1inch's $800B Ledger Has a Take-Rate Problem
$800 billion routed. Still not profitable.
That is the entire payload from a Crypto Briefing sit-down with a 1inch co-founder, and it should stop anyone who underwrites DeFi revenue for a living. Two numbers, six years, and they refuse to reconcile. A router that has pushed more value than most centralized venues move in their first decade cannot cover its own burn.
Here is the tell. The piece hands me a cumulative volume figure and a confession. It hands me no statistics methodology. No fee schedule. No revenue line item. No on-chain audit of the $800B itself. Gross flow is the only number that fits on a billboard โ and gross flow pays for nothing.
I learned this the hard way in 2022. Two weeks inside Terraform Labs' transaction logs, tracing the exact block where UST lost its ETH collateral peg, drilled in one lesson: the headline metric is always the one selected for you. When a protocol leads with volume and buries capture, volume is the advertisement.
Bear market rules apply. Survival over gains. So let's price the survival.
What 1inch actually is
Context matters, because "aggregator" hides the stack position.
1inch was born at ETHGlobal New York in 2019 as a hackathon build from Sergej Kunz and Anton Bukov. The thesis was narrow and correct: no single AMM holds the best price on every pair, so a router that splits and sequences swaps across many venues beats any single pool. That thesis only became load-bearing in 2020.
Uniswap V2 moved the needle. Here's how. Once constant-product pools fragmented into dozens of clones and forks, the routing problem became the actual product. Order books were abandoned, price discovery scattered, and the only durable piece of infrastructure left standing was the path between venues.
The token arrived in December 2020 โ 1INCH, 1.5 billion supply. A $175M Series B followed in late 2021 at a $2.25B valuation. Then the intents pivot: Fusion, 2022, where users sign an intent and competing resolvers bid to execute it. Cleaner UX, different economics. We'll come back to that.
Structurally, 1inch sits in the middle. Below it: EVM chains, AMM liquidity, the DEXs it routes to. Above it: wallets, front-ends, and the human clicking "Swap." That is a middleware position, not an infrastructure position. Middleware is where you get squeezed.
The competitor set is brutal and crowded. 0x, ParaSwap, Jupiter on Solana, CoW Protocol's solver auction, and โ the one that hurts โ every wallet that shipped native swaps and took the router's job in-house.
Front-end compliance is a cost line unique to consumer-facing routers. 1inch has gated access for US users, maintained geo-blocking, and staffed a legal function to keep a swap interface alive. A protocol that only wrote contracts would not carry any of that. A protocol that ships a website does. That distinction sits at the center of everything below.
The arithmetic nobody publishes
Let's run the model in public. All figures illustrative, all assumptions stated, so you can break them yourself.
$800B cumulative since mid-2019 is roughly $123B routed per year on average. Aggregator take rates, net of wallet and referrer rebates, land in the 1โ3 basis point band. Call it 1โ5 bps gross before rebates are returned. At 2 bps net on $123B, that's roughly $25M a year. Peak years better. Trough years far worse.
Now the current tape. DEX volume compresses hard in a bear market, and aggregator flow compresses harder, because the reflexive arbitrage that inflates it disappears first. Assume routed volume falls to $40โ60B annually. At 1โ3 bps net, that is $4Mโ$18M of protocol-level revenue.
Cost side. A top-20 DeFi protocol runs 60โ100 engineers fully loaded at $200โ350k, which alone is $12Mโ$35M. Add audits and bug bounties: $2โ5M. Legal, entity maintenance, sanctions screening on a user-facing front end: $3โ10M. Grants, RPC, indexing, data infrastructure: $5โ15M. Liquidity and ecosystem incentives: $10โ30M, and that line never shrinks in a downturn because it is the only lever that moves the volume number.
Realistic annual burn: $35Mโ$80M.
To break even against that cost base on $50B of routed volume, 1inch needs 16โ25 basis points net. The market clears routers at 1โ3. That is not a margin problem. That is a 5โ20x structural shortfall.
