Where We Are Going

The first three volumes answered "who we are and how we work." This volume asks a different question: where are we going.

The view up front. Eight years, unchanged: choose the North Slope of an industry, and look for the assets that will lead the next paradigm shift. This is not a slogan. It is the ruler we use to rank our attention.

Our attention narrows to four directions: AI foundation models and wholly new paradigms; physical AI, embodied intelligence and robotics; commercial space, quantum computing, controlled fusion and brain-computer interfaces; and one layer below, AI infrastructure — advanced compute, power and energy, optical interconnect, edge computing. A fifth line runs across all of them: intelligent manufacturing and hard-tech products that can compete globally. That is the whole list. No more, no less.

Why only these? Because we cannot cover more. We do twenty or thirty mandates a year, half of them new, half repeat business from existing clients, and a partner's attention only goes so far. The denominator is fixed.

Four Directions, Really One Chain

Why these four? Because they share one industrial chain. The last leg paid out on "electricity" — electric vehicles, lithium batteries, solar. This leg pays out on "intelligence". Same script, new cast. They also share the same group of phenomenal entrepreneurs. Once you see the chain, all four directions have coordinates, and ranking what comes first and what weighs most runs on one reusable frame of reference.

So what carries over from the last leg? People, production lines and engineers. That makes our own position clear: be the one who orchestrates technology, industry and long-term capital into a single arrangement. Being an Orchestrator comes down to three things — understanding, matching, closing.

Seeing It Clearly Is Not the Same as Acting Now

But one discipline comes first. Judging a direction is really two questions: is this direction right, and when is it right. Which one do you answer first? The second.

Do the arithmetic first. For directions like controlled fusion and quantum computing, the technical judgement may be entirely sound. But their commercialisation runs longer than a fund's life — a payoff window of ten years and more against a liability structure of seven or eight, with two or three fundraising rounds and staff turnover in between. So "not committing heavily for now" is not the same as "not believing in it". The real limit of a firm's capability is written in its fund life, not in its understanding. No need to be coy about that.

Does that mean we stop watching them? We watch. We watch from the bench. But money and people go first to where things resolve within three years.

The mechanics of this leg differ from the last. Internet channels had to be laid inch by inch. AI distribution is already in place, so a good thing spreads far faster than it did a generation ago. We are fairly sure of this one.

The Hotter the Field, the Cooler You Should Be

So should you bet more where it is hot? Just the opposite.

Money does flow towards heat and towards directions with room to absorb it. True enough. But heat is not the same as odds of winning. As things stand, the problem at home is not that nobody is doing the work. It is that nothing converges: a consensus barely forms around a direction and every company with a tenuous claim says it is in the field, resources thin out, and the price war arrives early. In a market like that, picking people and picking the moment weigh more than picking the direction.

So the practice is plain: pick a few good fields, and back phenomenal entrepreneurs heavily. Direction is only a coordinate. People are the variable.

The next few pieces lay out, one at a time, what we see along this chain: how robots move from demo to volume production, where the real question in driverless logistics sits, what low altitude actually is, and how to rank the four frontier lines.

No hurry. One piece at a time.