Not disagreeing at all with the specifics wrt publishing. I have some limited+distant experience as a supplier to the CTP side of print+publishing, where being best and fastest had obvious advantages, and I assume that technology is equally important in some other parts of publishing.
I also know many areas of business where there is no need for the latest greatest desktop company-wide, where the time+money spent on repeated company wide rollouts is largely money wasted. It helps the IT Director think he's important because his budget is ever growing, but actually his unsupervised spending can sometimes be a drag on the business at large. I've seen plenty of those.
The Vista disaster, and the number of companies consequently still running XP, has helped many companies understand that they do not necessarily have to keep up to date, and that there are actually risks as well as inevitable costs and potential benefits in the routine upgrade rollout philosophy.
To make money you have to invest in the infrastructure.
Indeed so. But whether the upgrade cycle is an investment or a waste of money is a company-specific decision that needs to be actively considered not just passively accepted "because everyone else does it and Dell offered us a good deal"
the vast majority of companies who suffer a serious failure of their computer systems go out of business within a year
True, but a different discussion to this. It indicates that planning is essential; it does not support the case for routine unnecessary upgrades.