> ## Content Index
> Fetch the complete content index at: https://blog.arbitrum.foundation/llms.txt
> Use this file to discover other available public pages before exploring further.

# How Arbitrum’s New Ordering Mechanism Creates More Opportunities for Traders & DeFi Primitives
- URL: https://blog.arbitrum.foundation/how-arbitrums-new-ordering-mechanism-creates-more-opportunities-for-traders-defi-primitives/
- Published: 2026-09-28T09:30:22.000Z
- Updated: 2026-09-28T09:30:22.000Z
- Author: Jessica Young

Blockchains lack native CRON jobs or automated triggers, so when market conditions change, like when a token price moves on one exchange and not another, DeFi’s stability relies on having economic incentive for balancing DeFi products. 

That work is MEV: arbitrage, liquidations, backrunning, etc. Without it, tokens would trade at varying prices across exchanges, and lending platforms would carry users with bad debt.

Recently, our DevRel [Jess](https://x.com/@0xjesstech?ref=blog.arbitrum.foundation) covered how users, apps, and chains try to separate harmful MEV from useful MEV 👇

[![](https://storage.ghost.io/c/d4/cc/d4cc8040-691e-4dd2-8282-0c08d45fc5a8/content/images/2026/09/image-1.png)](https://x.com/0xjesstech/status/2095468637268091128?ref=blog.arbitrum.foundation)

[Read the Article](https://x.com/0xjesstech/status/2095468637268091128?ref=blog.arbitrum.foundation)

**The problem is not that economic incentive for stabilizing DeFi products exists. It is who can see it and compete for it.**

Abritrum’s new mechanisms - Priority Gas Auctions (PGA) and Fast Feed - are designed to create more opportunities for traders and even the playing field for smaller MEV searchers to get involved in supporting the stability of DeFi. It also creates opportunity for emerging DeFi primitives, like proprietary AMMs (propAMMs), who need cheap, frequent, priority inclusion to keep onchain parameters fresh. More information on propAMMs 👇

[![](https://storage.ghost.io/c/d4/cc/d4cc8040-691e-4dd2-8282-0c08d45fc5a8/content/images/2026/09/image-2.png)](https://x.com/0xjesstech/status/2097688462488899836?ref=blog.arbitrum.foundation)

[Read the Article](https://x.com/0xjesstech/status/2097688462488899836?ref=blog.arbitrum.foundation)

It’s important to note - the mempool remains private on Arbitrum One. A private mempool is one of the simplest protections against harmful MEV like sandwiching and front-running without killing arbitrage or liquidations. On Arbitrum One, transactions stay hidden until they are sequenced.

### **Why PGA & Fast Feed Matter**

**For traders and searchers:** This upgrade reduces the barrier to entry and creates opportunity for new players to become searchers (MEV bot operators) without the previous requirement of needing the same hardware and latency as those already competing.

**For protocols and end users:** This upgrade creates a healthy market for arbitrage and liquidation bots - keeping DEX prices aligned and bad debt liquidated.

**Emerging DeFi primitives:** Newer DeFi applications like proprietary AMMs (propAMMs) need cheap, frequent, priority inclusion to keep onchain parameters fresh, which this upgrade supports

**For Arbitrum DAO & ecosystem:** This upgrade redirects value that previously leaked out to hardware providers and node infrastructure to the Arbitrum DAO treasury, to fund chain growth and public goods.

### **Introducing Fast Feed**

Searchers need to do two things: identify an opportunity, and act on it, quickly.

For example, a searcher could identify an upcoming DEX swap that will deviate the token price from its fair market value, and create a back-running arbitrage transaction that rebalances the token price across exchanges.

Fast Feed is built to support that first step: it’s a faster stream of already-sequenced transactions, helping searchers identify a state changing transaction that has been sequenced. 

Until now, maintaining a low-latency stream of ordered transactions usually meant running your own infra close to the Arbitrum Sequencer. Fast Feed turns the same data into a subscription product. 

**How this works**

Transactions are ordered by the Arbitrum Sequencer, and the public RPC feed publishes each transaction, its place in the sequence, and related metadata. Fast Feed is a permissionless subscription service that publishes this data a fraction of a second earlier. Access is managed by a smart contract onchain, where users can subscribe and pay a subscription fee. 97% of those fees go to the Arbitrum DAO.

### **Transaction Ordering Mechanisms**

Recall that searchers need to do two things: identify an opportunity, and act on it quickly.

Transaction ordering mechanisms determine how quickly users can act on opportunities, and have their transactions executed. Priority Gas Auctions offer a more fair opportunity to have transactions ordered quickly, without the need to build custom tooling, lowering the barrier to entry. 

For newer DeFi applications like proprietary AMMs (propAMMs) that need cheap, frequent, priority inclusion to keep onchain parameters fresh, PGA makes Arbitrum a compatible ecosystem.

To understand what ArbOS Elara’s PGA system is, let's take a look at what transaction ordering mechanisms have been used in the past, and their problems.

**First Come, First Serve (FCFS)**

Arbitrum One originally ordered transactions on a First Come First Serve (FCFS) basis.

