Wallet Report v2

Wallets are arguably the most widely used product in crypto. Whether explicitly or not, they are the gateway to nearly every onchain interaction, including swapping tokens, minting NFTs, and participating in governance. Yet despite their central role, wallets remain surprisingly underexplored in data-driven research.

Part of the challenge is that “wallet” isn’t a single category—it spans custodial and non-custodial models, EOAs and smart accounts, mobile apps and backend SDKs. Custodial wallets leave little onchain data, while non-custodial ones generate traceable signals, although these are fragmented across architectures (EOA, smart contract, EIP-7702 hybrids), standards (account abstraction has ERC-4337, Safe’s modular approach, Argent early adoption, etc), and tooling layers. Today’s wallet systems are often modular: Coinbase Smart Wallets may be deployed via apps like Zora with Privy managing access; Layer3 new wallet uses a smart account architecture built on abstraction layers like ZeroDev Kernel, Pimlico, Turnkey and Dune Echo; and tools like Privy or Reown’s AppKit abstract wallet logic entirely. The boundary between frontend, backend, and wallet is increasingly blurred.

Therefore, rather than presenting a single unified narrative, this report is structured as an anthology of the wallet ecosystem. We explore the space through a series of standalone but complementary sections, each grounded in distinct dashboards, datasets, and methodologies. This format reflects the true nature of wallets today: diverse, fast-moving, and shaped by parallel innovations.

To continue improving this picture, we invite wallet providers, analysts, and researchers to reach out and share additional data, insights, and context. By working together, we can build the most accurate and nuanced map of wallet usage across crypto.

This edition introduces improved datasets and a refined methodology for analyzing embedded swap activity from EOAs. We expand coverage with updated data on Phantom’s newest features, and add dedicated sections on Uniswap Wallet and Meteor in the EOA category. The smart wallet section now includes Biconomy’s evolving stack and Abstract Global Wallet’s chain-native architecture. We also present early adoption metrics for EIP-7702 in its first month live, and introduce Dynamic as a key player powering hybrid wallet infrastructure.

If you have read the V1 report, you can navigate through the table of contents for all the newly added sections.If you have read the V1 report, you can navigate through the table of contents for all the newly added sections.If you have read the V1 report, you can navigate through the table of contents for all the newly added sections.

1. Wallet activity is surging: From embedded swap transfers to UserOperations, nearly every metric points to rising onchain engagement through wallets. Whether via embedded DEX routers or smart account interactions, wallets are no longer passive interfaces; they’re where much of crypto’s daily activity happens.
2. Wallet adoption is geographically diversifying: Emerging markets are driving a growing share of wallet activity, with countries like Nigeria, India, Vietnam, and Indonesia consistently ranking among the top user bases across many wallets analyzed. This trend highlights wallets' expanding role as financial access points in regions underserved by traditional infrastructure.
3. UX is converging toward seamlessness: Passkeys, gas sponsorship, and chain abstraction are becoming baseline. This reflects an industry-wide push to reduce friction and make complex interactions feel familiar, signaling a maturation in smart wallet design and implementation.
4. Wallets are becoming self-contained superapps: The clearest trend is vertical expansion where wallets now embed swaps, staking, quests, bridging, and even game discovery. Rather than simply storing assets, they connect, curate, route, and increasingly own the user journey, blurring the line between tool and platform.
5. Infrastructure is becoming more modular: Solutions like Privy, Reown and Dynamic show how wallet functionality is being abstracted into plug-and-play infrastructure—provisioned at login, embedded into apps, and invisible to end users. This shift enables apps to deliver wallet-powered experiences without exposing the underlying mechanics.
6. Dune’s role in supporting wallet development: Data, whether used as analytics or realtime for backend development, is the key to wallet innovation. Dune is looking to expand its role in making the wallet landscape more vibrant with additional developer solutions on realtime wallet data APIs.
7. Hardware wallets (e.g., Ledger, Trezor) store keys offline for enhanced security but require physical interaction.
8. Software wallets (e.g., MetaMask) are mobile, desktop, or browser-based, easier to use but more exposed to online risks.
9. Custodial wallets (e.g., Binance, Coinbase) are managed by third parties who control users' keys, simplifying access but requiring trust.
10. Non-custodial wallets give users full control over their keys and funds, aligned with crypto’s ethos but shifting security responsibility to the user. Almost all centralized exchanges offer a self-custodial wallet besides the default custodial one.
11. MPC (Multi-Party Cobimputation) splits a key across multiple devices or servers to avoid single-point exposure.
12. Multisig requires multiple private key signatures to approve a transaction, common in DAOs and treasuries.
    
