This is quite frustrating. If "industry" really cared about the skill set of their employees, they would train them themselves. Just like they used to do many years ago. My grandfather was an engineer at GE, but he received his degree in English. GE felt it was a priority and an investment to bring his skill set up to snuff to achieve the goals of GE.
So, what industry needs to decide is: WHAT IS IT THAT THEY REALLY WANTED PEOPLE TO BE TRAINED IN? I know there is no single answer for this. But, what does industry want? Do they just want people, trained in a specific discipline, who are proven problem solvers? Do they want people who are well versed in lots of different techniques? Do they want people who have been trained in economies of scale? In recruiting new hires, WHAT ARE THEY LOOKING FOR?
If recruiters and industry leaders could be a little more clear on this problem, I think that most faculty would be willing to help their students become ready for their future career. And, I also think that more graduate students would be willing to push for the right kinds of experience to satisfy what "industry wants". But until there is something more explicit, this is NOT going to happen. If industry is just sitting back and cherry picking new graduates looking for experience in "Solid phase synthesis of lithium-based materials that perform at under 300K with expertise in XPS, AFM, SEM, TEM, etc, etc, etc" what they are showing that they really want is for a larger pool of applicants and a larger pool of graduate students and graduate student projects to pick from. We've already discussed what a bad idea this is.
What is it that industry wants and expects???For the most part, I think the call for better preparation can be answered by the students themselves. As @UnstableIsotope pointed out, they need to expose themselves to more industrial examples by reading the working literature (C&EN, J. Med. Chem., Org. Process. Res. Dev., other relevant industrially oriented journals.) They need to attempt to turn "unknown unknowns" (things they did not know that they did not know) into "known unknowns" (things that they know that they're not exposed enough to, but are aware of the field and its concerns.) Forbearance from the professor for absences from the laboratory to go and do an industrial internship would also be helpful, I would think. I am curious as to how many professors would allow a graduate student to use an industrial internship and the relevant projects for a portion of their dissertation; while I'm sure most wouldn't have issues, I am sure that some would vehemently object.
To respond on Anon0732a, there was a time when General Electric (and its shareholders) would allow for long-term thinking, e.g. beyond a 2-3 year time horizon. It seems that if you're a major corporation now, you don't think much beyond next year's earnings numbers. (I also assume that GE did not have a major research university infrastructure to draw trained engineers from, so they HAD to do it themselves.) I can do Anon0732a one better and show the example of Kettering University, once the General Motors Engineering and Management Institute. (I remember getting a brochure from them when I was applying to college.) It was set up in 1926 to train employees for just GM -- where's the pharma/chemical version of that?
Ultimately, I think our academic interlocutors are right, in that there is some attempt at cost-shifting here. I don't doubt Wuxi Apptec has an in-house training institute. Is it time for collaborative institutes to train the US industrial chemists of the future?