
Ethereum's Priority Battle: Who Is Paying for Block Building?
Author: Greenfield Capital (Christoph Rosenmayr & Jascha Samadi)
Compiled by: TechFlow
TechFlow Summary: Since 2024, the Ethereum block building market has undergone structural upheaval: the dominance of Telegram bots has been crushed by MEV strategies, wallet-level rebate mechanisms are distorting user incentives, and MetaMask users' single-transaction priority fees have skyrocketed from $0.15 to $1.20. Behind these changes lies the restructuring logic of Ethereum's value capture chain.
Core Findings
Priority fees drive block value. We analyze the landscape of priority fee order flow on Ethereum to understand which categories (MEV, Telegram bots, smart contracts, DEX frontends, wallet transactions, or solvers) and which participants within those categories are driving priority fee expenditure.
We find that the market is shifting from retail-oriented Telegram bots that dominated 2024 to MEV-related traffic, with MEV now accounting for approximately 66–80% of weekly priority fee expenditure.
We break down the three MEV markets by priority fee expenditure and find that statistical arbitrage and sandwich trading markets have long been dominated by single players (Wintermute and jaredfromsubway, respectively), but they now face fierce competition from new entrants. The atomic MEV market, however, exhibits a highly fragmented competitive landscape.
We aim to understand how opportunities across these different MEV strategies arise in distinct ways and how each strategy differs in terms of priority fee and DEX volume intensity.
Besides MEV, in consumer-facing products, we observe a broad shift from DEX frontends to wallet transactions (primarily MetaMask Swaps), along with a recent resurgence of Telegram bots, with GMGN standing out in recent weeks. We also examine how priority fee rebates may distort incentives between users and wallets, encouraging increased priority fee spending.
All data can be found in this Dune dashboard and will be continuously maintained to provide ongoing insights into the Ethereum order flow landscape.
Introduction & Motivation
Since 2024, the Ethereum block building market has experienced significant shifts. Changes in order flow allocation, increased monetization at the wallet layer, the rise of private transaction routing, and evolving competitive dynamics among builders have substantially altered how value is created and captured within the block production stack.
This article is part of our broader research on the Ethereum block building ecosystem. Readers interested in understanding market structure, key participants, and the value chain may wish to start with our recent introductory piece on the block building market.
Building on this, our analysis core is divided into two parts to deeply explore the economics of block building. We aim to understand the landscape of valuable and monetizable order flow on Ethereum, and how it shapes block building competition through builder-level outcomes. We also seek to understand how trends in order flow and its relationships have reshaped competition among block builders over the past few years.
Part One examines the state of Ethereum block building since 2024 from a demand-side perspective, analyzing priority fee order flow sources, user behavior, wallet monetization, and transaction priority fee payments. Part Two shifts focus to the supply side, exploring how builders compete, generate builder surplus, and capture value within the block production stack.
Together, these articles provide a comprehensive view of how value enters, flows through, and is ultimately captured in today's Ethereum block building ecosystem.
Our analysis is entirely based on observable on-chain data. All data can be found in this Dune dashboard and can be used to generate additional insights beyond what is highlighted in this article. We aim to make the data forward-maintainable to provide visibility into future market structure changes.
Other critical factors affecting builder competitiveness, such as latency differences or relay strategies, fall outside the scope of this analysis.
Since the Merge in September 2022, MEV-Boost (an out-of-protocol implementation of PBS) has become the dominant mechanism for Ethereum validators to monetize execution-layer proceeds by allowing them to sell sequencing rights in their blocks to specialized entities known as block builders in an auction. This has created a competitive market for transaction ordering.
Because consensus-layer rewards are funded by newly issued ETH and base fees computed via EIP-1559 are burned, execution-layer rewards generated through block building auctions are the only non-dilutive validator revenue stream directly tied to blockspace demand.
This means the structure of the builder market not only dictates who builds blocks but also impacts the long-term economic sustainability of validator revenue.
