[wenc]

The trouble though is that UT-Dallas is still not a spectacular school, so the talent does not rank among the best. It's serviceable for what TI needs, but not the best. And few people will relocate to Dallas just to go UT-Dallas. TI may have gotten by with that level of talent in the past, but if you're a company that is pushing the boundaries, that level of talent may not be sufficient.

If you locate near an already great university in an already attractive city (e.g. UT-Austin), you can tap into a talent magnet that already draws people from around the world. Remember, the talent pool is now global and tech companies that aim from anything short of that would be depriving themselves.

Talent pool size is a real problem. To cite an example: for years, many French-speaking universities in Montreal had a hard time drawing global talent due to the language of instruction; on that basis alone they were less competitive than English-speaking universities in the same city (e.g. global English-speaking universities like McGill draw a from a larger and much more competitive pool of students from around the world) simply because their pool of students are constrained to French speakers. Just as an example: most Asian students will probably not bother to learn French, but many are already learning English in grade school. This means French-only universities lose out on the two biggest pools of engineering graduate students: those from India and those from China. To make things worse, French students from France rarely want to study in Quebec because most of them believe their own grande-ecoles to be better (there is also an element of French snobbery involved when it comes to Quebec).

Anyway, my sense is by setting up shop in a city that already draws global talent, the premium you pay in salaries over the long run is going to be far less than setting up your own university (considering how expensive it is to run a really good university). The payoff is much greater. University reputations are expensive and slow to build, and as can be seen in the UT-Dallas case, one is uncertain to succeed even with sizable investments.

p.s. of course, this argument breaks down if your company does not require a very high level talent.