The correct deployment strategy would be to make a big purchase at the front end of a device's lifecycle and then only replacements from then on out so that you get the most out of every machine. Future capital purchases would be with a new device and termination date.
I think this point is really important, and allow me to go one step further: I work in the public sector of education and purchasing technology is such a complex issue that IT governance has to be involved with decisions like this. That's to say that, without a governing body to review purchases (outside of whoever handles the actual procurement, i.e. funds leaving the bank account), mistakes like this will happen.
We can be upset with planned obsolescence, but there's distinctly a human error here where there wasn't enough research and planning.