Edit: and Intel doesn't and won't play into the market for support chips. There's no way in hell that they can support Intel's required profit margins.
It's kinda both. To be clear, EUV directly refers to a specific type of lithography equipment, ones using extreme UV light sources. However, those changes come with a lot of changes in larger process flow that are all inter-related.
The change in EUV means smaller lithography lines, and also means needing photo resist, etching, cleaning, etc. - many other process steps, that support that size change. Add in that now, with EUV lithography, you are operating lithography under vacuum and its a huge challenge. Suddenly you have to have vacuum ready materials.
One of Intel's mid 2010 wins was getting immersion litho to work better than peers - similar thing just wet wafers rather than vacuum wafers.
EUV is a tool/technology, one used only on critical layers where that precision matters (e.g., gate patterning). But the larger (entire) manufacturing process has to be designed to enable the use of that tool/technology.
AMD and Intel have been trading the throne for years and every single time everyone says one of them is about to die. Yet here we are.
I'll use Intel's advertised sizes for processes below, but keep in mind that a smaller number just represents a new manufacturing process. Not that actual size.
Intel switched to a FinFET process with its "22nm node" in 2011, then a few years later in 2014 switched to a "14nm node" that is also FinFET. That was effectively the last profitable manufacturing innovation for Intel. A "10nm node" certainly exists, but is practically a black hole of money when you consider the time invested on it and just how far behind they are at this point.
The only concrete thing Intel did to address this was to just rename all their process nodes to be smaller. That's basically it. A few years from now is a full decade that Intel will have spent without actually introducing a new profitable manufacturing process. You can make all the press releases you want about how it'll be profitable "one day", it doesn't change the fact that Intel can't really seem to design new manufacturing processes.
Intel effectively watched the manufacturing technology move past them and shrugged. The proposed solution to this is to spend several more years doing basically the same thing, but this time they'll hope they get it right. This is magical thinking.