Just get one now. The price/performance ratio of the Samsung 850 EVO is hard to beat.
The thing you have to understand about SSDs (and HDDs for that matter) is that they're measured in MB/s. Unfortunately MB/s is the inverse of how we perceive storage speed - as wait time. This means the bigger MB/s gets, the less it matters. If you need to read 1 GB of sequential data:
125 MB/s HDD = 8 sec
250 MB/s SATA 2 SSD = 4 sec
500 MB/s SATA 3 SSD = 2 sec
1 GB/s PCIe SSD = 1 sec
See how each doubling of MB/s results in only half the reduction in wait time of the previous step? The bigger MB/s gets, the less it matters in how we perceive storage speed. Coming from a HDD, a SATA 3 SSD will give you a 6 sec reduction in wait time. If you wait a couple years for a NVMe PCIe SSD, you'll only see a 7 sec reduction in wait time. Or put another way, a SATA 3 SSD you get today will give you 86% of the speedup compared to if you wait another year or two for a NVMe PCIe SSD. Not (500-125)/(1000-125) = .428 = 43% like you'd expect if you erroneously calculate with MB/s.
On top of this, most read/writes to a drive are not sequential (unless you do something like real-time video editing). Most of your read/writes are going to be better reflected by the 512k and 4k speeds of the drive. For SSDs those are currently around 200-300 MB/s and 30-70 MB/s respectively. In other words, even SATA 2 is good enough. Real-world comparisons of SATA 2 vs SATA 3 bear this out - there's very little to be gained by moving to SATA 3, never mind PCIe.
http://www.tomshardware.com/reviews/sata-6gbps-performance-sata-3gbps,3110-7.html
SSD sequential benchmarks are rapidly becoming like memory benchmarks, which show dramatic 30%-50% improvements with each generation. But which only translate into 1%-2% improvements in real-world tasks. Unless you do nothing with your computer but run benchmarks (edit: or rrun tasks heavily dependent on sequential speeds), you're better off ignoring the sequential benchmark speeds.