Inside Trust Wallet: The Self-Custody Layer Eating the Exchange

What happens to a wallet when trading, prediction markets, and DeFi all move on-chain at once? Wallets were supposed to be the simple part of crypto: a place to hold keys and sign transactions. But as the entire stack pushes execution on-chain, the wallet is quietly becoming the most contested piece of real estate in the industry. Whoever owns the wallet owns the user. \ That is the position Trust Wallet has spent eight years building toward. With more than 220 million users across 100+ chains and roughly 35 percent of global self-custody wallet downloads, Trust Wallet is now the largest non-custodial wallet in the world, and roughly one in three crypto holders use it. Felix Fan , who joined as CEO after building products at OKX and co-founding several consumer technology startups, sat down for our Behind the Startup series to talk about why the wallet, and not the exchange, becomes the default interface for on-chain finance. \ \ Trust Wallet led global crypto wallet downloads in March 2025 with roughly 35 percent share, more than twice the share of its nearest competitor. The wallet now serves over 220 million users across 100+ blockchains, with more than $30 billion in user asset balances secured under self-custody. \ Ishan Pandey: Hi Felix, welcome to our "Behind the Startup" series. You've recently stepped in as CEO of Trust Wallet after building products at OKX and co-founding several consumer tech startups. Tell us more about your background and what drew you to Trust Wallet? \ Felix Fan: Those experiences taught me how to build products people actually use every day. What drew me to Trust Wallet was the gap between its foundation and its potential. Trust Wallet had already built one of the strongest self-custody distributions in the industry; millions of users, 100+ chains. But wallets have historically been passive: hold, send, swap. Trust Wallet already began a wallet evolution last year, adding more features for perps, prediction markets, DeFi etc. I saw the product challenge and opportunity to evolve and improve it even further. \ Ishan Pandey: Trust Wallet started as a place to store assets and you're now positioning it as a venue for on-chain trading. That's a fundamental shift in what a wallet is. How do you think about the trade-off between keeping a wallet simple and self-custodial versus loading it with trading functionality and associated capital risks for users that historically lived on centralized exchanges? \ Felix Fan: I don't think it's a trade-off, I think it's a false binary. Users don't want a simple wallet or a powerful wallet. They want a wallet that's powerful and feels simple. That's a design problem, not a philosophical one. Which is why I often use the phrase "simple by default, advanced by choice." \ But I want to be precise about what Trust Wallet is and isn't. Trust Wallet isn't a trading venue. We don't operate markets, we don't match orders, we don't set terms, we are not anyone's counterparty risk, and we don't custody anything. What we build is the interface that lets users reach protocols that do those things. Hyperliquid is the venue. Trust Wallet is the way users reach it without giving up their keys. \ The capital risks you mention, perps, prediction markets, those risks exist whether we integrate them or not. Users are already accessing these instruments, just through fragmented flows: bridging to a separate platform, creating a new account, moving funds off their wallet. Every one of those steps adds friction and, frankly, its own risks, phishing, approvals to unfamiliar contracts, funds sitting on platforms they don't control. \ Your positions sit alongside your portfolio. You see your full exposure in one place. The environment in which you access them, self-custody, full visibility of your positions, gives you more control. \ The design principle is: don't hide complexity, layer it. A new user sees a clean swap screen. A trader finds perps and prediction markets when they look for them. The wallet doesn't push anyone toward instruments they're not seeking out. \ Ishan Pandey: Let's talk about the HIP-4 integration. Trust Wallet is the first major wallet to support Hyperliquid's native outcome contracts directly in-app. Walk us through what was technically hard about integrating native prediction markets into a self-custodial mobile environment, and how you handle settlement and resolution without holding user assets. \ Felix Fan: The core challenge is that prediction markets have a lifecycle that's fundamentally different from a swap. With a swap, you execute and it's done. An outcome contract is open-ended: you take a position, it lives for hours or days, and it settles based on an external resolution. Building that into a mobile wallet means handling position management, real-time pricing, and settlement states within an interface that was originally designed for one-shot transactions. \ On the self-custody side, Trust Wallet never touches user assets at any point. The HIP-4 contracts are native to Hyperliquid L1; positions are opened, held, and settled entirely on-chain. Trust Wallet provides the interface and the signing surface. When a market resolves, settlement happens at the protocol level. We don't hold funds, we don't resolve markets, and we don't custody positions in between. \ The engineering work was about making the interface feel coherent rather than fragmented. Your outcome positions sit alongside your perps in the same place. The assets can be converted to USDC through the interface in one click, no external bridging. We wanted the experience to feel like a clean, unified way to reach the protocol, not like you're being redirected to a separate product. \ \ Monthly notional volume on Kalshi and Polymarket rose from roughly $2 billion in March 2025 to a peak of $26.8 billion in January 2026, a roughly 13-fold increase. Prediction markets are now generating monthly volumes in the same range as legal US sportsbooks. Trust Wallet's HIP-4 integration places the category one tap away inside a self-custodial environment. \ Ishan Pandey: A core part of the pitch is "no KYC, no extra accounts, full self-custody." That's compelling for users, but regulators globally are tightening their stance on prediction markets and on-chain derivatives. How do you reconcile a no-KYC, permissionless product with an environment where regulatory pressure on exactly these instruments is increasing? \ Felix Fan: To clarify, Trust Wallet itself doesn't conduct KYC, because we don't offer any services that require us to do so. Where required, for example at the fiat on-ramp level, identification happens through third party vendors. \ On the broader question, self-custody and compliance aren't opposed. In a sense, self-custody is the compliance feature: crypto failures regulators have spent years responding to were custodial intermediaries misusing customer funds (FTX, Celsius, 3AC). Self-custody actually removes a failure mode because users hold their own keys and transactions are transparent, verifiable