The question of "clustering" is an old one in economics, and also politics and public discourse. PG addressed it a few times, for example. He also write about h

The question of "clustering" is an old one in economics, and also politics and public discourse. PG addressed it a few times, for example. He also write about how to create new centers, fwiw. It's not a straightforward question, and I don't think any answer is entirely satisfying. A couple of cents though. 1¢ - Industries like tech are a cultural phenomenon, like a music industry or art industry. You can't just decide that Warsaw will have Seattle's grunge rock scene. It just grew there as a cultural outgrowth. It can't be planted elsewhere with predictable results. The thoughts and little decisions made by people are influenced by their cultural context. 2¢ - For the best (in business terms) tech companies, currently, costs like salaries mean nothing. Twitter and FB were Twitter and FB with 1% of their current visits and employees. The extremely profitable and businesses enabled them to hire tens of thousands of people and spend tens of billions per year. We're these people necessary to the ad part if the business that employs most of them? I suspect that they aren't. IE, if it turned out FB could only make $40bn or $8bn pa (currently making $80bn) from their site, they would manage to produce the same fb with few compromises on a much reduced budget. This isn't BMW, where operating costs and economic output (the number of cars) are tightly related to eachother. If google or FB want to increase profit margins, they can, easily, by a lot. They would if they needed to, but they don't. Capital markets like the idea that they are reinvesting for growth and capital markets don't have any better ideas for where to invest profits anyway. With no incentive to cut costs, why move somewhere cheaper?