Furthermore, it affects the efficiency with which blockspace is allocated among users and, depending on the additional services builders can provide, influences the user experience interacting with that blockspace (e.g., execution quality). This layer is where market-driven value for transaction inclusion and ordering is priced, routed, and allocated, shaping all critical downstream outcomes that matter to users and validators alike.
Order Flow & Priority Fee Definitions
Order flow refers to all transactions submitted for inclusion on-chain. Not all order flow is equally valuable. Many transactions consume gas and pay standard fees but create little additional value beyond that.
Priority fees are payments made by order flow originators to block builders for specific state access, ordering, and inclusion. These payments are separate from computed base fees (burned via EIP-1559) and are paid to secure a specific transaction's inclusion in a block at a particular order relative to others. Priority fees are paid because many profit opportunities on Ethereum are competitive, time-sensitive, and highly sensitive to exact ordering. For example, this could be a competitive CEX-DEX arbitrage opportunity where only the first transaction to interact with the AMM pool captures profit, causing the opportunity to vanish for others, or a retail user attempting to ensure inclusion before others when buying into an anticipated memecoin launch.
These fees are paid either as native priority fees or as direct builder payments via `coinbase.transfer`, both achieving the same objective in this study. We sum these two components per transaction and refer to them collectively throughout as priority fees (also known as builder tips or state access payments). Relevant priority fees are the primary driver of order flow value.
Overall View of the Ethereum Priority Fee Landscape
We begin by providing a raw, high-level overview of the Ethereum priority fee order flow landscape, grouped solely by order flow originator addresses, to see which entities are most willing to pay for priority state access in blocks.
Demand for priority is highly variable depending on market conditions (such as volatility, e.g., the crash on Oct 10) or extraordinary events (like the week of March 9, which saw the highest total priority fees in recent years and the highest single priority fee, resulting from a user on an AAVE frontend accepting a high-slippage trade that suffered a $50 million loss, leading to a significant peak of $34 million in reverse-run priority fees (purple). This singular event will appear across multiple metrics throughout the analysis.)
We can observe that the composition of total priority fees has shifted significantly over time, moving from a market where two players (jaredfromsubway and the Banana Gun Telegram bot) accounted for most priority fee expenditure to a less concentrated one. Total priority fee expenditure has also declined notably since early 2024 (measured in USD and ETH) due to softer market conditions (such as reduced retail speculation on Ethereum mainnet), leading to fewer opportunities.
Priorities Fees by Traffic Category Tags
We utilize tags provided by Flashbots on Dune to tag MEV and categorize priority fee spenders by order flow tags. This does not fully cover all untagged priority fees shown above, as maintaining up-to-date tags is difficult, particularly given the fast-moving nature of MEV, such as searcher teams changing strategy contracts. We expect this may even slightly understate the contribution of MEV strategies.
However, the relative breakdown over time highlights a major structural shift away from retail-focused Telegram trading bots and trading terminals (red), which were the largest priority fee spenders dominating early 2024, toward current MEV-priority spending dominance (dark blue). As shown above, this shift was driven by declining priority fee expenditure from Telegram bots due to reduced token-sniping opportunities on Ethereum that were previously driving spending. However, in recent weeks, we have observed strong growth in priority fee expenditure from Telegram trading bots and terminals.
For currently retail-driven traffic, we also observe a shift from DEX frontends (yellow) to wallet transactions (green), which has grown steadily over the years to command a larger share.
Finally, the "Smart Contracts" category refers to priority fees paid by users when interacting with major token contracts (e.g., USDC or USDT), such as simple token transfers or approvals. We see nominal spending in this category remain relatively stable throughout the year, ranging between $100k–$200k weekly, and recently gained share as nominal spending in other categories declined.
Interestingly, since these transactions are not seeking competitive priority state access or faster inclusion, there is little reason to pay additional priority fees. However, because these are the most used smart contracts on-chain, millions of transactions paying slightly above the required minimum result in this category accounting for 24% of total priority fee expenditure (with average priority fees 99% lower than other categories).
