<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[All in Web4]]></title><description><![CDATA[欢迎来到 All in Web4。由研究员 Blaze Wood（木木）主理，专注 Web3 基本面分析、AI 智能体网络与 Crypto 生态深度研报。剥离市场噪音，为你挖掘下个周期的前沿 Alpha 与百倍价值发现。
Explore "All in Web4" for in-depth fundamental research on Crypto, Web3 ecosystems, and the emerging AI-driven Web4 era. Uncover market Alpha with data-backed insights.]]></description><link>https://allinweb4.hashnode.dev</link><image><url>https://cdn.hashnode.com/uploads/logos/699961a0a20b74e093e54fc6/7cb621a1-c65b-4b1e-a65b-e6e3c8d11d53.jpg</url><title>All in Web4</title><link>https://allinweb4.hashnode.dev</link></image><generator>RSS for Node</generator><lastBuildDate>Wed, 18 Nov 2026 14:02:24 GMT</lastBuildDate><atom:link href="https://allinweb4.hashnode.dev/rss.xml" rel="self" type="application/rss+xml"/><language><![CDATA[en]]></language><ttl>60</ttl><item><title><![CDATA[I Analyzed Uniswap V3 Across Every Major L2 — Here's What the Data Actually Says]]></title><description><![CDATA[Data-driven analysis of Uniswap V3 performance across major L2 chains. Comparing Fee/TVL ratios, LP rebalancing patterns, inactive positions, and real]]></description><link>https://allinweb4.hashnode.dev/i-analyzed-uniswap-v3-across-every-major-l2-here-s-what-the-data-actually-says</link><guid isPermaLink="true">https://allinweb4.hashnode.dev/i-analyzed-uniswap-v3-across-every-major-l2-here-s-what-the-data-actually-says</guid><category><![CDATA[Uniswap V3]]></category><category><![CDATA[L2]]></category><dc:creator><![CDATA[Terry Liu]]></dc:creator><pubDate>Mon, 02 Mar 2026 13:19:45 GMT</pubDate><content:encoded><![CDATA[<p><em>By Blaze Wood · March 2026</em></p>
<p>I've been digging into Uniswap V3's performance across L2s for the past few weeks. Not the "which chain is pumping" kind of analysis — I wanted to understand where liquidity is actually being <em>used</em> versus where it's just sitting there collecting dust after some airdrop campaign ended.</p>
<p>The short version: <strong>Base is eating everyone's lunch</strong>, Arbitrum is the reliable workhorse that stopped growing, and most zkEVM chains are running on fumes.</p>
<p>Let me walk you through the data.</p>
<hr />
<h2>The Setup</h2>
<p>Uniswap V3 is deployed on 40+ chains now, but the real action happens on maybe five of them. I pulled data from DeFiLlama, cross-referenced with Dune Analytics queries, and looked at on-chain LP behavior patterns from 2025 Q1 through early 2026.</p>
<p>The question I kept coming back to: <strong>Is the TVL on a given chain actually doing anything, or is it just parked there?</strong></p>
<p>Turns out, the answer varies wildly.</p>
<hr />
<h2>TVL: The Numbers Everyone Quotes (and Why They're Misleading)</h2>
<p>Here's the raw TVL picture:</p>
<table>
<thead>
<tr>
<th>Chain</th>
<th>Early 2025</th>
<th>Early 2026</th>
<th>Change</th>
</tr>
</thead>
<tbody><tr>
<td>Arbitrum</td>
<td>~$800M</td>
<td>~$700M</td>
<td>Flat to slightly down</td>
</tr>
<tr>
<td>Base</td>
<td>~$200M</td>
<td>~$600M</td>
<td>Tripled</td>
</tr>
<tr>
<td>Optimism</td>
<td>~$150M</td>
<td>~$150M</td>
<td>Flat</td>
</tr>
<tr>
<td>Polygon</td>
<td>~$30M</td>
<td>~$25M</td>
<td>Slow bleed</td>
</tr>
<tr>
<td>zkSync Era</td>
<td>~$30M</td>
<td>~$15M</td>
<td>Halved</td>
</tr>
<tr>
<td>Linea</td>
<td>~$3M</td>
<td>~$750K</td>
<td>Down 75%</td>
</tr>
<tr>
<td>Blast</td>
<td>~$15M</td>
<td>~$3M</td>
<td>Down 80%</td>
</tr>
</tbody></table>
<p><em>Source: DeFiLlama API, March 1 2026 snapshot</em></p>
<p>But TVL alone is a terrible metric. A chain can have $500M locked up and still be a ghost town if nobody's actually trading against that liquidity. Which brings us to the metric that actually matters.</p>
<hr />
<h2>Fee/TVL: The Metric Nobody Talks About Enough</h2>
<p>Fee/TVL ratio tells you how hard each dollar of liquidity is working. High ratio = active trading against the pool. Low ratio = money sitting idle.</p>
<table>
<thead>
<tr>
<th>Chain</th>
<th>TVL</th>
<th>Annualized Fees</th>
<th>Fee/TVL</th>
</tr>
</thead>
<tbody><tr>
<td>Base</td>
<td>~$500M</td>
<td>~$380M</td>
<td>~76%</td>
