That may depend on the field. My experience (in bioinformatics method development) is that people in the industry can't afford to work on state-of-the-art problems. But once a problem has become established and it's important enough to be worth their time, they will eventually come up with a better solution due to their superior resources.