We will further segment these categories in the next section.
Top 100 Priority Fee Spenders in the Last 45 Days
When examining the top 100 priority fee spenders in May 2026 over the past 45 days, we see a strong power-law distribution, with two MEV-related addresses (jaredfromsubway and 0xbdb3) driving the majority of priority fee expenditure, surpassing $2.4 million in priority fees over the period, followed by a sharp drop-off after the top 15. MEV strategies account for 69% of total priority fee expenditure (dark blue), down from approximately 80% at the beginning of the year (as shown in the weekly bar chart above).
The highest-ranking non-MEV-related priority fee spenders are Telegram bots (GMGN Terminal, Maestro, and Banana Gun), with GMGN standing out particularly at third place, spending nominal priority fees comparable to jaredfromsubway and Wintermute. They are followed by top token smart contracts (USDC and USDT) and MetaMask wallet transactions.
Priorities Fee Share by Traffic Category
We further segment priority fee shares by order flow categories to distinguish: frontends, meta-aggregators, routers, and MEV tags. These tags overlap; for example, MetaMask acts as a frontend executing trades through the MetaMask Swap router.
Within frontends, we see the decline of large Telegram bots (Banana Gun and Maestro) and the recent rise of MetaMask Swaps as one of the largest priority fee spending frontends (orange). Recently, all Telegram bots have seen a resurgence, yet GMGN Terminal's entry stands out. Having recently expanded its product to Ethereum mainnet, it now leads both the frontend tags and router tags categories, capturing 41% and 37% of total priority fee expenditure in each, respectively.
When viewing nominal priority fees paid under the frontend category tag, we see GMGN gaining market share by bringing new users of its existing products to Ethereum mainnet. This increases total priority fee expenditure for the frontend category tag rather than taking share from others, exemplified by MetaMask Swap priority fee expenditure remaining flat while significantly losing "market share."
In the meta-aggregator tag, we see solver models (1inch Fusion, Cowswap, and Uniswap X) effectively keeping priority fee spending low. Solvers act as sophisticated counterparties managing this on behalf of users, alongside batched transactions (on Cowswap). In routers, the MetaMask Swap router has led since mid-2025 (orange), recently overtaken by the Maestro router, followed by Uniswap, 0x, Aave, 1inch, and Banana Gun routers with similar shares.
As noted, MEV tags are not comprehensive, but we can observe Wintermute's (dark blue) resurgence dominating priority fee spending in MEV-tagged contracts. We also see jaredfromsubway's (gray) impressive return in recent weeks, which we will clarify later.
While MetaMask wallet transactions have not seen substantial growth since early 2024, we observe their users are willing to pay an average of $1.20 per transaction in priority fees (when manually selecting MetaMask), up from approximately $0.15 in early 2024. More strikingly, GMGN users spend over $2.10 on average, nearly matching Wintermute's average of $3 per transaction (though Wintermute captures actual trade profits per transaction). What has caused this shift in willingness to pay priority fees among MetaMask wallet transaction users since 2024?
Order Flow Auctions (OFA) & Priority Fee Rebates
One factor difficult to capture from on-chain data—and potentially skewing priority fee analysis—is the growing prevalence of OFA priority fee rebates through aggregation and delayed refunds.
Beyond offering front-running opportunities to the highest-bidding searcher and returning that bid back to the order flow provider, OFAs have begun offering a fixed proportion rebate on all priority fees. Under this model, transactions flowing into OFA RPCs are shared among all eligible builders, who package them by returning a fixed proportion of priority fees (typically 90%).
Different OFAs vary in refund mechanics and profit models, as this research details extensively. Returned priority fees are generally not refunded to the originator within the same block but are distributed periodically. This makes attributing refunds to specific transactions impossible using only on-chain data. See here for details.
