope these are more obvious after reading this article.
User incentives and application incentives are seriously misaligned. Users have no understanding of how large a normal transaction fee should be given the issues with the Solana transaction pipeline over the past few years.
The most important variable for user welfare is to make the transaction fee system standardized and legible to the user so they can understand when they are being overcharged. One approach to this is MCP + Priority Ordering that should be coming to Solana in 2026.
Another idea is to add the concept of "market buy for inclusion" to the protocol. The user should be able to specify "I want to pay the minimum to get included up to X Lamports." This exists on Ethereum with dynamic base fees.
To compete with traditional financial systems, Solana needs to improve the efficiency of the system and take user welfare seriously.
If you enjoy this type of research, come work with me at Meridian.