That rate of rising (2x in one month) is rough but really isn't too insane. In the summer of 2011 (May-June) Bitcoin went on a massive price surge and over the span of 45 days the difficulty climbed by a factor of 15x from 92k to 1.4M.
You monitor this a lot more than I do madpacket, I'd be interested on your thoughts. I'm still rather bullish on mining vs buying and hold since barring a fundamental technology change (IE ASICs) I don't see 25TH/s month being sustainable. Unlike with BTC where 90nm ASICs were replaced with 65nm, etc, etc and increases in hashrate could continue long term, there's nothing to suggest that will happen with Eth since new chip generations come on 1-2 year cycles and won't bring massive efficiency increases anyway. Right now adding 25TH/s in a month makes sense as there's an economic case to do so since 25MH/s might bring in $120/month with static difficulty and $0.10/kWh power, but once difficulty rises a couple more times and the same hashrate is bringing in <$25/month that level of investment is bound to go down given the much longer equipment payoff and uncertainty regarding the switch to PoS.
AFAIK Ether's DAG (hashing algorithm) is ASIC resistant so we won't really have the same problems as BTC. However, we may arguably have worse problems as miners. The planned difficulty bomb (graphically represented well by the link in my last post) shows clearly where things are headed. Looking at the future difficulty of DAG benchmarks it appears video cards with TLB issues will suffer the most (Polaris) with drastically reduced speeds unless developers can figure a way to stop memory thrashing. So it appears 3GB 1060's have 3 - 4 months of mining left before the size of the DAG becomes an issue, however interestingly enough 3GB 1060's don't appear to suffer from TLB issues so the hashrate for them should at least stay consistent.
So for mining Ether 6G 1060's, or 1070's make the most sense ATM if you plan on mining longer term. All that being said, with the difficulty bomb it appears we may only have a few months left of mining before hitting a break even point, at least with average hydro North American costs (0.12 KW). If the price of Ether rises to match the chill period, or the difficulty bomb is diffused in Metropolis (next version of Ethereum) which is tentatively planned for September my confidence in long-term Ether mining will go up. Sure it may be possible to keep mining Ether until they realistically roll out PoS and depending on how they deal with the difficulty bomb, we may only be generating small fractions of coins. This isn't a problem if the price moves accordingly so I don't want to be too negative but it's something to keep in mind. Sure there are other coins to mine if Ether mining becomes unprofitable but see the next paragraph for my concerns.
I'm with you on mining Ether and holding for the time being. Buying Ether outright is also a good option if you look at the fundamentals of the platform and can ignore the fact coins were $15.00 (extremely undervalued) not that long ago. I don't think many people mining alt-coins like ZEC / Sia only to trade them for Ether realize the tax implications. Every time you sell one crypto for another it's apparently considered a taxable event, especially worrisome if you trade on places like Coinbase which have supposedly been subpoenaed by the IRS to give up customer info dating back a few years. Other exchanges with more crypto pairs will get hit, and likely other countries will likely follow suit. Big money casts a bigger shadow, a shadow where the tax man is keeping watch.
A similar problem with ICO's. So many people throwing money at them like the early dot com days without any real product or application to warrant the amount of money being collected. A handful of developers without a solid business plan or product do not need 100+ million dollars. Even if you invest and make good returns you'll be subject to short term gains if cashing out (which many people do with these ICO's, lot's of pump and dump) or trading for Ether. That being said there's a lot of good ideas behind many of the ICO's, but the way in how they're being launched, and how much money is invested in really concerns me. I think it's just a matter of time before one of these 100 million dollar ICO's has a DAO like event which could really hurt the price of the Ether, at least in the short term.
Aside all this ICO nonsense and tax negativity, I'm actually pretty bullish on the price of Ether moving forward. If we have a smooth release with Metropolis the platform will only strengthen and allow for more diverse DAPPS and better security and scalability. Ethereum's already won the mindshare of many talented developers in the blockchain space. The company Consneys is around 300 strong and growing fast with around half of them working with Fortune 500's to integrate/augment or replace traditional applications using Ethereum. AFAIK there's no other blockchain technology or group of developers that's anywhere near this level of maturity. Private or public it doesn't really matter, it shows tremendous confidence in the platform. If we eventually get working Proof of Stake and Sharding with Casper (probably at least a year away realistically) the sky really is the limit as that will fix one of the biggest issues (scalability).
Long term I think many of us here who started mining over a year ago planned on staking our coins once Proof of Stake becomes a reality. I still plan on mining, holding, and eventually staking a good portion of my stack. Jumping back a bit, the nice thing about the difficulty bomb and planned PoS is we're nearing the practical limit of how much Ether will (ever?) be generated. I think we're nearing 93 million coins and at around 100 million things will supposedly grind to a halt. The argument of BTC maximalists held against Ethereum WRT inflation or no cap should no longer hold much weight.
Ok enough rambling, happy mining!