    
13. Address Filtering: We removed mislabeled or ambiguous addresses and added newly verified router addresses for OKX.
14. Deduplication: In the transfer event subquery, we now deduplicate multiple “to” = routerAddress transfer events occurring within the same transaction hash. This ensures we count only one event per transaction to prevent inflated swap counts or volumes.
15. Chain-Specific Address Matching: Rather than treating a router address as universal across chains, we now match each address with the corresponding blockchain/s where the address could be verified. This allows for greater precision in attribution. As a result, we've introduced a new category, ‘Unknown’, which includes address-chain pairs we couldn’t confidently verify. These may reflect either unlabeled but valid router addresses, or unrelated contracts altogether.
16. Transfer Event Analysis (“to” = routerAddress). Captures explicit token transfers from users to router addresses, reflecting the input side of a swap.
17. Transaction-Level Analysis (tx_to = routerAddress). Captures swaps where no direct “to” = router event exists. Here, we select transfers where the transaction’s tx_to is the router, “tx_from” is the user but “to” is not the router, to avoid overlapping with the previous query.
18. Tron, Bitcoin, and Solana are not included due to limited label coverage and integration difficulty in the timeframe of this research. However, we were able to reliably source Solana data for Phantom, which is covered in a dedicated section below.
19. The completeness of wallet coverage depends on available labels: some router addresses may be missing or outdated (e.g. older versions or unclassified deployments).
20. This analysis does not capture swaps routed via external dApps, even if initiated by a wallet (e.g., a MetaMask user swapping directly on Uniswap)—nor does it include transfers between wallets.
21. While not a complete picture, the resulting trendlines align closely with those from Dune Metrics' broader "Transfer Volume " charts (which looks at all net USD value transferred, not just embedded routers), suggesting the data is directionally robust.
22. Binance Alpha, a token discovery and pre-listing access program, initially launched in December 2024 and may have contributed to a longer-term baseline of activity. However, this does not fully explain the March–May spike: Alpha’s initial version was limited to Binance Wallet users, and the larger-scale rollout (Alpha 2.0) in May 2025 actually made the feature available to all Binance platform users, including those outside the wallet ecosystem. This likely had a greater impact on Binance Exchange activity than on wallet-level metrics.
23. A more immediate and likely factor is Binance’s 0-fee swaps campaign, which began on March 17, and offers zero trading fees for swaps executed through the Swap, Bridge, and Quick Buy features in the wallet (excluding third-party dApps). This initiative directly incentivizes in-wallet activity and aligns more precisely with the observed growth pattern.
24. MetaMask Portfolio for multi-chain asset management
25. MetaMask Snaps to enable custom plugin-like features
26. Ethereum still leads by volume, accounting for nearly 70% of non-BNB swap value.
27. Base has emerged as the leader in transfer count, reflecting its growing role as a low-cost environment for wallet-native activity.
28. Polygon, Arbitrum, and other Layer 2 networks continue to contribute meaningfully across both metrics, reflecting the broader trend toward multichain distribution.
29. Wallets are becoming primary swap venues: The steady growth in weekly swap count confirms that wallets are no longer just onboarding gateways but core venues for token trading, driven by convenience and integrated UX.
30. Adoption grows faster than capital: While swap count has steadily increased, volume has fluctuated more across cycles and has grown less. This suggests a broader, more active user base but smaller average trade sizes, pointing to retail adoption and utility-led growth.
31. Market maturity: While market share has shifted significantly from MetaMask’s early dominance, all major wallets analyzed have grown in absolute terms. The result is a more diverse ecosystem, shaped by differing strategies across regions, user types, and design philosophies.
32. Binance dominates recent growth: Binance Wallet surged to 44M weekly swaps and $27B in volume, leading with over 90% of swaps and volume at peak. Almost identical shares are recorded for BNB Chain, showing the powerful flywheel between wallet and native chain, a model Coinbase and Base are optimizing as well.
33. Ethereum remains the value layer: Ethereum’s swap share fell to about 10% but still accounts for 40% of volume, leading high-value trades. Meanwhile, BNB and Base now drive the majority of embedded swap activity, reflecting their traction for high-throughput, cost-efficient transactions.
    ‍
34. Swap count peaked at 10 million weekly in late November, before stabilizing at around 3.4 million by May 2025.
35. Swap volume hit an all-time high of $5.7 billion in January 2025 driven by the peak of the memecoin frenzy, then settled to a steady $700-$800 million per week.
    