This can skew conclusions about who is truly paying for priority state access and how much they are actually willing to pay, as a significant portion of the paid priority fees eventually returns upstream to the order flow source. However, we do not expect this to materially impact our overall conclusions. MEV-related priority fees constitute the vast majority of priority fee spending and consist of calibrated bids. Priority fees are set with extreme precision to defeat competing bidders for the same opportunity in builder-level bundling auctions. From a builder's perspective, receiving a 90% priority fee rebate is economically equivalent to bidding 90% less for that specific opportunity.
We believe consumer applications are best positioned to capitalize on priority fee refunds from OFAs, as they handle user RPC selection and facilitate uninformed order flow where priority fees are often set loosely and imprecisely (leading to frequent overpayment). This likely explains the higher observed priority fee spending from wallet-originated traffic, such as MetaMask's average $1.20 priority fee per transaction.
Since these priority fee refunds are typically not passed to MetaMask users, this creates a potential misalignment of incentives between minimizing user fees and maximizing wallet revenue. Wallets have an incentive to encourage usage while generally responsible for suggesting default priority fees to guarantee packaging, which most users blindly accept.
We have recently also witnessed significant consolidation at the OFA layer, such as Blink acquiring Merkle, and Consensys/SMG (operating SERVO MEV Protection) acquiring MEV-Blocker from the CoW DAO. These relationships increasingly influence traffic routing and winner-take-all block dynamics, though much of their economic modeling remains difficult to observe directly from on-chain data.
Given the attribution challenges posed by off-chain priority fee refunds, consumer application priority fee measurements should generally be treated as an upper bound on what their users are actually willing to pay for priority state access and packaging.
We will explore this further in our upcoming consumer order flow research, including how retail-directed trading flows are evolving and what this means for consumer applications, business models, and profitability in Ethereum's trade supply chain.
MEV-Related Priority Fees
As seen, over 66–80% of weekly priority fees—and thus block value—are driven by varying MEV strategies. To better understand the MEV priority fee landscape, we divide MEV into the three most common forms on Ethereum: statistical arbitrage (stat arb, CEX-DEX), sandwich MEV, and atomic arbitrage.
For statistical arbitrage, we reuse classifications from prior Greenfield research and rely on existing classifications for sandwich and atomic arbitrage by Dune user @hildobby.
Given the competitive nature of MEV, searchers must typically pay a large portion of their extractable profits as priority fees to guarantee winning an opportunity through competition. Accurately calculating their actual profits (the remainder after paying priority fees) is difficult and error-prone. In previous Greenfield research, we used Binance tags to estimate profitability margins, assuming an average stat arb margin of approximately 10% (where 90% goes to priority fees) and maximum stat arb opportunity margins reaching up to 30%.
While different MEV strategies (and individual searchers within them) will have varying average profitability depending on competitive dynamics like economies of scale, we consider priority fees a highly useful and reliable proxy indicator for understanding the landscape, keeping in mind they do not perfectly correspond to actual MEV profits.
We first compare the relevant priority fee payments (right) and DEX volume (left) across these three MEV strategies. Statistical arbitrage is the largest of the three, leading on both metrics: 70–85% of MEV-related DEX volume and typically 50% of MEV priority fees originate from statistical arbitrage.
Atomic arbitrage priority fees show significant peaks while sustaining an upward trend since early 2025, frequently contributing over 25% of total MEV priority fees in recent weeks. We hypothesize that Ethereum's annual gas limit increases have made atomic arbitrage (which always involves multi-hop transactions) more competitive, correlating with the November 2025 gas limit increase aligning with a higher share of DEX volume driven by atomic arbitrage.
The priority fee share from sandwiches has shown a steady decline since early 2024 when it once led this metric, formerly dominating MEV-related priority fee spending with over 60%. This can be explained by a larger proportion of retail trades bypassing the public mempool as MEV protection infrastructure has matured over the years, such as private RPCs or dApp-level solutions like batch processing via Cowswap. This is evident in the increasing number of trades bypassing the public mempool, with nearly 92% of DEX volume now routed through private mempools in recent months.
