Concentration is the strategy: why a five-brand portfolio is the right shape.

A reasonable question for any studio is: why not launch twenty brands instead of seven? AI tooling makes small teams dramatically more productive, so in theory a studio could originate far more bets than we do. The honest answer is the most important thing we can say about how we operate: concentration is not a phase on the way to a bigger portfolio. It is the strategy.

The case against spraying bets

The default assumption in venture is that more bets is better: more shots on goal, more chances one becomes a category winner. That logic makes sense for a fund investing small checks into other people's companies, where the fund's job ends at the check and a board seat. It makes much less sense for a studio, because we are not writing checks. We are staffing, building, and operating every brand ourselves.

Every additional brand we start draws on the same scarce resource: the attention of our engineering, design, and operating talent. That attention is what makes the AI-native build model actually work, small teams shipping at large-team velocity because they have real support from a shared platform and real focus from studio leadership. Spread that attention across too many brands at once and it stops being a craft and starts being a checklist, and the whole advantage of building AI-native evaporates.

Twenty half-staffed brands is not a portfolio. It is a backlog with a landing page attached.

What running seven brands deeply buys us

1. The studio can name every customer that matters, per brand

Across seven brands, studio leadership can hold each brand's top customers, biggest churn risks, and clearest expansion opportunities in working memory. That changes the quality of every product and go-to-market decision. Capital and engineering-time requests get answered with real context, not a generic process.

2. Operator-to-operator learning is a real, running conversation

Seven operating teams is small enough that they actually know each other. A pricing experiment at RenewalPad informs a renewal-motion decision at TempoDocs. An engineer who built the AI feedback loop at Lucidly can spend a quarter helping Attribufi get its own loop right. At thirty brands, that kind of cross-pollination requires a formal program. At seven, it just happens.

3. Build capital and engineering time get allocated as real decisions

With seven brands, every allocation of build capital and senior engineering time is a specific, informed call about which brand needs it most this quarter. We wrote the mechanics of that in allocating build capital and engineering time across studio brands. At a much larger brand count, allocation collapses into a formula, because there is no other way to manage it. Formulaic allocation is exactly what a studio's hands-on model is supposed to make unnecessary.

Why concentration is honest about risk, not naive about it

The standard objection to concentration is that it is risky, and in a pure venture-fund context that critique is fair. In a studio context where we staff and run every brand ourselves, the risk calculus flips. Our real risk is not that one of seven brands underperforms. Our real risk is that studio attention gets diluted so thin that none of the seven gets the depth of build and operating support that made starting it worthwhile in the first place.

We also diversify deliberately across categories rather than brand count: fund accounting adjacent tooling, advisor onboarding, sales enablement, renewal management, document workflow, attribution, and fractional recruiting. Those categories share almost no end-market overlap, so a slow quarter in one is uncorrelated with the others. That is the diversification we believe in: structural, not numerical.

When we will and won't start a new brand

We start a new brand when three things are true at once. The idea clears our pre-build validation bar, which we describe in our pre-build validation checklist. The studio has the engineering and design capacity to build it without pulling meaningful attention from the existing seven. And there is a specific person, either from inside the studio or an outside operator with the right domain expertise, who wants to run it for years.

We will not start a brand to hit a headcount, to make a deck look more impressive, or because a category is trending. None of those reasons survive a decade of actually operating the thing.

• Yes: a workflow we understand deeply, with a studio engineer or outside operator ready to run it for the next five years.
• No: an adjacent feature to an existing brand that would tempt us to bolt it on rather than build it properly.
• No: a category with strong buzz but no operator who actually wants to live inside it.
• No: any idea whose primary case is "we could spin this up fast," with no plan for who runs it after launch.

Why this is the right shape for an investor

Anyone evaluating Cobalt Glacier is not betting on the breadth of the roster. They are betting on the depth of the build and operating support behind each brand on it. That depth is what turns a promising idea into a durable, AI-native company. Diluting it across more brands than we can serve well is the fastest way to break the whole model.

Seven brands today, run with the focus of a small, senior operating team rather than the routine of a checklist-driven back office. A handful more over the next several years, added one at a time against the same bar. That is the shape of the studio.

Investors who want to understand our pace and capacity can reach out here. Operators who want to co-found the next brand can start that conversation here.