Almost every fintech that talks to us begins in the same modest place. They want one product, usually yield on stablecoins, occasionally a crypto-backed loan, and they want to offer it to their users without hiring an entire crypto team to make it possible. It is a reasonable ask, and for the first few minutes of the conversation it is the only thing on the table.
Then the conversation drifts, as it always does, toward something larger. If we can offer yield, they ask, can we also do lending? What about tokenized equities, or gold, or a diversified basket, all held against the same balance? By the time the call ends they are no longer describing a feature they want to bolt on, but a wallet: a single place where their users hold everything and, increasingly, do everything.
We have come to read that drift as the whole story of the market. Neobanks, payment apps and exchanges are all, at different speeds and from different starting points, becoming wallets. Not key stores, but full financial applications in which yield, lending, perpetuals, tokenized equities, real-world assets and fixed income sit behind one balance and one login.
The Monzo model, moved onchain
It is worth saying plainly that none of this is new. It is the model traditional fintech has run for years, rebuilt on different rails.
When you open Monzo or Revolut and decide to put some money to work, the only choice you are really asked to make is how much risk you are comfortable with. You pick low, medium or high, and the matter is settled. What happens beneath that choice is invisible to you: a professional asset manager, often BlackRock, is allocating your money across real markets on your behalf. You expressed a preference, and an institution did the rest.
The onchain version has the same shape and a different set of plumbing. A user selects a risk level, and underneath, rather than a fund, their money is put to work through onchain protocols. The experience is identical, down to the reassuring sense that you made one decision and something competent took care of everything after it.
What has changed is not the experience but the manager itself, which has quietly moved from a firm into code.
Coinbase has already shipped a narrow version of this, placing yield from a single venue in front of its users. But a single venue is not what most platforms are after. They want access to the whole investable universe, and they want to decide for themselves which part of it their own users are shown.
That is the piece nobody has really built yet, and it is precisely the direction in which everyone is now moving.
The categories are dissolving
Not long ago the boundaries were easy to draw. Equities lived in one place and crypto in another; public markets were open to anyone, while private markets remained the preserve of whoever was wealthy enough to receive the call.
Tokenization dissolves those boundaries almost as a side effect.
Once a Treasury bill, a share of Apple, a barrel of oil and a stablecoin position can all exist as tokens in the same wallet, the categories that used to separate them stop meaning very much to the person holding them. What is left is a balance and the freedom to move between assets at will.
The line between asset classes is the first to blur, and the line between public and private markets follows soon after, since the same infrastructure that tokenizes a share can just as easily tokenize a fund.
The interface stops being a screen
When we released the Compass CLI, an agentic way into every one of our products, the response confirmed something we had half suspected. Teams are no longer asking merely for a better screen with more buttons on it.
They want their users to be able to say, in plain language, "move my stablecoins into the best yield you can find," or "buy me some tokenized Apple," and have the instruction carried out, with the agent reading the market, pricing the transaction and executing it.
Robinhood is already offering a version of this for equities through a conversational interface. Once the underlying products live onchain and are reachable through a single API, placing an agent on top of them is not an ambitious leap but the natural next step.
When the manager is only code
Follow that logic far enough and the manager itself changes character. When a strategy is expressed as an audited smart contract that runs on its own, there is no fund and no quarterly letter, only code doing the work, into which anyone can deposit and, more importantly, into which anyone can look.
That last quality is the one that matters most. After two years of hacks and depegs, almost every serious question we are asked is a version of the same one: what, precisely, am I exposed to?
A well-run fund will answer that question in its statements, and answer it well. What onchain infrastructure adds is not disclosure where there was none, but immediacy and verifiability, a position-by-position picture available in real time rather than at the close of a reporting period.
The remaining difficulty is legibility, because raw onchain data is a long way from a clear understanding of one's own risk, and narrowing that distance is a meaningful part of what we do.
Whoever holds the wallet holds the customer
When a single application holds everything and does everything, that application owns the relationship, and that ought to concern a good many incumbents.
For a large share of customers, the app in front becomes the thing they genuinely bank with, and the institution that once sat in the middle becomes optional. Whoever controls that front end controls the money. A handful of incumbents have understood this and are building onchain products directly into their own apps rather than waiting for the change to be done to them; most are not moving at all, and it is that inaction that leaves the opening.
The shift is real, and it is nearer than the headlines suggest. It is not that crypto is going mainstream, which is both true and uselessly vague, but something more specific: the consumer application absorbs the product layer, asset management is expressed as code rather than as a fund, and whoever owns the front end owns the customer.
Why it has not happened already
If the direction is so clear, it is fair to ask why every app is not already there.
The answer is that the work in the middle is genuinely punishing. To offer even a few of these products in-house, a team must integrate dozens of protocols, each with its own idiosyncrasies, handle gas, manage non-custodial signing, bundle transactions, monitor risk across every venue and keep the whole apparatus secure. That is six to twelve months of specialized engineering directed at something that is not the team's actual product, and the work is hostile in a way that is hard to appreciate from the outside.
When engineers with doctorates describe onchain integration as painful, it is safe to assume it is blocked for almost everyone downstream of them.
The missing middle
That missing middle is what we have built. Compass is a single API for yield, lending, trading, tokenized equities, global markets and the risk data that underpins all of them. Six products through one integration, roughly a week to go live, and non-custodial throughout: we assemble the transaction, the user's own signer approves it, and at no point do the funds pass through us.
This is not a promise about the future but something already in production, with Para relying on us as its onchain layer across more than a hundred applications, and Thor Wallet running on us as well.
The most compelling version of all this, we suspect, will not appear first in New York or London. It will appear where local inflation makes dollar-denominated yield and stable savings something people actually need, across Latin America, Africa and Southeast Asia, where the move from speculation toward ordinary, useful, everyday finance carries the most weight.
That is the direction of travel: every app becomes a wallet, every wallet carries the full menu, and the infrastructure that makes it possible recedes into the background, in the way card processing long ago became invisible.
We are not there yet, and business-to-business change moves more slowly than any thesis cares to admit. But the shape of the thing is clear, and we are building the layer it will run on.
Ready to build?
Your users are already asking for this, and you do not need a crypto team to give it to them. You need the layer underneath, which is precisely what we provide: non-custodial, a single integration, live in roughly a week, with your product still the one your customers see.
Start at docs.compasslabs.ai, or speak to us at compasslabs.ai.