Overhead cams make a lot more sense on inline engines, and all of GM's inline engines are OHC designs now, I believe. In the case of DOHC, you only have 2 cams per engine and the drive system for the cams is fairly straightforward. Inline engines are inherently narrow, so the increase in head width isn't a problem. Only the height can become a problem.
It's when you get to V engines that DOHC (or SOHC) comes with some serious consequences. 4 cams instead of two, and the drive system for them gets much longer and complex. Width becomes an issue too.
I'm no expert on the subject, but the statement that OHC designs use fewer parts just seems wrong to me. Even if there are fewer parts, the cost of the parts for a DOHC engine versus an engine with one in-block cam has got to be higher. I mean, sure you've got a lot of pushrods, but how expensive are those? On the other hand, producing 4 cams instead of one must be way more costly.
Could I get other knowledgable GMI folks to weigh in on this? Which design typically uses more parts, and and which is more costly?
The Gen III small-block was a complete redesign back in 1997, so all the castings were new then, and still GM went with pushrods. Similarly, GM has had two different DOHC V6's with their own blocks (including the new 3.6L), so they have invested in the castings. It's not about them wanting to use old or existing castings that keeps them developing pushrod engines.
There are definite cost/value advantages to pushrod V engines that make them appropriate for a huge portion of the North American market. Sure, the industry as a whole goes with DOHC designes now, but that's the easy way out. You don't need to do as much development on a DOHC engine to meet mileage and horsepower targets, but the customer pays with higher build and maintenance costs. For instance, the timing belt in my Probe GT would have cost about $450 to replace, and it should have been done every 60,000 miles. I've never replaced a timing chain on a pushrod engine.