Our founding thesis: money is becoming machine-readable

A useful way to see where value ends up in a technology shift is to wait a few years and ask who quietly came to own the boring part. Not the app everyone downloaded. The rails underneath it: the marketplace, the identity layer, the settlement network, the thing every other product had to plug into and could not route around. In the last cycle that boring part turned out to be worth more than most of the apps combined, and it was captured early by a small number of companies who understood what they were holding and refused to let go.

We started First Tech Group on a bet about where the boring, load-bearing part of the next cycle is being built. The short version fits in a sentence: at the place where programmable money meets machine intelligence. This is the long version, and the argument for why we decided to build and operate it rather than fund it from a distance.

Two curves that were always going to meet

The first curve is the slow conversion of value into something software can read and move on its own. Cash, treasuries, equities, credit — instruments that used to live in closed, permissioned ledgers are being represented as programmable tokens that settle on shared infrastructure in seconds, at any hour, without a chain of intermediaries reconciling each other after the fact. Stablecoins were the proof of concept, and a boring one at that: a dollar that a program can hold and send. The direction of travel is the rest of the balance sheet.

The second curve is the shift of AI from a thing you prompt to a thing that acts. A model that can read your intent, remember your situation, and take steps on your behalf is only interesting if those steps can include doing — and increasingly, doing something means moving value.

Lay the two curves over each other and the conclusion is almost boringly logical: value that software can read, software will eventually move. The interface between money and machine intelligence stops being a curiosity and becomes a control surface — one of the more important pieces of infrastructure of the decade, and one that barely exists yet.

The problem nobody is pricing in

Here is the part that convinced us to start a company rather than write a memo. The pieces of this future are being built right now, but they are being built separately, by companies that compete with one another and have no reason to cooperate. An exchange over here. A wallet over there. An AI lab somewhere else. Each is a reasonable business. None of them, alone, can capture the thing that will actually matter.

Because the value at this convergence does not sit inside any one layer. It sits in the seams between them — the shared identity that lets a wallet and an exchange recognize the same person, the shared memory that lets an assistant carry context from a conversation into a transaction, the settlement guarantees that let one service trust another's money. Seams are worth enormous amounts and they are structurally impossible to build if you own only one side of them. A wallet that has to treat the exchange as an untrusted third party, an AI that cannot settle into a real market, a user reintroduced as a stranger at every app boundary — these are not edge cases. They are the default outcome of a market that builds the layers and forgets the joints.

That gap is the whole opportunity. A fragmented digital economy will leave the connective tissue on the table, and someone is going to pick it up.

Why funding it is not enough

The instinct of most of our peers is to treat this as a portfolio problem: back a wallet, back an exchange, back an AI company, and hope the market assembles them into something coherent. We think that is the wrong shape, and not by a little. You cannot build a seam by owning shares in the two companies on either side of it. The identity layer, the shared memory, the settlement trust — these have to be designed together, operated together, and answerable to one roadmap. A board seat does not get you there. Only ownership and operating control do.

So First Tech Group is an operator-investor, and the hyphen is load-bearing. We take a small number of deliberate positions in the layers that have to exist, and then we do the unglamorous work of running them — the matching engine and custody discipline of an exchange, the key security and recovery of a wallet, the memory and orchestration of an AI system. We build the connective tissue on purpose, because it is the one thing a market of rivals will never build for us.

What we own, and why it is one thing

We hold one position in each of three layers — a venue where value trades, a wallet where it is held, and an intelligence that acts on both — and we connect them with a single identity-and-memory backbone. Read individually, they are three companies. Operated together, they are one system with a compounding loop running through it: liquidity makes the wallet useful, the wallet distributes the intelligence, the intelligence deepens the markets, and every interaction enriches a memory the whole group shares. We are not assembling a portfolio. We are assembling a stack, and holding the joints.

The bet, stated plainly

The wager underneath all of it is that in infrastructure — where the hard problems are latency, custody, key security, model orchestration, and regulatory trust — the teams that build and run the thing will out-compound the teams that merely finance it. Operators learn what is real faster. They can originate the companies a thesis requires instead of waiting for them to walk in. And only they can own the seams. If that bet is right, the advantage was never any single layer. It was the connective tissue, held long enough to compound.

What we are not claiming

We are early, and we say so. Our exchange is in active build, not live. Our wallet and AI products are in development. This is a long game measured in years, the outcome is not decided, and anyone who tells you the winners of this convergence are already known is selling something. What we are claiming is narrower and, we think, defensible: that the most valuable infrastructure of the next era lives at the seam between money and machine intelligence, that a fragmented market will leave that seam unbuilt, and that the way to capture it is to own and operate the layers on either side of it rather than to fund them and hope.

That is the founding thesis. Everything else we publish is a footnote to it.