However, in recent weeks, sandwich MEV-related priority fee expenditure appears to have risen considerably, with 35–51% again coming from this strategy. We will consider potential reasons in the next chapter.
Differences between the aforementioned DEX volume and priority fees imply that each of these three strategies varies in priority fee intensity per opportunity. We display the average priority fee per transaction and priority fee as a percentage of DEX volume for each MEV category.
Statistical arbitrage has the lowest priority fee intensity per transaction and as a percentage of DEX volume, likely because opportunities occur more frequently but carry lower total extractable value. Atomic arbitrage features exceptionally dense priority fees, generating 10 times the DEX volume per $1 of priority fee spent, with DEX volume associated with atomic arbitrage averaging 0.4% used in priority fees.
Currently, roughly 38% of Uniswap volume comes from stat arb-related trades, having dipped as low as ~5% in early 2026 and peaked above 45% in early 2025. We can observe a strong correlation between the percentage of Uniswap volume attributed to stat arb and the BTC DVOL index on Deribit (this correlation strengthens if looking at ETH 12-second realized volatility between blocks), as the range of stat arb opportunities depends on the volatility of traded assets across CEXs and AMMs.
Overall, approximately 49% of Uniswap's total volume currently stems from these three identified MEV strategies, having reached up to 56% in early 2024.
Statistical Arbitrage (CEX-DEX) Market Distribution
We now look at which players dominate the statistical arbitrage market through DEX volume (left), priority fees paid (right), and relative fee share (bottom).
The relative distribution is particularly interesting. The priority fee market was once dominated by Wintermute (blue) and SCP/beaverbuild (pink), who together controlled ~50% of the share from early 2024 to September 2024, with the remainder split among a fairly dispersed long tail. After September 2024, a stat arb searcher previously tagged as "Kayle" rapidly gained market share, but both Wintermute and SCP managed to regain and even expand their market share before April 2025, squeezing out other stat arb searchers and controlling over 75% of the market.
Following SCP's gradual transition from beaverbuild to BuilderNet (addressed again below), SCP's searchers began losing market share starting December 2025, leaving Wintermute to dominate 70% in Q1 2026, followed by SCP at ~5%, and 0x7cDa58 (likely a new contract from the team previously operating the "Kayle" tagged contract). This period marked the most concentrated phase of stat arb market share.
In recent months, 0xbdb3ba (peach) has emerged as a challenger to Wintermute, holding roughly similar market share by priority fee expenditure. Together, they account for 83% of total priority fee spending in the stat arb category. This contract employs a unique strategy bundling all stat arb exchanges into a single transaction, featuring over 50 individual legs, as demonstrated in the sample transaction.
The team behind the contract currently challenging Wintermute and slightly领先 in stat arb opportunities remains unclear. Manual inspection suggests 0xbdb3ba, similar to Wintermute, shows a preference for routing to Titan, indicating neither is affiliated with or vertically integrated into any other existing builders.
Sandwich Market Distribution
We examine the same metrics for the sandwich market, where we observe the aforementioned greater decline in sandwich MEV-related priority fees. Furthermore, unsurprisingly to anyone who has been sandwiched on Ethereum mainnet, we see how impressively jaredfromsubway (dark blue) has maintained dominance over the past few years, with minimal fluctuation in market share and consistently hovering between 70–85% of paid relative priority fees.
Note: Data displayed here was collected prior to the theft of jaredfromsubway's primary contract, which exceeded $15 million. How this event impacts sandwich competition in the coming months remains to be seen.
On the other hand, manually excluding jaredfromsubway reveals a market that changes quite rapidly, with different players rising and falling quickly. Compared to the long tail stat arb market when Wintermute and SCP were excluded, the change is far more dynamic. The sandwich market is more stable at the top but significantly less stable below it compared to stat arb.
We see 0x01FD (dark green) recently succeed in securing second place behind jared, spending approximately 20% of total sandwich MEV priority fees and holding this position for several weeks, appearing to be jaredfromsubway's most credible competitor in some time.
