Geographic Access and the Future of Portfolio Intelligence

Institutional-grade allocation technology has been concentrated in a handful of zip codes for three decades. Manhattan, Greenwich, Mayfair, Geneva, a thin stripe of coastal California. That concentration is a historical artifact of how information was distributed, and software is quietly dissolving it.

The dissolution is already well under way, but the implications for distribution — who gets to offer what to whom, and at what margin — are only now becoming visible to operators outside those zip codes.

Why the concentration existed

Sophisticated portfolio management grew up in a physical world, and the price list of that world is worth writing down. A Bloomberg terminal lists at roughly $27,000 per seat per year, and there are on the order of 300,000 of them — a distribution layer worth several billion dollars a year that had to be bolted to desks somewhere. Research desks were rooms full of people reading wire-service notes that arrived in printed form. Relationships between research producers and capital allocators were maintained through in-person meetings, because no other channel had comparable fidelity.

One economic consequence followed directly: any client outside a reasonable travel radius of a major financial center received a degraded version of what clients inside that radius received. They received it later, they received less of it, and they received it filtered through intermediaries whose margin was compensation for the geographic distance.

A client of a regional wealth advisor in, say, Columbus or Lyon or Perth was not getting worse advice because the advisor was less skilled. They were getting worse advice because the advisor was further from the source of the information, and the source operated at a speed that favored proximity.

What changed

Two things changed, almost simultaneously, over the past five years:

1. Institutional-grade research and allocation tools became deliverable as pure software, and the input costs collapsed with them. Compute that a desk once bought as six-figure hardware rents for around $0.10 an hour. Macroeconomic data that once justified a terminal seat — FRED alone serves over 800,000 series — is free, as is every SEC filing back to 1996 and, since 2019, commission-free execution on US equities at the major retail brokers. The terminal is now a browser tab, the research desk an API, the rebalancing logic a cloud function.
2. Distribution latency collapsed from "overnight by courier" to "sub-second by websocket". A signal computed in Stamford is available in Stockholm, São Paulo, and Seoul at the same millisecond.

Together those two changes dissolve the geographic concentration. When both the tools and the signal are delivered by software, the location of the receiver stops mattering for the quality of the product. What remains is the question of who has the distribution, and that question is open for the first time in a generation.

The moat around sophisticated allocation was never a moat of skill. It was a moat of proximity. Software removes the moat without removing the skill.

Three consequences for distribution

Consequence 1 — Tiered access collapses toward a single tier

Tiering by wealth — family offices get one level of product, mass affluent another, retail a third — was an economic necessity when the product required physical distribution, and it was enforced by hard thresholds. Hedge fund access in the United States sits behind the accredited-investor tests ($1 million in net worth excluding the primary residence, or $200,000 in individual income), private banks commonly set minimums of $5 million and up, and the strategies themselves add investment minimums on top. None of those numbers measures the sophistication of the question a client is asking. Tiering becomes an economic choice when the product is software, and firms that keep pricing on asset size rather than on the question being asked will lose share to firms that do not.

Consequence 2 — Intermediaries compete with sources

A regional wealth advisor in Columbus is no longer a filter between the client and the source of research. The source is accessible directly. The advisor's value proposition has to shift from "I bring you information you couldn't otherwise get" to "I help you interpret information you could easily get yourself". That is a different business, with different margins.

Consequence 3 — Non-US capital is no longer under-served

Historical under-service of capital outside the US, UK, and Switzerland was partly a staffing problem: firms didn't have local analysts, so they didn't have local products. Software doesn't require local analysts. The quality gap between a $50M portfolio in São Paulo and a $50M portfolio in Greenwich shrinks to near-zero, and the fee gap becomes harder to justify on any basis other than inertia.

The question is no longer whether sophisticated allocation reaches smaller and more distant clients. It is which firms figure out how to profitably distribute it to them first.

What this means for Quark

Our architecture is built for this dissolution, not around it. Our research stack runs in the cloud, our signals are delivered over HTTPS, our model portfolios are accessible from any browser, our checkout works with any card that operates in international commerce. The client in Columbus and the client in Zurich receive identical product. The question that determines whether they subscribe is not geographic — it is whether the product is worth its price.

For a modern research firm this is the correct baseline rather than a differentiator. Incumbents whose pricing depends on geographic scarcity have a more difficult adjustment ahead.