    
36. Solana-Native, Cross-Chain Capable: While Phantom now supports Ethereum, Polygon, and Base, 97% of its embedded swap activity and volume occur on Solana, confirming its position as the leading self-custodial wallet in that ecosystem.
37. Massive Onchain Footprint: Swap count peaked at 10 million weekly, with volume hitting $5.7 billion in January 2025. At its peak, Phantom embedded swaps accounted for over 10% of all Solana transfer volume, reaching up to 20% during several weeks in late 2024.
38. Expanding Into Platform Territory: With features like in-app staking and PSOL (its native LST) and social feeds, Phantom is evolving into a vertically integrated DeFi and social hub. Not just a wallet, but a full-stack application driving liquidity and discovery on Solana and beyond.
39. Seamless delegation via EIP-7702 to Uniswap’s own smart wallet contract
40. Support for ERC-5792, enabling batched transactions such as approve + swap in a single click.
41. Future plans for additional batching (e.g. approve + LP, or bridge + swap).
42. ERC-4337 compatibility, paving the way for gas sponsorship and ERC-20 gas payments.
43. Native support for passkeys and session keys, allowing users to authorize new signers and unlock more complex trading flows.
44. DEX-Native Advantage: Uniswap Wallet flips the traditional wallet model by building from a leading DEX outward, integrating deeply with UniswapX to offer users gasless, MEV-resistant swaps across on-chain and off-chain liquidity sources through intent-based execution.
    
    
45. Smart Account Ready: With the upcoming Smart Wallet rollout, Uniswap is blending EOA familiarity with smart wallet flexibility, unlocking batched transactions, gas sponsorship, delegated signers, and passkey support, all designed to streamline complex trading into a seamless user experience.
46. USDC → NEAR with $600k in the past 90 days
47. NEAR → USDC with $338k.
48. NEAR → SOL: $395K
49. NEAR → ETH: $231K
50. SOL → NEAR: $143K
51. Chain-Native Advantage: By fully leveraging NEAR’s intent-based architecture, Meteor enables seamless, gas-abstracted swaps and bridging, positioning itself as a programmable liquidity hub for multichain coordination, especially between NEAR and Solana.
    