Here is the mechanism underneath it. The marginal cost of adding one more AMM to a router is effectively zero. The marginal revenue of adding one more AMM is also effectively zero. When both are zero, competitive equilibrium pins price at zero โ no cartel, no coordination, just Bertrand logic compiled into Solidity. Uniswap's own router is free. Wallets' built-in swaps are free to the wallet and monetized through the wallet's own order flow. The router is priced like a commodity because it is one.
Now the intents pivot. Fusion is genuinely better execution โ I've tested it, and the gas abstraction and MEV protection are real. But architecturally the protocol becomes an auctioneer. Auctioneers don't earn spread; they earn the minimum clearing fee that keeps solvers solvent. You have moved execution risk off the user and moved profit off the protocol. Solvers compete, surplus converges to zero, and 1inch collects the auction house's cut โ structurally thinner than the spread it replaced.
There is also the switching-cost problem, and it cuts one way. Wallet defaults are sticky for the wallet, not for the protocol. A front-end can swap its default router in a sprint. Users don't notice, because users never chose the router in the first place โ they chose the interface. The route is an implementation detail to them.
And there is the execution tax. Every extra hop in a route costs gas, and on L1 that overhead eats the price improvement that justified routing. Gas spike detected. Run. Routers win on net outcome, not gross price improvement โ and net outcome is what a wallet's native single-hop swap can increasingly approximate.
CoW Protocol runs the same auction logic with a narrower mandate. Jupiter owns the Solana flow with a token that has a real fee surface. 1inch competes against both while carrying L1 gas overhead the first avoids and chain dominance the second holds. That is a hard corner.
The $800B denominator problem
The number itself deserves stress-testing.
Cumulative routed volume, as a category, counts repeat bots, MEV backruns, arbitrage loops, points-farming round-trips, and internal routing hops. In my Terraform audit, a meaningful share of "activity" was the same small capital base cycling through a reflexive loop. ERC-20 rush vibes. Proceed with caution.
If even 20โ30% of that cumulative $800B is reflexive arb and incentive-driven churn, the economically meaningful base shrinks. And worse: the flow that disappears first in a downturn is precisely the flow that paid the least. Incentive-farmed volume is negative-margin volume. You pay grants and points to acquire it, then it leaves and takes the fee revenue with it.
No fee schedule was published. No revenue figure accompanied the volume figure. No methodology behind either. That absence is the real disclosure failure, and it is more informative than the milestone ever was.
The contrarian fix that will not work
The consensus next move is B2B. Institutional API revenue. Sell routing to desks.
Stress-test that.
A trading desk's build-versus-buy decision on a smart order router is a two-week project for two competent engineers, sitting on top of FIX connectivity and a legal venue list that already exists. Institutions do not need a public-chain router. They need a cleared execution path and a custody arrangement. An aggregator's API, for them, is a market-data probe for pre-trade color โ not a venue they will send size through. RWA on-chain has been a three-year storytelling exercise for exactly this reason: the institutions being pitched already own the rails they trust.
The B2B revenue that does exist is not technology revenue. It is order-flow revenue โ paying a wallet to make your router the default slot. That is distribution rent, a marketing expense wearing a SaaS costume. It scales linearly with payment, and the moment you stop paying, the default flips.
Second false fix: the fee switch. A fee switch on a protocol with no net revenue is a tax on users with a one-click exit. You cannot tax flow that can leave in a single click, and routing is the most switchable behavior in all of crypto.
Both fixes share one flaw. They assume the router's problem is pricing. It isn't. The problem is that routing generates no defensible surplus. The honest pivot is not selling routes โ it is selling the only thing that isn't commoditized: exclusive order flow, or verified execution quality that a desk cannot replicate in two weeks. Everything else is repackaging.
What to watch
Three signals, in order of value.
One: a quarterly protocol revenue disclosure with an actual methodology attached โ fee line, rebate line, solver payout line. Whoever publishes that first re-rates the entire sector.
Two: whether Fusion and resolver payouts surface as a cost center rather than a marketing bullet.
Three: default-slot deals with named wallets, disclosed in size. That is the only revenue that isn't theory.
If the next milestone 1inch announces is another cumulative volume number, that's the answer. The first DeFi protocol to publish audited net revenue will be the first DeFi protocol anyone can actually value. The race is no longer for volume. It's for a denominator.