The pitch arrived in my inbox on a Tuesday, wrapped in the familiar language of inevitability. A B2B swap provider had taken the stage at TOKEN2049 Singapore to announce that wallets should stop building their own exchange infrastructure. The reasoning was clean, almost seductive: self-built routing is expensive, small teams can't sustain it, and the sensible answer is to plug into a third party, take a cut of the flow, and get back to the business of being a wallet.
I read it twice. Then I did what I always do when a vendor hands me a conclusion and calls it a market trend: I checked who was paying for the microphone.
The speaker was the B2B and partnerships lead at ChangeNOW, a swap-as-a-service provider that has been running instant exchanges since 2017. Every substantive claim in the presentation โ that 98% of swaps execute within roughly 0.5% slippage, that wallets lost a billion dollars in the past half year trying to build this themselves, that partners receive a starting revenue share of 0.4% โ originated from that single source. There was no third-party benchmark. There was no audit. There was a Dreamster logo on a slide, a music-vertical wallet that had joined something called Fast Track, and a Bitget exploit reference to lend the whole thing urgency.
This is not a story about a company lying. It is a story about an entire revenue model being sold to wallets that cannot afford to interrogate it, and the questions that go unasked when the pitch deck doubles as the press release. The wallet monetization conversation has been hijacked by the people who profit from the most commoditized layer of it.
Let me walk you through what that means in actual numbers, because the numbers are where the story lives.
The Build-or-Buy Question Is Real, and the Answer Is Mostly Buy
Here is where I have to be fair to the pitch. If you are running a wallet with four engineers and a product roadmap that already looks like a hostage situation, building your own swap router is genuinely a bad idea. I have watched teams try. The work is not the flashy part you see in a demo โ it is liquidity venue integration, quote aggregation across chains, bridge orchestration, slippage protection, MEV mitigation, reconciliation ledgers, and a support queue that fills up the moment a user's transaction stalls on a congested chain at 3 a.m.
The engineering cost is real. A credible self-built swap stack requires at least two dedicated backend engineers, a smart contract specialist for the routing and approval logic, and an on-call rotation that can debug cross-chain failures. Add the compliance surface if you touch fiat ramps, and you are looking at a burn rate that would eat a seed round in eighteen months. For a small wallet, buying is not surrender. It is arithmetic.
But here is the part the pitch smooths over. The decision to integrate is not a technical decision. It is a pricing decision, and the pricing is invisible to the user who is actually paying it.
When a wallet routes its swaps through a third-party provider and takes a 0.4% revenue share, that 0.4% does not appear out of the provider's margin. It is layered on top of the provider's own spread. The provider's public claim is that 98% of swaps settle within about 0.5% slippage โ but slippage in the instant-swap business is almost never measured against the venue's own quoted price. It is measured against some external reference, usually a mid-market benchmark, and it excludes the spread the provider itself captures.
Do the arithmetic that the slide deck avoids. If the provider retains a spread and the wallet takes 0.4% on top, the user's all-in execution cost lands somewhere between 0.8% and 1.5% depending on the pair and the size. Compare that to a self-custodial DEX aggregator route at 0.1% to 0.3% including LP fees, or a major centralized exchange taker fee. The wallet has traded its user's execution quality for a revenue line item that the user cannot see.
I have audited routing stacks where this exact structure was buried three layers deep in the fee disclosure. The wallet looked competitive on the surface. The user was paying nearly triple what a direct route would have cost, and nobody in the product team knew because the provider owned the quote engine.
The 98% Number Is a Marketing Artifact
I want to spend time on this metric because it is the load-bearing claim of the entire pitch, and it collapses under the lightest scrutiny.
"98% of swaps execute within 0.5% slippage." Within what sample? Over what time window? Which asset pairs? What transaction sizes? Was this measured during the low-volatility Asian session on major pairs, or does it include the long-tail tokens that the same presentation insists are the wallet's differentiation opportunity?
There is an internal contradiction here that the presenter either did not notice or hoped the audience would not. The pitch tells wallets to differentiate by supporting long-tail assets โ the tokens that big wallets ignore. But long-tail assets are precisely where liquidity is thin, where slippage is worst, and where the 0.5% claim is least likely to hold. The differentiation strategy and the performance claim are in direct tension, and the presentation presents both as if they were compatible.