    
52. Ecosystem-Embedded Design: Rather than acting as a generic multichain wallet, Meteor is tightly integrated with NEAR-native protocols and tokens, aligning its UX and liquidity flows with the unique structure of the NEAR ecosystem.
    ‍
    ‍
53. Social login onboarding (e.g. email, Google), avoiding traditional seed phrase friction
54. In-wallet fiat onramps, supporting payments via card, Apple Pay, and Google Pay
55. Batch NFT transfers for gaming assets
56. In-wallet game discovery, including an ecosystem calendar
57. Developer tools for seamless wallet provisioning and quest-based user engagement
58. Weekly swap count peaked at 270,000 in January, before stabilizing around 40,000.
59. Swap volume topped $55 million weekly in early 2025, with a current YTD average of $20 million.
60. November–December 2024 saw a notable uptick in onchain activity, likely linked to broader market optimism around election season.
61. In January 2025, activity surged again—this time tied to the launch of Tama.meme, Ronin’s own take on the “pump.fun” meme coin meta. The wave of meme coin speculation led to a surge in wallet interactions, swaps, and user onboarding.
62. Designed for gamers, Ronin Wallet has sustained ~400K weekly users with spikes tied to ecosystem events like Tama.meme. All activity remains native to the Ronin chain, showcasing the power of a laser-focused approach: optimize for one niche, and keep shipping.
63. After consolidating its core user base and use cases, Open Ronin is an ambitious attempt to export that success across chains —bringing its streamlined UX and embedded tools to ecosystems like Base, Arbitrum, and BNB Chain.
    ‍
64. In China, MetaMask controls 87% of users and 61% of funds, but OKX (20%) and Bitget (15%) play a larger role in balance share than user count suggests.
65. South Korea sees MetaMask (68%) and OKX (25%) lead on users, but Trust and Bitget also register notable balance shares.
66. In Singapore, users split between OKX (60%) and MetaMask (35%), while MetaMask slightly edges OKX in balance share (53% vs. 45%).
67. Indonesia shows a similar pattern: MetaMask leads with 59% of users and 57% of balances, with OKX close behind (23% users, 41% balance).
68. In India, MetaMask dominates across the board: 63% of users and 79% of funds, with Phantom surprisingly holding 17% of balances despite low user share.
69. Nigeria also favors MetaMask heavily—71% of users and 75% of funds —followed distantly by OKX and Bitget.
70. Germany: MetaMask leads with 70% of users, but Trust Wallet surprisingly holds 57% of the country’s wallet balances—signaling a strong trust in its storage capabilities.
71. France: MetaMask commands 76% of users, but OKX captures a notable 38% of capital—suggesting it appeals to fewer but higher-balance users.
72. In Brazil, MetaMask holds 71% of users, but Trust Wallet surprisingly holds 74% of total balances, hinting at regional capital trust preferences.
73. Argentina shows an unusual split: users are evenly divided between MetaMask (50%) and OKX (43%), but 96% of balances sit with MetaMask, pointing to deeper financial usage.
74. Some wallets like OKX and Bitget show strong regional clustering, particularly in Asia. Coinbase dominates in North America and parts of Europe. Others (MetaMask, Phantom, Rabby) have a more globally distributed user base, with usage common across emerging markets and capital concentration in developed ones.
75. Switching from user count to wallet balance dramatically shifts the map. Capital is more concentrated in developed markets, regardless of the number of users.
76. Emerging markets drive adoption, developed markets bring liquidity. For teams and protocols, this split may influence both growth strategy and monetization focus.
77. MetaMask and OKX are the most common wallets across most major countries for both usage and funds held.
78. Account recovery without a seed phrase, using trusted entities or "guardians" (known as social recovery) to help restore access if keys are lost
79. Multi-signature access control baked into the wallet logic
80. Gasless transactions via relayers and paymasters
81. Bundling and batching of multiple actions into one transaction
82. Modular upgrades with customizable permissions and extensions
83. Base began rapidly gaining ground in July 2024 and now accounts for over 65% of weekly deployments.
84. Arbitrum also experienced a notable spike in March 2025, briefly reaching 60% of deployments.
85. Biconomy was the dominant factory throughout most of 2023.
86. (Gnosis) Safe began steadily increasing its share in early 2024 and currently leads in deployment volume.
87. Zerodev surged to the top position briefly in October 2024.
88. Stackup recorded two distinct peaks: July 2024 and March 2025.
89. Coinbase emerged as a significant factory in 2025, consistently ranking among the top 5 and showing continued growth.
90. In the second half of April 2025, Base processed over 3 million UserOps per week, representing 87% of total weekly operations.
91. Polygon held over 90% market share until April 2024, decreasing to 60% by July as Base gained traction.
92. Arbitrum remains strong in third place, with intermittent surges in Q4 2024 and March 2025.
93. Polygon led the initial wave of smart wallet deployments in 2023, but Base now dominates, accounting for over 65% of new deployments and up to 87% of weekly UserOperations in April 2025.
