Picture the moment you step off a plane in a city you've never worked in before, carrying a résumé that would get you a meeting in thirty minutes back home. You send the emails. You wait. Nobody here knows your name, your reputation, or the two years of hard-won judgment you carry around in your head. The interviews are polite and go nowhere. This is not bad luck. This is the cluster working exactly as designed, just not for you.
The question of why industries concentrate in single cities, often cities that seem poorly suited to the task, is one of the more interesting puzzles in economics. The short answer is that geography almost never explains the cluster. History, accident, and a set of self-reinforcing mechanisms do.
The first mover sets the trap
The initial location is usually a historical accident, and that part is widely understood. What's less appreciated is how quickly the accident becomes irreversible.
When a critical mass of firms in the same industry settles in one place, three things start happening simultaneously, and each one makes the cluster stronger. First, a specialised labour pool forms. Workers with niche skills, the kind you can't train in a semester, migrate toward the concentration because that's where the jobs are. Firms then locate near that pool because hiring from it is cheaper and faster than relocating specialists from elsewhere. This loop closes fast.
Second, specialised suppliers appear. A city with forty biotech labs will attract the one firm that makes custom cell-culture equipment, the one that handles biohazardous waste on short notice, the one that calibrates centrifuges. Those suppliers wouldn't exist in a city with two labs. So the fortieth lab in the cluster gets services the first lab in a competing city simply cannot access at any price.
Third, and this is the part most guides skip: knowledge leaks. Not through corporate espionage, but through lunch. Through the engineer who leaves Firm A on a Friday and joins Firm B the following Monday, carrying two years of tacit knowledge in her head. Through the industry conference that only makes sense to host where everyone already is. Through the shared vocabulary that forms when enough smart people work on adjacent problems in the same zip code for a decade. Economists call this a knowledge spillover, but that phrase undersells how visceral and fast it actually is.
Why 'just move' is harder than it sounds
The obvious objection: if land in San Francisco costs three times what it costs in Phoenix, why don't firms simply relocate and pocket the difference?
Some do. The calculation, though, is more brutal than it appears. Imagine a mid-sized venture-backed company, twelve engineers, a sales team, a product that competes in a market where the top twenty firms are all within a forty-minute drive of each other. The CEO prices out offices in a secondary city. Rent drops by sixty percent. Then she models the actual costs: the three senior engineers who won't move, the recruiting disadvantage in a city where candidates don't already know the company's name, the loss of the informal intelligence network, the harder path to the next funding round from investors who prefer to back companies they can reach for coffee. The savings evaporate. Often they go negative.
This is the trap. The cluster is expensive precisely because it works, and it works in ways that don't show up on a balance sheet until you leave and notice their absence.
The geographic disadvantage that stops mattering
Here's the wrinkle: the industries that cluster most tightly tend to be the ones where tacit knowledge matters most and physical inputs matter least.
Film production in Los Angeles is the textbook case. There is no geographic reason whatsoever for Hollywood. Timber for sets can be shipped. Cameras are manufactured in Japan and Germany. The sun, which early directors did actually need, shines in Florida too. But film, like software and finance, is a business built on knowing who can solve which problem, who owes whom a favour, who will green-light a project if the right producer attaches their name. That knowledge lives in people. The people are in Los Angeles because everyone else is.
Contrast this with steel production, which historically located near coal and iron ore deposits because moving the raw material was prohibitively expensive. Once those physical constraints ease, or once the industry's value shifts from raw processing to specialised design and services, the cluster can migrate toward talent rather than geology. Which is exactly what happened to the American steel industry over the course of several decades. The industries that defy their geography are, almost without exception, the ones that have already made that transition.
What people get wrong about this
The popular version of the clustering story leans heavily on the founding myth: Stanford and Shockley Semiconductor in Silicon Valley, the garment traders who settled the Lower East Side and built New York's fashion industry, the coffeehouses of seventeenth-century London that became Lloyd's of London. These stories are true and they're useful. But they invite the wrong conclusion, which is that a city can deliberately replicate the process by building the right university or the right incubator.
Consider two colleagues, Marcus and Priya, who both worked at the same fintech startup in New York in its early years. Marcus eventually moved to a well-funded tech hub that a regional government had built from scratch, complete with subsidised office space and a new engineering school. Priya stayed. Five years later, Marcus's hub had attracted a few dozen companies but struggled to retain senior talent, which kept leaving for cities where the deal flow was thicker and the serendipitous collisions more frequent. Priya's network had compounded. Her second company was funded in a week, partly because three of her potential investors had watched her operate for years.
The mechanism that makes clusters work, the accumulated social capital and knowledge density, takes a generation to build and cannot be conjured by a ribbon-cutting ceremony. Governments that fund these projects are not wrong to try. They are, however, almost always wrong about the timeline, and the history of top-down cluster-building is littered with expensive ghost campuses that planners still describe as works in progress.
The cost that concentrations impose
None of this is free. The people paying the highest price are not the firms.
When an industry locks into a single city, it creates a geography of exclusion. The talented person who can't afford to live in the cluster, or who has family obligations that make relocation impossible, is cut off from the industry's best opportunities regardless of ability. The same density that produces so much knowledge spillover and innovation also produces a housing crisis, a commute crisis, and a sorting mechanism that correlates career success with the willingness and ability to move to one specific, expensive place. A cluster, seen from the outside, looks less like a thriving ecosystem and more like a walled city that charges admission in the form of rent and rootlessness.
The remote-work shift that followed pandemic-era disruptions offered a genuine test of whether knowledge-intensive industries could deconcentrate. The evidence is mixed. Some dispersion occurred, but the most senior roles and the densest networks remained stubbornly tied to the original clusters. It turns out the lunch conversation and the hallway collision are harder to replace than anyone admitted when the offices were empty.
The cluster persists not because firms are irrational, but because the advantages are real and the exit costs are steep. Geography didn't create it. Geography won't dissolve it. So ask yourself: the next time a politician promises to build the next Silicon Valley in a mid-sized city with a generous tax break and a ribbon-cutting ceremony, who, exactly, is that promise actually for? The cluster endures because it serves the people already inside it. The question worth sitting with isn't why the industry is there. It's who gets left out by the fact that it is, and whether anyone in power considers that a problem worth solving.