That's not exactly conclusive because VRAM reporting tools report the "allocated" VRAM, not the VRAM in use, they didn't use a 2080 8GB and 11GB for an apples to apples comparison since we know from the Vega II tests that nVidia and AMD cards utilize VRAM differently, and none of the cards they used were equal in power since the 2080 Ti is -far- faster than Vega 64 so conclusions really can't be drawn from it, especially as BF V at 4K full details only the 2080 Ti is able to put out playable frame rates.
Also since VRAM measurements were taken with the 2080Ti, it is quite possible the game did not shift unneeded resources into RAM (which was the real point of the article) and doesn't represent actual VRAM usage.
