>> But the US missed to invest in its infrastructure
> By what metric?
I'm not a civil engineer, and I don't know if anybody can tell reliable numbers at all for that issue. But the US infrastructure hasn't received proper maintenance in the last decades if one can trust certain studies. And this is getting serious for some areas in the next couple of years.
http://edition.cnn.com/2010/US/06/22/map.infrastructure/inde...
Looking at the downward trend for most cathegorized domains since 2001 is not a good sign for US infrastructure. One may say that everybody involved in that business has his own interests (government(s), construction businesses, consultancies i.e.) and is trying to defend them or make profit out of certain decisions. But stretching the infrastructure to the upper limit of the designed lifetime without proper maintenance until then just accumulates infrastructure investments, mostly towards a higher figure. (It's always difficult to find reliable sources regarding that issue, with a lot of lobbying going on in that area. In that case I simply trust the American Society of Civil Engineers for judging about that issue because my own lack of knowledge and expertise.)
> That's not how these kinds of things usually work. Usually the longer you can put off an upgrade, the cheaper it is to maintain a system when you amortize it. And I can't see why this would be any different in the grid (or roads, or whatever else).
It's not only about the equipment working until or even over the designed lifespan and/or just replacing them, but the aging grid is characterized by a couple of points. There's an interesting book called 'Aging power delivery infrastructures' from Willis et al. (2001) about it: http://books.google.com/books?id=1GcxSDpvdzYC
Basically in that book the authors point out 5 main factors: 1. Old equipment, 2. Obsolete system layouts, 3. Old engineering methods, 4. Uncoordinated and non-optimal use of distribution, 5. Old cultures and ideas.
I think there hasn't been any major improvements since the release of the book, at least no I'm aware of. And this is definitely not an US-only issue. But I think there's definitely a need of rethinking and reengineering the power grid.
It will be a big effort for western industrialized nations to bring the grid, which is mostly pre-70's, into the new century. And just by pushing the lifetime of this old and inefficient grid doesn't make it cheaper (higher failure rates, inefficient equipment, blackouts, i.e.) in the future.