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.