IT has occurred to me that if you replace the bearings occasionally on an outrunner , the motor should last basically forever.Is there anything I am missing ? Do magnets lose power? Does the wiring get brittle ? The job of replacing bearings has shown itself to be a minor task , with little cost.
Is there anything that will degrade in the E.S.C.?
This certainly beats I.C. power by a whole bunch .
Keith Varley. 

Some motor failure possibilities:

Bearings, yes.

The magnets are glued in, and can get debonded.  Stories of this abound.

The cylindrical part of the motor bell is made of steel, to provide a back iron for the magnets.  Nearly all the motors that I know of use an aluminum end on the bell.  So the join between aluminum and steel can become unstuck over time, vibration, and temperature cycling.  There are a variety of possible ways to hold these together; I don't know what's used in what motors, or what's best (any machinists, mechanical engineers, or people with broken motors should probably pitch in here).  I have seen a very few (just one or two) stories of this happening (with pictures).  In that (or both of those) cases it was a cheap motor; I think the cap was pressed onto the bell rather than being glued, threaded, or perhaps shrunk on.

Normally, each of the three power wires is soldered onto the armature windings.  Solder joints are subject to vibration failures.  I've seen teardown pictures of cheap vs. good motors: good motors pot that joint in some sort of silicone; cheap motors just let it flap around (and fail).

Wiring probably does get brittle over time, but barring really high heat it'll be decades.  Unless you've got a real el-cheapo, the motor will probably be obsolete before the wiring cracks -- and wiring can be replaced if you've got more time and skill than $$.

AFAIK, there is little, other than way too much heat, that'll demagnetize a rare earth magnet.  When I was starting my career in control systems design, crusty old control engineers would chuckle at my current limits and tell me that if I tried that with an AlNiCo motor I'd degauss the magnets the first time out, and have some serious 'splaining to do to the boss.  With a rare-earth magnet motor you can run enough current to melt the wiring to a puddle, and as long as the magnets never exceed their Curie temperature they'll be fine.

Some ESC failure possibilities:

The input caps are aluminum electrolytic.  Aluminum electrolytic caps have, as a critical component, some electrolytic paste that's made with water, salts, and binders to make it pasty.  When the water dries out, the cap stops working.  These dry out over time, and they dry out a lot faster when they're run hot.  Good ESC manufacturers use good caps, which helps a lot.  There's a cap type called an "Aluminum polymer" coming out that's replacing aluminum electrolytics, but they're spendy: I don't know if even the top-end ESC makers are using them.

When the input caps fail, you'll start getting big voltage spikes on the ESC input; this'll degrade the ESC transistors over time, which will eventually make one or more of them fail short (meaning: the transistor innards essentially melt together into highly conductive slag that shorts drain and source together into a short, expensive piece of wire).

All semiconductor electronic components are subject to damage by static discharge.  Long before you can feel the spark, the spark is damaging the parts.  This should be less of an issue with a properly-designed ESC, but all the anti-static stuff adds weight and cost.

Program memory is stored in flash, which (essentially) charges up little capacitors more or less permanently.  Over time, heat, and radiation the "more" becomes "less", and the memory can get corrupted.  Most flash is guaranteed for ten years, which means that most processors will probably hold memory for many times longer than that -- but a few won't, and it only takes one flipped bit in the wrong spot to render your whole ESC useless.

Pretty much all of the above failures are going to result in the ESC working fine for most of it's life, then one day mysteriously refusing to work or bursting into flame.  Without a boatload of failure analysis (at which I am not good) I couldn't tell you what to expect when -- and even with a boatload of failure analysis I'd still tell you to watch what actually happens if you want to know truth.

Gee whiz Tim, you sure know a lot about the subject.  Thanks.  See you at the Fall Follies?Keith

In a weird sort of way -- my professional specialty is motion control, which is all about motors and control rules and software and circuits.  But I haven't flown a lot of airplanes on electric motors.

So I'm kinda heavy or theory and on practice in the field next door, but if you want to know specifics (like the best way to affix a steel bell to an aluminum front housing) I can only bloviate, not tell you for sure.

Thanks for adding those, John.  I tend to think in terms of the electrical issues that might cause the Dread Finger of Blame to point at me -- most of the problems you mention are manufacturing defects, while cracked traces and whatnot are either normal wear and tear, the result of too-high vibration levels, or boards that are not mechanically stout enough (which causes the DFoB to point to Engineering, but to the mechanical engineer standing next to me, so that's all right).