The industry is now developing 100G per lane wavelength optics. For a good update consider the all day Ethernet Alliance and OIDA Workshop on this topic in June. The industry is also defining 50G per lane electrical technology in the OIF. 100G per electrical lane is some years off, but expect people to start looking at it in 3 to 5 years.
So that's a total of 7 general purpose 10G form factors, counting the similar X2 and XPAK separately.
There was also a high density SFF-8470 InfiniBand form factor which supported 2.5G lanes, which could have been used for XAUI 10G interfaces but was not.
The general purpose 100G form factors in chronological and density order are:
CFP, CFP2, CPAK, CFP4, QSFP28
That's a total of 5 general purpose 100G form factors, counting the similar CFP2 and CPAK separately.
There is also a high density CXP InfiniBand form factor, which is restricted to DAC and MMF interfaces, for both multi-channel 10G and single-channel 100G. If CXP is included in the general purpose form factor counts, the totals are 8 for 10G and 6 for 100G. So today there are more 10G form factors than 100G form factors.
Because SFP+ is so dominant, we tend to forget that it is not the first 10G form factor. The first 10G module (300-pin MSA) was introduced in the late 90s. By today's standards, it is incredibly clunky because it's ~20x the size of the SFP+. However, when introduced it was a breakthrough product because it replaced full size cards.
100G skipped 300-pin MSA as first form factor. We should also not expect QSFP28 to be the end of the road for 100G. A possible next form factor is SFP100, utlizing future 50G or 100G per lane technoogy. This is normal evolution. As optics and IC integration improves, the size shrinks and density increases.
The Other Tesla David Blaza5 comments I find myself going to Kickstarter and Indiegogo on a regular basis these days because they have become real innovation marketplaces. As far as I'm concerned, this is where a lot of cool ...