I think the use of motion is problematic for clients. Clients want video but don't know where or how to use it. There are two issues: 1. the viewing source/outlest; and 2. Content.
I've looked at the metrics and what we've seen is that click-throughs on videos do not translate into sales. Front page of the NYT, big ad, motion, whole nine yards. No one watches the whole thing. No one stays at the site and shops online or visits a store, less so than a compelling still in the same location. The reason is, partially, the viewing source and the time investment. people are more likely to watch a clip and follow through if its on a TV, or a large display, when they are at home. Few stores in the US are equipped to have video displays large enough and numerous enough. Asia is a different story, but in the states, well, Unless you are shooting commercials motion is essentially a vanity project for commercial clients. The iPad and digital editions may change this because its less browing and more directed viewing. But we aren't there yet.
As to content, video is viewed but passed over before they end, partially because the content is bland. The holy grail is a viral video but clients are afraid of viral content. Only rarely do clients accept ideas that are brilliant, and even rarer still that those ideas are executed well enough for the briliance to shine.
The view from the agencies is that motion is not a gimmick, but its importance is not there yet. This is a growing realization. We will see a flood of video with html 5, which clients don't want to pay for, extorted out of everyone with a 5D2 or 7D. The content will be as bland as the stills they produce, creating glut, similar to the use of stock by horridly designed web sites ranging from Yale University to Chico's.
All this to say I don't think Phase should go after Video unless they believe that they can OWN it, beating Red and Canon and Sony and Arri. There will be too few people willing to buy it over the more established options.
I agree with your reflections on that matter. It is true that the demand is growing, but nobody is really sure about what to do with it.
Viral is a big cake for advertisers and my bet is that html 5 will provide the hability to interact. I think that the interaction will be the clew and I won't be surprise to see the marquet parts eating alive the traditional tv ads.
For example, you shoot a video of a car and you ask the viewer for active participation, so you can choose to drive the car in the city, then stop the movie and choose the colors, back to drive in the moutains etc...
The visual experience will be totally boosted. In my example, stills are used for showing colors.
But that will be heavy productions.
As you pointed, what will profileferate will be those unpaid extra videos (and they are proliferating right now here!...) with whatever.
When I say they are proliferating, I'm refering to some serious ediotorial business that is starting to use video as a "obliged gift". Sign of the times.
So I think will see 2 tendencies: the unpaid mess, and some truly big stuff very elaborate.
edit: and agree with Feppe, we start to see screens more and more. Advertising with big screens in shops so the summer collection shooted with the sound, the models and the wind and all that is much more impacting. Then, NO LIMITS in the claims, you can ad big texts, etc...
And not only fashion, arquitecture projects, the list is unlimited.