My own experience testing instant-swap venues against aggregator benchmarks is that the headline slippage figure holds for the top twenty pairs at retail size. Push into anything with a thin order book, or size up beyond a few thousand dollars, and the realized cost diverges sharply from the marketing number. This is not fraud. It is the standard practice of quoting the best case and calling it the average.
What would make the claim credible is a published methodology, a sample size, a time window, and an asset distribution. None of that was offered. In B2B sales, an unverifiable performance metric is not a feature. It is a negotiating lever.
The Unit Economics Nobody Runs Before Signing
Here is the calculation that should be taped to every wallet founder's monitor before they take a partnership meeting.
Take a mid-sized wallet. Fifty thousand monthly active users. Assume 8% of them swap in a given month, which is generous for a storage-first product. Assume an average ticket of three hundred dollars. At a 0.4% revenue share, the monthly math looks like this: 50,000 users times 8% participation times 300 dollars equals 1.2 million dollars in monthly volume. Multiply by 0.4%, and the wallet earns 4,800 dollars a month. Annualized, that is 57,600 dollars.
Now subtract the cost of running the wallet. Two engineers, infrastructure, customer support, compliance overhead. You are looking at a monthly burn north of 100,000 dollars if you are being honest about it. To break even on swap revenue alone at a 0.4% share, that wallet needs 25 million dollars in monthly swap volume. At a 300-dollar average ticket, that is roughly 83,000 swaps per month. That is no longer a mid-sized wallet. That is a top-decile wallet with a trading-heavy user base.
The 0.4% revenue share is the lowest-density monetization model available to a wallet, and it is being sold as the solution to wallet monetization.
The presentation makes a rhetorical move here that I find genuinely frustrating. It claims that "the community doesn't pay" โ that users who simply hold assets generate no revenue, and therefore wallets must find ways to monetize activity. But this is a false binary. Users who do not swap still spend through cards, still on-ramp and off-ramp fiat, still pay for premium features, still generate data and order flow value. The claim that storage-only users are worthless is only true if the only revenue line you are willing to build is swap commission. It is a definitional trick that narrows the solution space to exactly the product the vendor sells.
And the revenue lines the presentation does not mention are the ones with the highest density. Card interchange runs 1% to 2% of spend volume. Order flow and RFQ rebates can run mid-to-high single digit basis points on routed volume with far better unit economics than a 0.4% swap share. These are the actual profit pools in the wallet business. They are also the areas where a B2B swap provider has no product to sell.
I do not think that omission is an accident. The presentation is not a market analysis that happens to conclude wallets should integrate swap APIs. It is a swap API sales motion that happens to be shaped like a market analysis.
The Security Argument Is Backwards
This is the part that made me put down my coffee.
The pitch positions third-party integration as the safer choice โ let the provider handle the trading infrastructure, let the wallet focus on custody and user relationships. The provider absorbs the operational complexity. The wallet keeps the keys.
But integration does not reduce a wallet's attack surface. It expands it, and the expansion happens in three places the presentation never mentions.
First, every integrated swap requires users to sign token approvals. More swap surface means more approval transactions, which means a larger phishing and approval-drain surface. The wallet's security team now has to reason about approval hygiene across a routing stack it does not control.
Second, integration introduces API key and secret management risk. The wallet now holds credentials that, if leaked, could expose fee structures, route manipulation, or worse. This is a new class of operational secret that did not exist before the partnership.
Third, cross-chain swaps introduce bridge risk. The presentation explicitly encourages wallets to differentiate through cross-chain swap support. Every bridge is an additional trust assumption, and bridge exploits are the single largest category of DeFi losses in the past three years. The wallet's security narrative and its actual security posture diverge the moment it integrates.
I have seen the aftermath of a leaked partner API key. The wallet did not lose user funds directly, but it lost control of its routing configuration for eleven hours, and the resulting fee anomaly cost it more in trust than the leaked credential cost in dollars. Nobody puts that in the onboarding deck.
There is also a structural governance issue. The provider sets the revenue share, the routing whitelist, and the fee schedule unilaterally. A small wallet has no negotiating leverage โ the 0.4% is described as a "starting rate," which means larger partners get better terms and smaller partners get the worst terms in the deal. The wallet is not a partner. It is a distribution channel with a revenue share attached.
