If everything else is done right (bearing mounted correctly, lubricated, no dirt, correct loading, etc.), bearings ultimately fail due to high cycle fatigue at the rolling element tracks on the races. Repeated flexing caused by high Hertzian (contact) stress between ball and race makes tiny divots fracture out of the race surface. Particle count goes up, more divots break out...bearing fails.
The grooves they put in the races of these bearings will act as stress concentrators and accelerate fatigue failure compared to ungrooved bearings. Still, if low friction is more important in the application than long life, this may be a useful new twist on rolling element bearings.
In a greased bearing the problem is that the balls can touch each other (causing sliding friction) and /or pack together on one side of the race causing uneven load or the inner race to shift relative to the outer. The solution is typically to put a cage in so they stay evenly spaced. This cage can be metal or plastic and the balls can rub against it. The grease is mostly used to lubricate the cage from the balls. The rolling resistance of the balls on the race is very low as they aren't sliding across the race, but _are_ creating friction and wear when they touch the cage.
This system seems to work by allowing the bearings that are packed together to touch briefly as they enter the divot causing them to space out afterward. I guess it would perform similarly to an un-greased geared roller bearing?
I'm not sure if this will work well under load. Grease provides lubrication to prevent pressure welding of the bearing to the race and it can also act as a heat sink. Besides those, I would assume the main problem would be fatigue at the site of the divot.
The best part about all of this is if you follow the links to the Coo Space web site and then mouseover the links: their web site is hosted on, of all things, geocities.jp.
I wonder how this works under load. If a vibrating or asymmetrical load is on the shaft (the normal condition for a bearing) then is the small deceleration/acceleration effect still working to separate the balls?
Modern, decent quality ball bearings have a high chrome content (not quite "stainless steel") and don't rust so easily. The primary reason for grease (oil in a few specific applications) is definitely lubricating the ball/race interface.
Water and grease are textbook immiscible things. There isn't water in grease. There may be water in the environment and there usually is in air, though.
Grease is needed to lubricate the balls so they are not in direct contact with the races. Without grease the balls/races will fail rapidly when under load. There is no load on the cages so I fail to see how this development helps at all.
Let me quote a key passage: "Engineers must check that films are sufficiently thick so asperities — peaks of surface roughness between mating surfaces — don't touch. Otherwise, bearings can rapidly wear and fail."