</tr>
<tr>
<td>Arbitrum</td>
<td>~$700M</td>
<td>~$250M</td>
<td>~36%</td>
</tr>
<tr>
<td>Optimism</td>
<td>~$150M</td>
<td>~$50M</td>
<td>~33%</td>
</tr>
<tr>
<td>Ethereum L1</td>
<td>~$1.5B</td>
<td>~$300M</td>
<td>~20%</td>
</tr>
<tr>
<td>zkSync Era</td>
<td>~$15M</td>
<td>~$1M</td>
<td>&lt;10%</td>
</tr>
</tbody></table>
<p>Base's fee efficiency is <em>double</em> Arbitrum's. That's not a rounding error — it means Base liquidity is genuinely being used by real traders, not just farming rewards.</p>
<p>Why? Two reasons: Coinbase funnels retail users directly into Base (they don't even know they're on an L2 half the time), and gas is so cheap ($0.01 average) that high-frequency rebalancing actually makes economic sense.</p>
<hr />
<h2>How Gas Costs Shape LP Behavior</h2>
<p>This is where it gets interesting. V3's concentrated liquidity model means LPs need to actively manage their positions — set tight price ranges, rebalance when price moves out of range. But rebalancing costs gas. So the chain's gas cost directly determines LP strategy:</p>
<p><strong>On Ethereum L1</strong> ($15-50 per rebalance): LPs set wide ranges and pray. "Set and forget" is the dominant strategy. Result: 50-65% of positions are out of range at any given time.</p>
<p><strong>On Base</strong> ($0.01 per rebalance): LPs set tight ranges and rebalance frequently — average of 6 times per month. Automated Liquidity Management (ALM) protocols like Gamma and Arrakis are everywhere. Only 25-40% of positions are inactive.</p>
<p><strong>On zkSync/Linea</strong>: Nobody's home. 60-75% of positions are inactive. Most of the TVL is leftover from incentive programs that ended months ago.</p>
<p>Here's the rebalancing frequency data:</p>
<table>
<thead>
<tr>
<th>Chain</th>
<th>Avg Rebalances/Month</th>
<th>% of LPs Rebalancing &gt;5x/Month</th>
</tr>
</thead>
<tbody><tr>
<td>Base</td>
<td>6.2</td>
<td>52%</td>
</tr>
<tr>
<td>Arbitrum</td>
<td>4.8</td>
<td>38%</td>
</tr>
<tr>
<td>Optimism</td>
<td>3.9</td>
<td>29%</td>
</tr>
<tr>
<td>Polygon</td>
<td>1.7</td>
<td>8%</td>
</tr>
<tr>
<td>zkSync</td>
<td>2.1</td>
<td>12%</td>
</tr>
</tbody></table>
<p>The pattern is clear: cheaper gas leads to tighter ranges, more rebalancing, and higher capital efficiency. V3's design works best on chains where adjusting positions is nearly free.</p>
<hr />
<h2>Who's Actually Providing Liquidity?</h2>
<p>The composition of LPs tells you a lot about a chain's maturity:</p>
<p><strong>Arbitrum</strong> has the most balanced ecosystem — roughly a third large LPs (&gt;$100K), 40% retail, and 25-30% automated strategies. It's the "institutional" L2 for DeFi.</p>
<p><strong>Base</strong> skews heavily retail (50%+), which makes sense given the Coinbase pipeline. But ALM coverage is also high (~28%), meaning even retail users are getting access to professional-grade position management.</p>
<p><strong>zkSync and Linea</strong> are whale-dominated (40-70% large LPs). When the whales leave — and they will, because there's nothing to trade against — the TVL will crater further.</p>
<hr />
<h2>The V3 vs V2 Question</h2>
<p>One thing that surprised me: Polygon's V3 penetration is only 45%. QuickSwap (a V2 fork) still dominates there. Compare that to Arbitrum at 93% V3 penetration — it's basically a different universe.</p>
<p>This matters because V3's concentrated liquidity model isn't universally superior. On chains with low trading volume, the complexity of V3 position management isn't worth it. V2's "set and forget" simplicity actually wins when there aren't enough trades to justify active management.</p>
<table>
<thead>
<tr>
<th>Chain</th>
<th>V3 Penetration</th>
<th>Why</th>
</tr>
</thead>
<tbody><tr>
<td>Arbitrum</td>
<td>93%</td>
<td>Mature ecosystem, professional LPs</td>
</tr>
<tr>
<td>Optimism</td>
<td>88%</td>
<td>V3 successfully replaced V2</td>
</tr>
<tr>
<td>Base</td>
<td>83%</td>
<td>V3 dominant, some V2 long tail</td>
</tr>
<tr>
<td>Polygon</td>
<td>45%</td>
<td>QuickSwap V2 still king</td>
</tr>
<tr>
<td>zkSync</td>
<td>94%</td>
<td>V3 only game in town (but the town is empty)</td>
</tr>