What Fast Track Actually Is
The program called Fast Track, launched in March of the same calendar year and featured at the conference, is described as a way for wallets to get traffic and support. Dreamster, the music-vertical wallet, is the showcased participant.
Read that again with the economics in mind. A wallet with no monetization infrastructure gets free integration, traffic promotion, and revenue share. In exchange, it becomes a case study. The provider gets distribution, a logo for the next conference slide, and a reference customer for the next sales cycle.
This is customer acquisition cost, dressed as a partnership program. There is nothing inherently wrong with that โ every B2B company subsidizes early adopters. But wallets should understand what they are receiving. Fast Track is a subsidy, and subsidies have expiry dates. When the promotional traffic stops and the revenue share reverts to standard terms, the wallet discovers whether it built a business or rented one.
My concern is that the wallets entering these programs are not modeling the post-subsidy state. They are modeling the launch state, with the traffic boost and the conference spotlight, and they are treating that as the steady state. I have watched this movie in the DeFi summer of 2020 and again in the NFT cycle of 2021. The projects that survived were the ones that modeled what happened when the spotlight moved on.
The Contrarian Read: The Real Money Is in the Flow, Not the Trade
Here is the angle I have not seen anyone make in the coverage of this announcement, and it is the one I think matters most.
The wallet industry has spent years chasing transaction revenue because transaction revenue is legible. You can point at a swap and say: there is a fee, here is the split, here is the revenue. It is easy to model, easy to pitch to investors, easy to put on a slide.
But the durable profit in financial infrastructure has never been in the trade. It is in the flow. Payment networks do not make their money on foreign exchange spreads; they make it on interchange. Brokerages do not make their money on trade commissions anymore; they make it on order flow, margin, and asset management. The wallet that anchors its entire business model to a 0.4% swap commission is competing in the most commoditized, most easily undercut segment of the value chain.
And it is already being undercut. The presentation itself acknowledges that large wallets offer fee-free stablecoin swaps and that this is not their primary revenue source. Think about what that implies. The largest players have already decided that swap fees are a customer acquisition tool, not a profit center. A small wallet entering the market with a 0.4% commission model is entering a price war against competitors who have decided to charge zero.
The revenue share model has a shorter half-life than the wallet's product cycle. Either the wallet follows the large players to zero and its revenue evaporates, or it maintains the fee and loses on user experience. There is no third path in this structure.
The actual moat for a wallet is not the swap fee. It is the user relationship and the signing authority โ the custody, the account abstraction, the recovery flow, the trust that makes a user comfortable holding significant value in your product. That trust is the scarcest asset in the industry, and the pitch is proposing to monetize it at the thinnest possible margin while ceding the actual execution layer to a third party.
There is a version of wallet monetization that captures the flow instead of the trade. Fiat rails with interchange. Card programs with spend-based revenue. Payroll and settlement integrations that capture a basis point on money movement rather than a commission on money exchange. These are harder to build, require licensing and compliance, and do not fit neatly on a conference slide. They are also where the actual margin lives.
The wallets that figure this out will look different from the ones chasing swap commissions. They will look more like fintech companies and less like trading terminals. And they will not need a B2B swap provider's partnership program to find their revenue.
What I Am Watching Next
The immediate question is whether the wallets entering Fast Track are modeling the post-subsidy state or the launch state. Ask the founders directly and you will get a slide, not a spreadsheet. That tells you everything.
The medium-term question is whether the 0.4% starting rate survives contact with the fee-free stablecoin swap trend. My read is that within twelve months, swap commission will be a rounding error for any wallet with meaningful scale, and the providers selling it will have pivoted their pitch to something else โ probably cross-chain routing, probably with a new name and a new conference stage.
The longer question is the one the industry keeps avoiding: if wallets are going to be the distribution layer for the next cycle, who captures the value of that distribution? The answer will determine whether wallets become financial businesses or thin clients with a revenue share.
I have been covering this industry for twenty-seven years, and I have learned that the revenue models that get pitched the loudest are usually the ones with the thinnest margins. The quiet revenue lines โ the interchange, the order flow, the settlement fees โ are where the survivors build their businesses. The wallets that understand this will still be here when the next cycle's conference slides are being written. The ones that don't will be a case study in the following presentation, used to illustrate why wallets should be careful about who they let define their business model.
The microphone belongs to whoever is holding it. The business belongs to whoever is doing the math.