There's no way to square the ETCS circle now electrification is so catastrophically behind any sort of schedule, but signalling immunisation on an existing installation isn't all wasted money, the same cannot be said for DMU stock which may well be withdrawn by 2040.
Much will surely depend if there's sufficient AC immune equipment compatible with the existing installation available - for example signalling relays? Not an insurmountable problem but it all starts pushing the costs north if the project has to modify existing equipment housings simply to provide mounting for new relay bases rather than exchange the existing for the AC immune variety.
But a rather large assumption is that the condition of the relevant parts of the existing signalling assets are fit and able to withstand physical alteration. It is often the case with installations already way past their use by date that they'll just about cope with their daily functionality given appropriate TLC by the knowledgeable local techs but go near it with a not so clued up contract installation team and there's a real risk of bits of it unintentionally dropping off - e.g. wire insulation, creating all manner of potentially dangerous outcomes. In practice of course the project will be aware of such risks long before their installation teams go near and therefore might be a risk which can be adequately managed by the, often at extortionate additional cost proportionate to the immunisation work required, for example wholesale short term renewal of the worst state equipment housing with immunised spec kit compatible with the existing signalling system, but would then be done again when resignalled with 'digital' systems a couple of years later delivered by a different supplier. It all boils down to scale, costs, time and risk to operational reliability while working around old (fragile) equipment to determine if immunisation in lieu of resignalling is a deliverable strategy in the specific circumstances rather than finding out belatedly through bitter experience.
Yes, GWML electrification project managed too often to locate the buried signalling cable route with their size 10 pile driver but buried route isn't commonplace everywhere and one might hope if the patch concerned has known buried route the project finds the damn thing first with a comprehensive site based cable routing survey using all available techniques before leashing the piling team loose - lessons learned and all that?
I grant you though. The integration of signalling equipment into the cab environment while retaining large parts of the lineside and operating centre system architecture makes migration strategy that more complex, particular where one or more sides of the equation come with time / condition constraints forcing the agenda. Oh what joy for those grappling with it all.
