We would need 2x the hashpower right now to double difficulty again, which would be at this point an insane amount of new deployments. I see this being very unlikely as mining profit is already highly unfavorable, even for large farms that also need their own custom high-energy data centers to build and maintain around them. I think we'll see the network start evening out, if not now then soon, which makes profit projections less volatile with a stable-ish difficulty.
The network will still continue to grow, just not at nearly the pace it has been as these thresholds are reached.
Things are definitely slowing down but there's still a very good chance that we'll be at almost 5x the current difficulty by the end of the year. That can be funded from just the current levels of mining rewards (running at about $1.75M per day right now), BTC price, retail hardware costs and allowing for moderate energy costs and we'll hit 2x by the end of August. If the hardware vendors drop the retail prices or the BTC price increases then > 5x by the end of the year is very likely (and this will likely happen because they always need to price things to make it appear possible to achieve an ROI).
Data centre builders won't be paying anything like retail hardware prices and can move to locations with low energy costs so they have even more scope to increase the hash rate; if they behave rationally and believe that others are doing the same then they'll actually have an incentive to bring as much capacity online as fast as they can in order to gain a short-term advantage; this is a fascinating Prisoner's Dilemma problem.