While the exact cause of the aforementioned sudden spike in sandwich-related priority fees (45–60% of MEV priority fee spending over the past 3 weeks) is unknown, we can speculate that intensified competition may have triggered a priority fee war between jaredfromsubway and 0x01FD (resulting in compressed margins), as they are the only true drivers behind the increased priority fee spending.
This is reflected in increased priority fee expenditure intensity per dollar of DEX volume (manually selected), with jaredfromsubway (dark blue) spending roughly 3 times more per $1 of DEX sandwich volume than four weeks ago, likely a response to competition from 0x01FD (orange).
As detailed below, 0x01FD is very likely operated by the same team behind BobTheBuilder.
Atomic Arbitrage Market Distribution
Examining the atomic arbitrage market reveals a quite different picture. As previously shown, the entire arbitrage market is slightly smaller but opportunities are markedly more volatile, resulting in sharper peaks.
This makes sense because large atomic arbitrage opportunities typically stem from individual users executing high-price impact swaps on AMM liquidity pools (accidental or deliberate). This enables atomic arbitrage to realign prices across all AMM pools (due to deviations caused by high-impact swaps elsewhere) without corresponding CEX price divergence.
An example is the aforementioned AAVE frontend swap event (swapping $50 million for $35,000), which triggered the highest atomic arbitrage priority fees we have observed since early 2024.
Such anomalous events usually arise from user input errors and occur less frequently. This contrasts with statistical arbitrage, which consistently occurs whenever CEX prices diverge from DEX prices across the 12-second block interval, happening at fixed frequencies based on volatility. Or, when a certain number of users simply choose to trade without sandwich protection, sending their transactions to the public mempool. This explains the inherently more volatile nature of atomic arbitrage priority fees.
Additionally, we see this is the only one of the three distinct MEV markets characterized by highly dispersed competition, with no single entity regularly dominating the space.
This can likely be attributed to the relatively smaller overall share and the natural lack of strict economies of scale in pure atomic arbitrage (e.g., flash loans lower capital barriers). These market conditions differ significantly from stat arb (governed by latency, inventory, and CEX fee tier structures where Wintermute excels) or even sandwiches, where jaredfromsubway's inventory-heavy strategy provides similar economies of scale, enabling long-term dominance in their respective categories.
Part One: Conclusion
Part One Summary
Part One demonstrates that Ethereum priority fee demand is not a single unified market, but rather composed of multiple independent order flow segments with distinctly different economic drivers. Since 2024, the composition of valuable order flow has shifted substantially across both categories and individual players, with each segment evolving under its own competitive dynamics. This is most evident in the MEV domain, where statistical arbitrage, sandwich attacks, and atomic arbitrage exhibit significant differences in concentration, opportunity frequency, and priority fee intensity.
Part One focuses on the demand side: who generates valuable order flow and why they pay priority fees. Part Two will shift to the supply side, exploring how builders compete for these order flows, how they convert into builder surplus, and how Ethereum's increasingly specialized order flow providers and their unique service requirements shape the block production market.
If you have feedback or interest in these topics, please feel free to reach out via Telegram (https://telegram.me/@thirdeye33) or email ([email protected]).
All charts and underlying queries can be found in this Dune dashboard and will be continuously maintained to provide ongoing insights into market structure changes.
Disclosures
Greenfield holds investments in Titan's parent company Gattaca, CoW Protocol, and NuConstruct. This analysis relies entirely on publicly available on-chain data obtained through Dune Analytics. Findings reflect on-chain observations and independent research viewpoints and do not constitute portfolio promotion or investment advice.
The analysis may mention projects, protocols, builders, searchers, and market participants where Greenfield may hold direct or indirect interests. Any such mentions are solely for research and informational purposes and should not be interpreted as endorsements, recommendations, or statements regarding future performance.















