If we look at this as objectively as possible:
PC notebook and desktop will be slow growth market, but x86 is entrenched here and would be the last segment to fall to ARM. There will be pockets of losses such as Apple (that 'lifestyle' company) making 'ok' processors for their own machines. Assume glass half full and presume Intel regains process parity, if not leadership, and is able to stop share losses to AMD. Still, it isn't enough, so go to 2)
In Server space glass half full says that Intel still has competitive x86 offering. AMD is taking share like crazy. But will any of this matter? The writing is on the wall, ARM trajectory in cloud and servers is very very strong. It fundamentally is better mips/watt. So, this could crater Intel sooner than losses in client. Shockingly, in the last earnings announced Sever team was barely break even. Shareholders were aghast.
Manufacturing. Much has been written about this. To summarize, Intel simply cannot fill it's own fabs or afford the cost of each new generation. It must either divest fabs or compete with one of the best companies on the planet -- TSMC. There are many core competencies that Intel lacks -- tools and library ecosystem, low cost manufacturing, ability to run high sku mix at low volume, ability to support many legacy nodes. I am at a complete loss to understand how the BOD ever signed up to such a 'moonshot'. Anybody who understand the skills required to beat TSMC would never have signed up for this kind of wasteful capital investment. The argument that Intel must do this for National security interests is a pig in a poke. Trying to beat TSMC has (and will) destroy shareholder value.
For these reasons, we can easily predict that Intel is on a decline that may take as few as three years and as many as ten years. Intel may be able to extend its life a bit if it gives up on the 'moonshot'. Either way, it won't be able to survive too long