There is no "death of distance", no matter how cheap communicating becomes. Rather, the many advantages of clustering will continue to be attractive -- but available over a larger space so that the many costs of crowding (including congestion and high-rise development) can be avoided. Lower Wall Street represents a tight 19th-century cluster whereas Silicon Valley is a spread-out modern cluster (even spanning SF Bay by some estimates). Pierre Desrochers has been writing about this evolution for some years. He evokes Jane Jacobs, among others, describing the variety of actors and roles that create fertile soil for innovation.
The trouble is that planners and policy makers around the world have embraced the idea that they can create and/or manage successful clusters. What if their heavy hand stifles creativity instead? What if this variant of industrial policy pushes resources to the wrong places at the wrong times? To ask the question is to answer it.
In the face of all this, many scholars are looking at models of increasing returns and path-dependence. These forces cannot be denied. Rather, the question has to do with how important they are.
Globalization is one way to describe the unpredictable dynamics of the movements of labor and capital. Another is to look at U.S. county-level private sector job growth. Looking at just the 100 largest counties (by pop. in 2000), thirty-year growth rates ranged from -36% (St. Louis) to just over 400% (Orange county, FL); the median growth was 90% (Fulton county, GA). Is there momentum? Inevitably, yes. Momentum is a default last-resort investment strategy only.