The more global and powerful the vendor, the better the view of the future…
Err… no. If you want a view of the future, you find the emergence events and this is post-emergence. Big company events like this are only held once they can safely say what everyone already knows.
As to Intel involvement, remember Intel was a board member of the Web3DC (X3D, VRML, Collada, etc.). Rankled by the participation agreements that prevented pushing encumbered-IP into the open standard, they resigned and worked their own consortium (see 3DIF).
Note something important here: if competition is important to this industry’s legitimacy, then the Web3DC people still have the only open competitive standards and technologies. For X3D/VRML, there are multiple clients and server systems available with a push toward virtual world standards for interoperating systems such as H-anim and the Network Sensor. While Koster, Intel, and Rosedale talk this stuff, the Web3DC is actually doing it.
Back to Intel:
The aim is of course, to sell hardware, and any good graphics professional knows the score: more polygons moving at the speed of light need more hardware. For that reason, I'm glad to see Intel parroting the rest of the industry about their discovery of the 3D Web. (Jeez... you just can't make this stuff up.)
The difference in 3D at AOL or the other failed attempts is that AOL had one bit right: a massive MU is a server-farm technology more than a graphics technology. Real-time on the web is not a real thing. Latency == Web. Real time MU is mostly a smoke and mirrors trick but it works.
The need for interoperable web 3D is quite real in the market. The reason is not thrills: it is thrills maintenance. 3D content is high cost content unless you pick your technology and the standards well. I do this stuff for fun but I do know how to do it cheaply until I have to host on an expensive server farm. If I have to throw it away for every client or client version, I can't afford it and neither can even deep pocketed companies. We learned our lessons from the early monolithic WYSIWYG systems for documents: translations, rehosting and scrubbing timely information is an expensive nightmare. Because X3D/VRML really is a competitive market, I may target one browser such as BS Contact, but then I adjust the content to make it playable on competitors even if in a downlevel mode. I learn how to make it more interoperable and the vendors flush out some bugs. Everyone wins. That is how a competitive ecosystem with more symbiotes than parasites works.
Fortunately, the beginning steps of VRML ensured that despite its highs, lows and outright cat fights, the community culture is a real survivor and healthy because it remains open and cooperative. So almost a decade and a half later, the content is still working and has improved substantially. That'd pretty good for amateurs. :-)
Where should Intel be going? Away from the data center concepts and into the peer-to-peer standards, or at least into local hosting. The reasons for businesses are easy as I've said before: don’t do private business in public phone booths.