Arbitrum weekly net outflow of $130 million: L2 fund migration and reshaping of DeFi liquidity landscape

In early May 2026, a set of on-chain bridging flow data sparked widespread discussion in the crypto community. According to monitoring data from the blockchain analytics platform Artemis, in the week ending May 6, the Arbitrum network recorded the largest cross-chain bridge net outflow among all public chains, approximately $131.59 million. Meanwhile, the derivatives trading chain Hyperliquid had the highest net inflow of about $133.56 million, and Base followed with a net inflow of approximately $34.39 million.

What’s more noteworthy is that Arbitrum actually attracted a total bridging inflow of about $577.75 million over these seven days, ranking first among all networks, but at the same time, there was about $709.34 million in bridging outflow. The offset between these two figures resulted in the largest negative net flow.

This data quickly became a focal point for market observers: Is Arbitrum’s liquidity “escaping,” or is it merely undergoing structural reconfiguration?

From a Standout to Multiple Powers

To understand the current data changes, it’s necessary to review the recent two-year evolution of the Layer 2 (L2) space.

From 2024 to the first half of 2025, Arbitrum, leveraging its first-mover advantage and a robust DeFi ecosystem, maintained the top position in total value locked (TVL) on Ethereum L2s for a long time. During its peak, it attracted deployment from many leading protocols, including GMX, Radiant, Pendle, and other core applications, building the most mature on-chain financial ecosystem at that time.

The turning point came in the second half of 2025. The Coinbase-supported Base network, with its deep integration of fiat onramps with centralized exchanges and a surge in social finance and consumer-grade applications, saw a rapid increase in user activity. By the end of 2025, Base had surpassed Arbitrum in DeFi TVL, capturing about 46% of the L2 market share, while Arbitrum’s share was around 31%.

Entering the first quarter of 2026, the competitive landscape further diversified. Hyperliquid, a Layer 1 (L1) network focused on derivatives trading, emerged strongly, with its perpetual contracts market share continuously expanding. Institutional participants began accumulating its native token HYPE through digital asset financial instruments—according to analyst Aletheia’s disclosure on May 5, this proportion had approached 9%. This evolution from “L2 dual giants competing” to “multi-polar differentiation” forms a key background for understanding the current bridging capital flows.

Capital Flows Are Moving, But Not Unidirectionally “Escaping”

Cross-referencing Artemis’s latest data with historical trends reveals several key structural features.

Net Flow Panorama: Winners and Losers

The table below summarizes the net bridging capital flows of major networks in the first week of May 2026:

Data compiled from Artemis’s monitoring of seven-day cross-chain bridge flows (as of May 6, 2026).

This table reveals a fact that can be easily obscured by simplified narratives: Arbitrum remains the chain with the highest total bridge inflow among all networks. A large amount of capital is still entering its ecosystem; however, the outflow rate is even higher. This indicates that the issue isn’t “lack of attractiveness,” but rather a phase where capital retention within its ecosystem has weakened.

Volatility in Historical Data Suggests Cycles

Looking at a longer timeline, clearer fluctuation patterns emerge. In early March 2026, Arbitrum experienced a weekly cross-chain bridge net inflow of about $615.75 million, ranking first among all networks. However, just two weeks later (mid-March), it showed a net outflow of about $120 million, echoing the current data. By mid-April, Arbitrum had returned to a net inflow state.

This “net inflow—net outflow—re-net inflow” oscillation pattern indicates that Arbitrum’s capital flows are more characterized by cyclical pulses rather than a continuous outflow trend. Each fluctuation is often closely related to specific protocol incentive cycles, cross-chain arbitrage opportunities, or market risk-avoidance sentiment shifts.

Market Sentiment and Narrative Breakdown

Based on the latest data, three typical narrative logics have formed in the market.

Narrative 1: “L2 Capital Rotates to High-Performance Derivatives Platforms”

This narrative suggests that the continuous net inflow into Hyperliquid indicates traders are migrating from general-purpose L2s to derivatives trading networks with specialized performance advantages. Hyperliquid’s trading volume continued to rise in Q1 2026, with TVL steadily increasing to about $1.56B—according to DeFiLlama data as of May 6. The ongoing accumulation of institutional financial tools is creating structural demand for this asset, tightening the available circulating supply.

The core argument here is: as DeFi narratives become more complex, traders prefer to pay for throughput and actual trading experience. Capital is flowing back from the complex abstraction layers of DeFi to more direct and efficient infrastructure layers.

