The right questions to be asking aren’t “why does Silicon Valley create so few jobs;” it’s “why doesn’t everyone move to the Bay Area” (the rent is too damn high) or “how come there’s only one high-tech cluster.” After all, if industrial age capitalism had just created the prosperity of the Detroit area in its heyday, we’d look on it as a huge bust. But we had lots of industrial production clusters, of which the Detroit automobile industry was just the most famous.
I think there’s a standard geographical reason why capital intensive production of material goods exhibits polycentrism: the cost of transport matters. Many of the early industrial nuclei were located relatively close to the inputs for manufacturing. Additionally, once the goods were produced they had to be distributed as cheaply as possible, so location was another essential fixed parameter. Big eastern industrial centers loom large in the public imagination, but the same logic applies in other regions of the nation. Cheap electricity and abundant clean water is why many tech-oriented manufacturers are based out of the Pacific Northwest. It isn’t as if you could just relocate the Columbia river.
The same constraints don’t apply to virtual products and services. These rely on one primary input: humans. Not only that, it seems possible that the returns on density of humans may exhibit increasing marginal returns, at least for a period. One can imagine that when you put a lot of bright minds in the same town you have a “bidding war” in terms of audacity and innovative spirit. Not only that, but a free exchange of ideas in a hyper-charged intellectual milieu transforms creativity from a purely individual activity to a communal one, albeit one which balances the wisdom of smart crowds with individual competitiveness and excellence. One can attain this sort of social critical mass online, but the reality is that there are things you miss currently through pure virtual interaction. Nerds are still human beings, and there’s a lot that happens in the “meat space” that you don’t want to miss out on. So the incentive is that if possible you want to get as close to the heart of the “action” as possible. Geography still matters on the margin for individuals.
Once this process starts in a mind-hub positive feedback loops and winner-take-all dynamics drive the system so that polycentrism shifts toward one-hub-to-rule-them-all. That’s one reason that Route 128 has been overshadowed by Silicon Valley. Though Silicon Valley has some advantages by virtue of geography (the weather is better), I suspect its current advantageous position is due more to chance and contingency. Once it achieved a small heart start in terms of the “place to be” if you wanted to be a tech entrepreneur, then the unstable equilibrium of parity collapsed.
This sort of dynamic generalizes to mind-hubs overall. London has the City and New York has Wall Street. Unless non-human inputs start to loom larger in financial services (e.g., the computers gain significant advantages from cheaper and more reliable power sources on the tiniest margin and it is critical that the humans interface as close to the metal as possible) I assume that their commanding position in the national economy will remain as is. Of course course my caveat about the national level is critical. Despite a globalized world because of the fractionization along cultural, linguistic, and national lines, you will have a polycentrism on an international scale. For now.