Perhaps the GOP had precognition, and saw all of these rumbles coming this decade. If so, then this is exactly why the GOP packed GA-5 and put it where they did, along with making the other 2 VRA districts just a couple of points blacker than was needed. Compared to below, potentially sacrificing GA-6 would be a bargain.
It's really not that much of a difference (3 points on average), but in conjunction with moving GA-5 to the north just a bit, you potentially get a totally different result in today's climate.
I'm largely using interstate and major highway boundaries wherever possible, since that's just as relevant of a boundary in metro ATL as anything. I started out drawing the 3 VRA metro districts in a reasonable fashion, followed by GA-5 to the north. I left GA-2 & 14 identical and effectively identical, respectively, since Bishop's VRA district is legit and 14's configuration is about as fair as one can get being tucked in the corner of the state. Filled in the blanks from there.
GA-5 becomes a majority-white VAP (55%) district that went 65% Obama (59.5% agg Dem), becoming a fifth reliable Democratic district. Prior to Trump, that'd be where the interest in this map ends for the most part.
The problem is that if this current climate is indicative of trends that'd be present throughout the broader metro, then you'd have
three more vulnerable metro districts.
McCain's 08 margin in GA-6, GA-7 and GA-11 below are smaller (20, 17 and 18 points, respectively) than Romney's 12 margin in the real-life GA-6 and are of similar consistency. Potentially, you now have 7 solidly Democratic and/or vulnerable districts in the metro: every single one of them.
Of course, you also still have Bishop's district on top of that. So we're at 8 non-safe GOP seats.
But it doesn't stop there necessarily. GA-10 is arguably at least as competitive as some of those northern metro districts (McCain 52-46; agg Dem 47%). It includes Augusta, Columbia County (if you have suburbia around ATL rebelling, then you likely have movement here, too), a small portion of Athens and a decent chunk of Barrow's final stomping grounds. It would make for an odd coalition and would require an ideal candidate perhaps, but potentially vulnerable nonetheless.
Then, you have GA-1, which McCain won by 8 (agg. Dem 46%). Not too different in raw numbers from the current GA-1, but considering that the more polarized portions of the real-life GA-1 are cut out + you have a higher than average percentage of suburban moderates in both Chatham and Effingham, and suddenly this district doesn't look quite that safe, either.
Anyway, the point is that one can't just look at the lines in a GOPmander and say it's gerrymandered or not.
Allocate the Dem votes just a tad bit differently with boundaries that make sense/are neat and you suddenly have reduced the GOP to a minority of seats being guaranteed for them - especially in Trump's climate. Losing GA-6 being the worst-case scenario doesn't look like that bad of a deal.