This article, and many like it, seem to use "startup" and "consumer-facing web startup" interchangeably. If you're doing the latter, you will at least be in better company in the Bay Area. There are certainly companies that benefit from having a presence there, but there's something to be said about being outside a consumer-web echo chamber as well.
In Founders at Work, the one founder who I remember stressing the importance of location the most was Mike Lazaridis, who started his company in a smallish town an hour or two from Toronto. They have a higher market cap than all but a few of the companies Jessica interviewed.
The smallish town was/is Waterloo. The RIM HQ is close enough to the University of Waterloo campus that the two entities have sold buildings to one another. In a speech he gave once I heard him draw the analogy about putting the refinery right next to the well. He also had this interesting bit about how it was more important that the "RIM" logo on the building be more visible from the campus than the street.
Most places have to convince talent to stay whereas the Valley has it come to them, I can see it being an easier battle.
Google wasn't interviewed in Founders at Work, but they stress the importance of location starting with Noogler orientation, with periodic reminders every time you try working on a distributed team.
They stress the importance of location of team members relative to each other, or the importance of being in the valley? From my understanding, Google has a number of engineering sites outside of the Valley, even in the USA.
When you're a startup, that's pretty much one and the same.
The overall principles are that team members should be close to one another, and that you should focus very intensively on making sure you have the right team. In cities with a lower density of technical people than the valley, it's very difficult to have both a critical mass of highly-skilled people and have those people all colocated together. You occasional see successful startups like 37signals or Flickr that have smart people, but do so by having them work remotely. And you see lots of startups that attempt to get lots of local people, but sacrifice team quality as a result. You don't tend to hear much from those. But it's when you combine smart people with critical mass that you get something like Google.
They do, but I think they still keep the individual teams close to one another. I know that the team here in Pittsburgh works on only certain products, at least.
You must be thinking of RIM in Kitchener/Waterloo. When I was at University of Toronto in the early nineties, U of Waterloo was the place to go for CS or Systems Engineering if you wanted to get hired by Microsoft.
Sooo ... it's an interesting counter-example in that it kind of supports the general argument that it's good to go where the smart and savvy people are, even if that's not just SV.
In Founders at Work, the one founder who I remember stressing the importance of location the most was Mike Lazaridis, who started his company in a smallish town an hour or two from Toronto. They have a higher market cap than all but a few of the companies Jessica interviewed.