94. Account deployments peaked in July 2024 with over 1 million in a single week, but have since declined to ~120K weekly, suggesting early saturation or a shift toward more targeted onboarding.
95. UserOperations continue to grow overall, despite fewer new account creations. This indicates increased usage of existing smart wallets and deeper integration into DeFi and apps.
96. Bundler activity is rising, as bundles per UserOp decrease, pointing to a maturing infrastructure and a more competitive bundler landscape.
97. Factory-level dynamics are highly competitive: Biconomy led for much of 2023, while Safe gained traction in 2024. Zerodev briefly took the lead in October, and Coinbase has steadily climbed in 2025.
98. Factory breakdowns are limited to labeled contracts, meaning some activity is undercounted.
99. Growth patterns correlate with ecosystem strategy. Base’s rise, for example, mirrors active ecosystem investment and strong support for AA infrastructure by players like Coinbase.
100. V1 (non modular account): 762,000+ accounts (largely dormant)
101. V2 (modular account): 890,000+ accounts (active across multiple chains)
102. Nexus (modular account with native orchestration): 244,000+ accounts (launched in April 2025)
103. TVL captures the amount of capital held within smart accounts at a given time.
104. TVP tracks the total value that has moved through accounts—i.e., how much has been used, swapped, bridged, or transacted.
105. Modular Evolution and Diverse Metrics: With over 1.9M smart accounts deployed across three architectures, Biconomy exemplifies how modular design, versioning, and orchestration reshape what usage looks like.
106. Smart Accounts as UX Infrastructure: Nexus, Biconomy’s latest architecture, powers one-click, intent-driven flows across chains, revealing a shift from wallets as static storage to programmable execution layers that abstract complexity for users and developers alike.
107. Base accounts for ~66% of UserOps
108. Optimism: ~20% of UserOps
109. Arbitrum and Polygon: smaller but steady usage
110. Newer chains like Ink and Berachain have also seen rising adoption
111. Since launching its own swap feature, Gem Wallet has generated over $81,000 in swap fees, with 2025 revenue already surpassing $73,000 as of late April.
112. The bulk of fee revenue comes from Ethereum ($70K), followed by BNB Chain and a growing contribution from Arbitrum, particularly in Q2 2025.
113. Most historical swap activity has been driven by BNB, although Ethereum has also been an important contributor, and more recently Base and Solana.
114. On a protocol level, Thorchain stands out as the top revenue generator, with consistent activity also emerging across PancakeSwap and other newly integrated routes.
115. Human-readable transactions, reducing user error.
116. Gas fee optimization, through transaction batching and stablecoin-denominated payments.
117. A “gas tank” system for prepaying fees.
118. Integrated swaps, fiat on/off-ramps, and broad network support (including Ethereum, Polygon, Base, Optimism, Arbitrum, Scroll, Avalanche, and Gnosis).
119. Different architectures, different strengths: Ambire and Layer3 use smart accounts to enable features like gas sponsorship, session approvals, and passkey login. Gem—while not a smart account wallet—shows how lean, protocol-level integrations (e.g. Thorchain, PancakeSwap) can drive usage and fee capture.
120. Incentives matter: Ambire’s XP-based rewards boosted feature exploration, while Layer3 benefits from launching its wallet inside an already active, incentivized ecosystem. Embedding a wallet within a platform that already drives user actions can create a powerful flywheel—accelerating smart wallet adoption from day one.
121. Base: All three wallets show growing traction on Base, suggesting it’s becoming the default proving ground for new wallet products. Layer3 sees 66% of UserOps on Base, Ambire shifted activity there in Q1 2025, and Gem recorded a growing share of swaps on the chain.
122. Execution styles vary, but user growth is visible: While Ambire leads in early adoption of new standards like EIP-7702, Layer3 optimizes flow within its app, and Gem targets modular liquidity access. These approaches illustrate how wallets are diversifying not just features—but also the ways they position themselves in the broader crypto UX stack.
     ‍
123. Create and manage accounts through the Universal Wallet interface
124. Benefit from modular security features like multi-sig and recovery modules
125. Interact across wallets, dApps connected, or external protocols directly (e.g. Uniswap)
126. Optimism dominated early in 2024, peaking at nearly half of all deployments.
127. Base quickly gained momentum during the summer and overtook Optimism by August, now accounting for 60%+ of new accounts, a clear sign of developer and app migration.
128. Polygon experienced spikes in October, coinciding with a sharp rise in prediction market activity (notably Polymarket), and briefly led all chains with 45% of deployments.
129. Ethereum has shown a steady rise in share since late 2024, reaching ~20% by April 2025—perhaps signaling growing demand for mainnet-grade security and permanence.
130. A sharp October uptick on Polygon, again possibly tied to US election-related usage.
131. A gradual rise on Ethereum, with consistent growth pushing it to 15–20% share of active accounts by April 2025.
132. Arbitrum has remained steady throughout, hovering just under 20%.
     
