The AI revolution is still in its early innings. And here’s the strange thing about this moment: intelligence is getting cheap. Genuinely cheap. Capital is sloshing around looking for somewhere to go. So when intelligence and capital are abundant, what’s actually scarce?

Knowing which problems are worth the work, and having the standing to do them from the inside.

Most firms don’t get to choose. They pick an instrument first and then go looking for problems shaped like it. A seed fund needs companies at seed. A buyout shop needs cash flow to lever. The structure comes first, and the problem has to fit it or it gets passed on.

We built TriEdge the other way around. We start with a problem we understand, and choose the form afterward.

Where we come from

I spent more than a decade building and investing in healthcare businesses before TriEdge, and that experience gives us deep roots there, along with finance and real estate. So yes, we’re a family office by structure. But we never stopped thinking like builders – mostly because we don’t know how to think any other way.

Permanent family capital is a peculiar thing. On paper, it’s just a structure. In practice, it’s freedom. We can back a founder. We can buy a business and reshape it. We can start something from scratch when that’s the better path. And because the capital is our own, we can move fast when something is obvious and stand behind work that matters even when the payoff isn’t right around the corner.

That freedom comes with a responsibility to make it count. We’re grateful to work alongside founders tackling problems that could genuinely change people’s lives. The long horizon also changes how we show up for them. We can be useful before there’s a deal on the table, and without needing to turn every conversation into one. Oftentimes, the most valuable thing we bring isn’t a check. It’s a phone call, a warm introduction, or an honest read on a hard problem.

What we learned

We’ve spent much of the past two years in build mode, which is a polite way of saying: testing ideas, making mistakes, and finding out where we were wrong. The big one? We badly underestimated how hard AI adoption would be, even inside businesses we knew cold. Better tools alone don’t solve it.

What we took away is a conviction we now hold tightly: technology alone almost never changes a system. People do. You have to get the right people and institutions in the room, then build enough trust that they’re willing to rethink how the whole thing works, together. That part doesn’t scale, which is exactly why it’s valuable.

How the form gets chosen

Our venture incubations, including Mongoose, Arbiter, Anthuria and Hypha, all exist for the same reason. We kept running into problems firsthand and eventually stopped waiting for someone else to fix it. None started as an investment thesis. All three started as something that should have existed already.

And sometimes the right form isn’t a company at all. Patience matters most when the opportunity is bigger than any one company, when it lives in the institutions and infrastructure around it. In Israel, for example, mental health solutions rarely travel from pilot to deployment – not for lack of solutions, but for lack of a road: no agreed measure of a good outcome, no data access to validate against, no path from something that works to something that reaches people. No single company fixes that, because the returns from fixing it accrue to everyone.

Na’im is our attempt to build that road for the Mental Health system, deploying philanthropic or investment capital depending on what the work calls for. Much of the value will be captured by companies we have no stake in. That’s what building a precondition means.

Different forms, same sequence. Start with a problem we genuinely understand, get our hands dirty, and let the structure emerge from the work rather than forcing the work into a structure.

Where we’re going

That sequence isn’t tied to a sector. What it needs is a particular kind of problem: regulated, high-trust industries with entrenched incumbents, long procurement cycles, real switching costs, and buyers who won’t move without a reference from someone they already trust. Industries where invention is rarely the constraint and adoption always is.

Healthcare, real estate and financial services are where we’ve operated, so it’s where we know the leverage sits. Cyber, defense tech, physical AI, deep tech and applied AI have the same shape, and we’re building there deliberately – with people who already have the depth we don’t yet have ourselves.

None of it requires us to be at the center. What we prioritize is getting the right people in the same room: buyers, operators, institutions, investors, technologists, each seeing a different piece of the problem. The opportunities that matter most rarely announce themselves in advance – they come into focus once the work is underway, and by then what matters is whether the right people are already in it.

Most capital can’t choose its own form. We’d like to be the kind of firm that can.