and auditable. There is obviously a lot more work to be done here but in principle, a regulator focused on consumer protection from intermediary failure has fewer concerns, not more, with a self-custodial architecture. \ On prediction markets specifically; the instruments users access through Trust Wallet are native to Hyperliquid's protocol. Trust Wallet is an interface, we provide access, we don't operate the markets, hold the funds, or set the terms. That distinction matters both practically and from a regulatory perspective. \ That said, we're not naive about where the regulatory landscape is heading. We watch developments closely across jurisdictions. And our general philosophy is to build on open, transparent infrastructure, i.e. on-chain settlement, public smart contracts, self-custody. \ We're not trying to avoid regulation. We're trying to build in a way that works with where regulation is going, not against it. \ \ Customer-impact losses across the major custodial intermediary failures of the last decade add up to roughly $28 billion, with FTX ($8 billion customer asset hole at filing), Celsius ($4.7 billion), Genesis ($2.4 billion), Voyager and BlockFi all triggered by the same 2022 cascade. Self-custody removes the failure mode that produced every one of these events. \ Ishan Pandey: Prediction markets are speculative by design, and your own disclaimer notes users can lose their entire stake. There's a tension between "making crypto accessible for mainstream users" and putting leverage-free but still high-risk speculative instruments one tap away inside a mainstream wallet. How do you draw the line between accessibility and exposing inexperienced users to instruments they may not fully understand? \ Felix Fan: This is a fair question and one we think about a lot. The line we draw is: make it accessible, but never make it misleading. \ Outcome contracts have a clear and bounded risk profile, i.e. you can lose what you put in, nothing more. There's no leverage, no liquidation, no margin calls. That's actually a simpler risk model than many instruments already available across crypto. But "simpler" doesn't mean "safe," and we don't pretend otherwise. \ In the product, we're clear about what outcome contracts are, how they settle, and what the risks are. More broadly, I think the crypto industry has to stop treating users like they can't make their own decisions while also stopping pretending every product is risk-free. The honest position is: here's what this is, here's how it works, here's what you can lose. If you want to participate, we'll give you a clean and self-custodial interface through which to reach these protocols, not to push them toward instruments, and not to gatekeep adults out of decisions they're entitled to make. That's the line. \ Ishan Pandey: The business model is interesting: Trust Wallet currently charges zero markup, and Hyperliquid's outcome market fees are currently zero. "Currently" is doing a lot of work in that sentence. If the long-term strategy isn't fee extraction at the wallet layer, where does the durable revenue actually come from and how do you avoid the race-to-zero that has commoditized so many crypto products? \ Felix Fan: Fair observation. The zero fees on outcome markets reflect where the ecosystem is today, Hyperliquid's protocol fees on these markets are currently zero, and we haven't added a wallet-layer service fee on top. As the ecosystem matures, fee structures across the stack will evolve, and we're transparent about that. \ Ishan Pandey: You're betting that the wallet becomes the primary interface for on-chain finance, not the exchange. But that means competing with the very exchanges that often have far larger balance sheets and distribution. What is your honest read on whether self-custodial wallets can win the mainstream user, or whether most people will always default to the convenience of a centralized app? \ Felix Fan: I'd reframe the question, because I don't think Trust Wallet competes with exchanges in the way it implies. My honest read is that both will coexist, but wallets will thrive and they'll thrive more than what most people expect. \ Centralised (or decentralised) exchanges are operators. They run order books, hold customer funds, match trades, take the other side of liquidity, and act as the counterparty to their users. That's a specific business with a specific risk and regulatory footprint. Trust Wallet isn't that business. As mentioned above, we don't operate a market, we don't hold customer funds, we don't match orders, and we're not anyone's counterparty. We build a self-custodial software interface that connects users to the protocols and services they choose to use. \ So the issue is not wallet-versus-exchange. They are more about self-custody and the perceived efficiency of a single-use platform. All these players will continue to exist. Our bet is that as the on-chain ecosystem matures, more users will choose self custody because the value proposition of self-custody becomes more compelling as the surrounding infrastructure improves. \ The exchanges that do well will be the ones that lean into being excellent at what they are. \ \ DEX share of total crypto spot trading volume rose from roughly 4 percent in 2023 to 20 percent in 2025. The perp DEX share of crypto derivatives went from 3 percent to 26 percent over the same window. Hyperliquid alone now processes up to $30 billion in daily perp volume and holds roughly 70 percent of the perp DEX market. The aggregation layer is migrating to the wallet. \ Ishan Pandey: Finally, looking three to five years out, what does the wallet look like when embedded trading, prediction markets, and DeFi are all native? And what advice would you give to founders building consumer crypto products who are trying to balance genuine decentralization with the UX expectations of mainstream users? \ Felix Fan: In three to five years, a wallet won't be a thing you open to "check your crypto." It'll be the primary interface for on-chain finance, trading, earning, lending, predicting, and delegating to AI agents that navigate complexity. We're already building in that direction. \ The wallet becomes the aggregation layer. Not because it does everything itself, but because it connects you to everything, protocols, markets, agents, through one self-custody interface. The protocols compete. The wallet connects. \ Some crypto users don't read whitepapers. They don't understand gas. They don't want to think about which chain they're on. If your product requires any of that knowledge to use, you've already lost the mainstream user. \ Genuine decentralisation and great UX aren't opposed, but you have to be willing to do the hard work of abstracting complexity rather than just exposing it and calling it "transparency." The products that win the next hundred million users will be the ones where decentralisation is the architecture, not the interface. \ Don't forget to like and share the story! :::tip Vested Interest Disclosure: HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR ::: \