My predictions (worth exactly what you paid for them):
Non-business, non-gaming users will have largely moved off of PCs to mobile devices, continuing the trend that is already well underway. PC market will split, with mainstream business-oriented desktops being the majority of sales and workstation/enthusiast systems becoming a small (and shrinking) niche market.
Triple or quad channel DDR4 will be the norm for all but low-end systems. Several DRAM chip makers will have filed for bankruptcy and/or been acquired by larger rivals, as a result of guessing badly on the timing of the transition from DDR3 to DDR4.
Intel's upward march of CPU performance will continue, but at a slower pace due to lack of competition.
AMD will still have a small presence in the server (and possibly the low-end desktop) market, but AMD-based enthusiast systems -- if they exist at all -- will be limited to builds based on re-purposed server platforms, and will be cost-effective only for highly parallel workloads, due to continued mediocre single-thread performance. They will have sold off the GPU division to raise cash (licensing back the tech needed for their APUs), in a last ditch attempt to save their CPU business. Their foray into ARM (and the Seamicro acquisition) will putter along without ever quite achieving critical mass, doing well enough to ensure their continued existence, but never quite pushing them over the line into sustainable profitability.
GPU performance will continue its upward trend, bringing us ever closer (but still not quite reaching) the Holy Grail of Hollywood-quality effects in real-time. Most PC games won't take advantage of the full GPU capabilities anyway, as many will continue to be ports of console titles. High-performance scientific computing will be almost entirely GPU-compute based.
SSDs will dominate the storage market. Mechanical hard drives will still be around, but used primarily in business and data center applications. Home use of mechanical drives will be mostly limited to NAS and media center type applications (which will experience a surge in popularity when someone finally finds the critical hole in Blu-Ray DRM and facilitates easy ripping of Blu-Ray titles).
The mechanical keyboard craze will have plateaued, with the current bumper crop of "me too" vendors of Cherry-based 'boards whittled back to a handful. Popularity of mechanicals will remain above where it was 5 years ago, but there won't be as many choices as there are now.
The discrete sound card market will effectively continue to be stuck back in 2002, with fewer and fewer enthusiasts opting to upgrade from their motherboard on-board audio. Creative, no longer able to make a profit from hardware sales, will resort to a pure patent trolling business model. Asus will opt to exit the discrete sound card market rather than litigate.
VIA will continue to produce SFF hardware that almost nobody cares about, and Matrox will have unsuccessfully tried (yet again) to break back into the mainstream market with a graphics-related peripheral along the lines of their TripleHead 2 Go
device from a few years back.
The years just pass like trains. I wave, but they don't slow down.
-- Steven Wilson