Okay. I am just going to quote both of these as it's a bit of a contradiction and say nothing else as it isn't required. (Graphics Core Next is modular.)
It really isn't and GCN being modular isn't relevant at hand ... (just about every graphics vendor have separate display/video engine units but all of them need time to integrate these hardware features)
AMD hasn't definitively beaten Intel in 15 years in the CPU space until just now. They survived.
I don't know if I'd say AMD is 'beating' Intel when they still have some gaming performance advantages but at the very least it could be considered a 'stalemate' ...
AMD needs to break the stalemate and improve their gaming performance fast before Intel reacts (they're good at this) but maybe just maybe they'll finally go down and quietly at that ...
The VAST MAJORITY of current laptop display resolutions today are overwhelmingly sub-1080p! (heck, it's not all that rare to see 768p/900p as well) The fact that Navi can match a Radeon VII using HBM2 memory in 1080p graphics performance while using ~25% less power is a substantial improvement and that especially counts when most systems won't be able to supply the GPU with tons of power like we see on desktop ... (Navi still uses less die area than the Radeon VII coming in at a ~25% smaller die size)
Delivering a Zen 2/Vega combination just to rush out a product is downright sub-optimal. 1080p graphics performance in the portable space will matter for some years to come down the road since reviewers hold lower standards for them in terms of resolution capability in comparison to desktop systems ... (being more efficient at lower resolution is especially important since most portable systems won't have 100s of watts of cooling power to play with)
As for 10nm being used for desktop CPUs, Intel's current roadmap might say otherwise since they don't plan on having sub-14nm desktop CPUs at least until the end of 2021 ... (definitely not expecting sub-14nm in the near future)
By the end of 2020, I expect either a make it or break it revelation from Intel. Intel are either going to have a breakthrough (make it) or finally call it quits and then go fabless (break it) because they can't keep going in like this when their competitor is going to be eyeing 5nm well before they even get to 7nm ... (Intel can't afford to have sagging revenues due to a fierce competitor when logic fabrication R&D is more capitally expensive than ever)