</tbody></table>
<hr />
<h2>Dune Queries: Verify It Yourself</h2>
<p>I'm including two of the core SQL queries I used. These run on Dune Analytics — feel free to fork them and check my work.</p>
<p><strong>Inactive Position Analysis (Arbitrum example):</strong></p>
<pre><code class="language-sql">WITH current_pool_price AS (
  SELECT pool, tick_current
  FROM uniswap_v3_arbitrum.pools
  WHERE block_time = (SELECT MAX(block_time) FROM uniswap_v3_arbitrum.pools)
),
positions AS (
  SELECT
    p.pool, p.owner, p.tick_lower, p.tick_upper,
    CASE
      WHEN pp.tick_current &gt;= p.tick_lower AND pp.tick_current &lt;= p.tick_upper
      THEN 'ACTIVE' ELSE 'INACTIVE'
    END AS status
  FROM uniswap_v3_arbitrum.positions p
  JOIN current_pool_price pp ON p.pool = pp.pool
  WHERE p.liquidity &gt; 0
)
SELECT
  pool,
  COUNT(*) AS total,
  ROUND(SUM(CASE WHEN status = 'INACTIVE' THEN 1 ELSE 0 END) * 100.0 / COUNT(*), 2) AS inactive_pct
FROM positions
GROUP BY pool
ORDER BY inactive_pct DESC
LIMIT 50;
</code></pre>
<p><strong>LP Rebalancing Frequency:</strong></p>
<pre><code class="language-sql">WITH position_events AS (
  SELECT 'mint' AS event_type, owner, pool, block_time
  FROM uniswap_v3_arbitrum.mints WHERE block_time &gt;= TIMESTAMP '2025-01-01'
  UNION ALL
  SELECT 'burn', owner, pool, block_time
  FROM uniswap_v3_arbitrum.burns WHERE block_time &gt;= TIMESTAMP '2025-01-01'
),
rebalances AS (
  SELECT e1.owner, e1.pool, e1.block_time AS burn_time, e2.block_time AS mint_time
  FROM position_events e1
  JOIN position_events e2
    ON e1.owner = e2.owner AND e1.pool = e2.pool
    AND e1.event_type = 'burn' AND e2.event_type = 'mint'
    AND e2.block_time &gt; e1.block_time
    AND DATEDIFF('hour', e1.block_time, e2.block_time) &lt;= 6
)
SELECT
  owner,
  COUNT(*) AS rebalance_count,
  DATEDIFF('day', MIN(burn_time), MAX(mint_time)) AS active_days,
  ROUND(COUNT(*) * 1.0 / NULLIF(DATEDIFF('day', MIN(burn_time), MAX(mint_time)), 0), 3) AS per_day
FROM rebalances
GROUP BY owner
ORDER BY rebalance_count DESC;
</code></pre>
<hr />
<h2>So What?</h2>
<p>If you're an LP trying to decide where to deploy capital:</p>
<ol>
<li><p><strong>Base</strong> is the best risk-adjusted choice right now. Highest fee efficiency, lowest gas for rebalancing, strong retail flow from Coinbase. The catch: you're betting on Coinbase's continued growth.</p>
</li>
<li><p><strong>Arbitrum</strong> is where you go for depth. If you're moving serious size and need minimal slippage, nothing else comes close. But don't expect growth — it's a mature market.</p>
</li>
<li><p><strong>Unichain</strong> is the wild card. Uniswap's own L2 is already capturing ~50% of V4 volume. Worth watching closely over the next 6-12 months.</p>
</li>
<li><p><strong>Stay away from zkEVM chains</strong> for LP purposes. The TVL numbers look okay on paper, but 60-75% of positions are inactive. You'd be providing liquidity that nobody trades against.</p>
</li>
</ol>
<p>One last thing worth noting: across all chains, research suggests roughly 80% of active V3 LPs underperform simple buy-and-hold over time. Impermanent loss is real, and concentrated liquidity amplifies it. The 43% TVL sample I looked at showed LPs earning \(199M in fees but suffering \)260M in IL — a net loss of $61M. ALM protocols help, but they're not magic.</p>
<hr />
<h2>Caveats</h2>
<ul>
<li><p>TVL figures are estimates based on DeFiLlama snapshots; intraday fluctuations can be significant</p>
</li>
<li><p>LP behavior metrics (inactive %, rebalance frequency) are derived from on-chain patterns and published research, not exhaustive position-by-position analysis</p>
</li>
<li><p>MEV and wash trading inflate volume numbers on all chains to varying degrees</p>
</li>
<li><p>This covers 2025 Q1 through 2026 Q1 — crypto moves fast, and these conclusions have a shelf life</p>
</li>
</ul>
<p><em>Data sources: DeFiLlama API, CoinLaw Research, Keyrock L2 Analysis, published academic literature on Uniswap V3 LP performance.</em></p>
<p><em>This is research, not financial advice. Do your own homework.</em></p>
]]></content:encoded></item></channel></rss>