Searchers were competing for arbitrage, liquidation and back-running opportunities through hardware. Whoever had the fastest hardware and lowest-latency connection to the sequencer identified the opportunity faster, submitted the transaction faster, and earned the rewards.

Although arbitrage and liquidation bots support DeFi protocols, this model pushed spending outside of the ecosystem and towards hardware companies, as opposed to the Arbitrum DAO, and had a high barrier to entry for new Searchers wanting to participate.

**Timeboost**

Timeboost was introduced to Arbitrum in April 2025, and was created to combat the battle for hardware. This was the first iteration for Arbitrum that redirected a searcher's spending towards the DAO.

Timeboost introduced offchain auctions between searchers. Each searcher would bid for control of an ‘express lane’. For 60 seconds, the winning searcher would get to submit any transaction through an express lane endpoint, which got sequenced immediately. All other transactions were held for 200ms before being included in a block. The winning searcher maintained this position across all blocks within the 60 second round. 

For searchers, timeboost swapped the race for better hardware to a bidding war, where the proceeds were sent to the DAO, ultimately supporting the underlying protocol.

**Why transition away from Timeboost?**

In practice, Selini and Wintermute alone won 77.41% of Timeboost auctions from April 17, 2025, to April 13, 2026 [\[1\]](https://arxiv.org/html/2509.22143v2?ref=blog.arbitrum.foundation). This meant that smaller searchers couldn’t compete, and stopped bidding.

Instead, smaller searchers seemed to start competing with others outside of the express lane, accepting the 200ms delay. This means that instead of targeting MEV opportunities that the searcher in the express lane targeted, they found alternative opportunities and competed for those. 87% of atomic arbitrage and \~98% of liquidations bypassed the express lane entirely, meaning Timeboost captured only \~1.7% of the atomic arbitrage revenue [\[2\]](https://forum.arbitrum.foundation/t/constitutional-aip-transition-arbitrum-one-ordering-policy-to-priority-gas-auctions-pga/30942?ref=blog.arbitrum.foundation).

### **Priority Gas Auctions**

Priority Gas Auctions give users the opportunity to add a priority fee if inclusion speed matters. 

When the Sequencer receives transactions, they are stored in the private mempool, tagged with the arrival timestamp, and their priority fee. 

On Arbitrum One, typically blocks are every 250ms, and spans two 125ms PGA rounds. 

In each round, pending transactions are ordered by priority fee. If two transactions have the same priority fee, whichever one has the earliest arrival timestamp is ordered first. 

The sequencer then fills a block with the ordered transactions until it hits one of its limits: the 32 Mgas target is hit, the calldata limit of 95,000 bytes is reached, or the round expires. 

The sequencer collects the calculated priority fee per gas according to EIP-1559.

Typically, a block (250ms) spans two PGA rounds (125ms each). However, if the transactions in a single PGA round fills a block (it hits the gas target, or calldata limit), the block is sealed, and the sequencer begins building the next block. Block production is capped at a maximum of 8 blocks per second.

When a transaction is submitted during a PGA round, it is held and enters the sequencer’s mempool at the start of the next round.

### **Introducing the Anti-Starvation Mechanism**

What happens to transactions with no priority boost?

If the mempool has transactions remaining after a PGA round, the outstanding transactions receive a virtual priority boost as part of the anti-starvation mechanism.

This fee is virtual, and isn’t actually charged to the sender. It’s used to ensure that standard user transactions with low or zero priority fees don’t sit in the mempool indefinitely during sustained periods of high volume.

### **Maintaining Nonce Cache**

When a single user submits multiple transactions, it’s important to maintain the original order, so transaction nonces are used to maintain ordering. If a user submits multiple transactions, while the first is still in the mempool, the sequencer lists the remaining transactions in the ‘nonce cache’ until the earlier transaction is included. Transactions in the nonce cache do not receive the priority boost.

### **Conclusion**

Previous transaction ordering mechanisms have had their challenges. For searchers, the barrier to entry was high: participating in an ahead-of-time auction requires building custom tooling and forecasting MEV for an entire upcoming round.

Meanwhile latency-sensitive applications weren’t supported: emerging DeFi primitives such as proprietary AMMs (propAMMs) need cheap, frequent, priority-ordered inclusion to keep onchain parameters fresh. An express lane held by a single controller for 60 seconds at a time is incompatible with these new AMMs.

By pairing Priority Gas Auctions with Fast Feed, ArbOS Elara captures revenue for the Arbitrum DAO treasury, lowers the barrier for entry in the MEV market, and introduces support for latency-sensitive applications.

[Find Out More Here](https://arbitrum.io/products/pga-and-fast-feed?ref=blog.arbitrum.foundation)

---

\[1\] [https://arxiv.org/html/2509.22143v2](https://arxiv.org/html/2509.22143v2?ref=blog.arbitrum.foundation)

\[2\] [https://forum.arbitrum.foundation/t/constitutional-aip-transition-arbitrum-one-ordering-policy-to-priority-gas-auctions-pga/30942](https://forum.arbitrum.foundation/t/constitutional-aip-transition-arbitrum-one-ordering-policy-to-priority-gas-auctions-pga/30942?ref=blog.arbitrum.foundation)