Narrative 2: “Base Is Replacing Arbitrum as the New DeFi Hub”

Supporters of this view point out that Base has not only surpassed Arbitrum in TVL but also holds about 46% of the market share, ahead of Arbitrum’s roughly 31%. Its deep integration with mainstream fiat onramps provides retail user access advantages, making its user growth somewhat “scene-driven” rather than “incentive-driven.”

Narrative 3: “Data Noise Theory”

Some also believe that a single week’s bridging flow data is insufficient to determine a trend. Cross-chain bridge capital flows are influenced by many short-term factors—protocol incentive events, large address asset reallocation, market volatility-driven risk shifts—that can distort net flow signals in the short term. The historical volatility of Arbitrum’s data also supports this cautious stance.

Fact-Based Review of the “Capital Escape” Narrative

While the narrative that “capital is fleeing Arbitrum” has some propagative value, several data structures and ecosystem fundamentals suggest the need for careful examination.

First, total inflows remain substantial. In the same week, Arbitrum attracted about $577.75 million in total bridge inflow, ranking first among all monitored networks. If capital were truly “escaping,” this data would be hard to explain. A more accurate understanding is that Arbitrum’s ecosystem still holds strong appeal, but the “stickiness” of capital staying within it has weakened at this stage.

Second, inflows and outflows coexist. The pattern of high outflows coupled with high inflows reflects that Arbitrum is actively undergoing capital reallocation rather than simply experiencing liquidity exodus. A healthy financial ecosystem often involves frequent capital movements; the key is whether the net difference maintains a consistent direction.

Third, historical data does not support a unidirectional trend. As previously mentioned, Arbitrum’s bridge net flows have fluctuated between positive and negative over the first five months of 2026. Extrapolating a single data point into a trend may overlook this oscillation.

Fourth, the DeFi ecosystem’s depth remains strong. As one of the most mature L2s, Arbitrum’s ecosystem of DeFi protocols remains deep and diverse—an inherent structural advantage that is difficult for many emerging networks to replicate in the short term.

Structural Impact of L2 Competition

Although the “escape” narrative warrants cautious interpretation, the current data indeed reflects three structural shifts in the L2 space.

Change 1: From “General-Purpose L2 Competition” to “Specialized Layered Segmentation”

In 2024, the main axis of L2 competition was “who can offer cheaper transactions.” The Dencun upgrade significantly lowered rollup operation costs, making transaction costs nearly negligible across all L2s. Cost advantage no longer serves as a competitive moat. The competition logic shifted toward “specialization”—Base focusing on consumer applications and fiat entry points, Arbitrum deepening its focus on institutional DeFi and multi-virtual machine ecosystems, Hyperliquid targeting derivatives trading, each building differentiated value propositions.

Change 2: “User Entry” as a Growing Priority

Base’s leading position in TVL is reshaping the market’s valuation of L2s. Previously, industry habitually used TVL as the core ranking metric. Base’s rise indicates that user reach and ecosystem activity are becoming equally critical. The competition among L2s is evolving from “capital efficiency race” to “user entry race.”

Change 3: The Squeezing Effect from L1 and Emerging Networks

Ethereum mainnet’s transaction fees have dropped significantly after the Glamsterdam upgrade. According to CoinW’s weekly report in May, on-chain DEX trading fees on Ethereum have fallen to about $0.01, weakening the “price advantage” narrative that once supported L2’s position. Meanwhile, high-performance L1 networks like Hyperliquid are exerting capital and user diversion pressures from outside the “L2 exclusive track.” The L2 ecosystem faces dual pressures from “above” (L1 performance improvements) and “sideways” (competition from high-performance L1s).

One of Arbitrum’s responses is the Stylus upgrade—introducing WebAssembly virtual machines, enabling developers to write smart contracts in Rust, C, and C++, fully compatible with EVM. This multi-virtual machine architecture aims to lower the entry barrier for traditional Web2 developers and reach a developer base of over ten million Rust and C/C++ programmers.

Conclusion

Interpreting Arbitrum’s net outflow of about $131.59 million in the first week of May as “capital fleeing” has propagative value but conflicts with the data structure in several ways. A more accurate description might be: the L2 space is undergoing a profound reshuffle from “general competition” to “specialized layering,” with capital flows between networks increasing in frequency and scale. Arbitrum remains one of the most liquid and protocol-rich networks in the L2 ecosystem, but it faces competitive pressures—from Base’s user scale challenge, Hyperliquid’s trading scene diversion, to Ethereum mainnet’s performance catch-up—that require new strategic responses.

For market participants, the focus should not be solely on weekly net capital flows but on the long-term performance of each network in more fundamental dimensions such as user stickiness, protocol quality, developer activity, and revenue sustainability. These factors will ultimately determine the final landscape of L2.