     
133. Base has shown steady and sustained momentum, rising from 18% of smart wallet transactions in June 2024 to 43% by April 2025, establishing itself as a key execution layer for smart wallet activity.
134. Optimism maintained a strong presence throughout 2024, reaching a peak of 55% in December.
135. Polygon’s share varied across the period, with notable spikes, such as a jump from 3% in January to 38% in April, suggesting responsive adoption during key moments and campaigns.
136. Arbitrum continues to play a solid role in wallet deployments and active usage, and its share of transactions remains around 10%, pointing to potential for future growth in execution volume.
137. Ethereum is increasingly central for smart account deployment and active accounts, reflecting its foundational role. Its share of transaction execution also massively increased in May. Lower gas fees and more efficient execution thanks to various recent EIPs and the Pectra upgrade may further accelerate this trend.
138. Browser-based connections have dominated throughout the year, consistently making up around 80% of sessions.
     
139. Mobile wallets (including QR-based interactions) account for about 20%, with a 75/25 Android-to-iOS split.
140. Universal Wallets Blend Access and Abstraction: By combining Safe’s smart account architecture with WalletConnect’s AppKit, Reown bridges the gap between advanced smart wallet capabilities and intuitive, app-agnostic user interfaces.
141. Built for Mainstream-Ready UX: Reown’s focus on reusable sessions, cross-device continuity, and seamless smart account delegation reflects a broader shift toward wallets that feel like familiar login systems rather than crypto tooling.
142. Wallet creation speed: <200ms average time to provision a wallet after login
143. Signature speed: <20ms average latency for executing wallet signatures
144. Wallet activity: Over 183M embedded wallet signatures and 180M transactions processed year-to-date (YTD)
145. Email is still dominating with 40%
146. Twitter (35%)
147. Followed by Telegram (9%), Farcaster (6%), the only web3 method used.
148. Social apps (53%)
149. Followed by enterprise (17%), gaming (11%), fintech/trading (11%), and AI products (8%)
150. Frictionless Onboarding at Web2 Speeds: Privy enables wallet creation in under 200ms through familiar login methods like email (41% share), accelerating adoption without sacrificing self-custody.
151. Ubiquity Across App Types: With 183M signatures YTD and top usage across social (53%), fintech, and consumer apps, Privy shows that embedded wallets are now foundational to a wide range of verticals.
152. Hybrid by Default: Dynamic empowers developers to support both EOAs and embedded wallets via one SDK—allowing seamless onboarding for new users and preserving flexibility for power users.
153. Usage Reflects Fragmentation, Not Winner-Take-All: While EOAs like MetaMask and Phantom dominate by chain, embedded wallets are growing 3x faster—showing developers are building for a “wallet layer,” not a single winner.

‍

At Dune, we believe data is the backbone of innovation in web3. Our mission has always been to make crypto data accessible. With this report, we’ve showcased how wallet data isn’t just a tool for analytics—it’s the foundation for building smarter, more intuitive applications that empower users to own and navigate their digital lives.

Now, we’re taking that vision further.

On May 20th, Dune launched Sim, a brand-new multichain developer platform designed to become the backend infrastructure for every wallet, portfolio tracker, and user-facing application in web3. Sim provides instant, structured access to the most critical wallet data, with six initial endpoints tailored for seamless integration:

‍

Building a wallet?

Power your wallet app with sim.dune.com

‍

We hope you enjoyed this report!

To continue improving this picture, we invite wallet providers, analysts, and researchers to reach out and share additional data, insights, and context. By working together, we can build the most accurate and nuanced map of wallet usage across crypto.

Data inquiries:

filippo@dune.com

Want to be featured in upcoming reports like these?

Email alsie@dune.com or TG @alsieliu

Get started on Dune:

‍dune.com

Get started on Sim:

sim.dune.com

The wallet data must flow.

DisclaimerDisclaimerDisclaimer

This report reflects the current views of Dune analysts and may not be updated as circumstances change. Past performance does not guarantee future outcomes, as market conditions, regulations, and methodologies evolve over time. The information is provided “as is” without warranty, and Dune, their affiliates, or contributors shall not be liable for any damages arising from its use. No graph, model, or analysis herein should be the sole basis for investment decisions. This report does not constitute an offer or solicitation to buy/sell, and investors should seek professional advice tailored to their needs.

‍