I couldn't get over the suspicion that in 1 year or 3 years, APS-C finally may be phased out. The manufacturers are swearing that isn't the case. Perhaps they are right, but I would imagine that the advanced amateur or pro who currently buys a 50D will eventually expect full-frame in the X0D series. The APS-C will probably be relegated to the ghetto of entry-level DSLRs.
You're overlooking one central and critical piece of things: the cost of silicon imaging sensor rises
exponentially with area.
Image sensors don't follow Moore's Law. All other silicon applications, such as computers, can double performance every 18 months for the same price level. Or they can halve the price every 18 months for the same performance level. The reason this is possible is that transistors are shrinking.
Image sensors don't get any cost benefit from shrinking, because they cannot shrink. The area of the sensor has to remain the same. Although commercial full frame image sensors have been getting cheaper for the last 7 years, it's happening at such a slow pace that it will be much more than 3 years before they even come close to the cost that APS-C is now.
The 5D2 has a $1,500 premium over the 50D, despite the fact that the 50D is more advanced in several ways (gapless microlenses, autofocus, etc.). This is due to the large sensor. The 5D1 had a similar price premium over earlier APS-C cameras. This price premium has shrunk very little in the past, and I see no reason for it to accelerate in the future.
I think that the 50D has proven that increased resolutions and better low-light performance on the smaller sensor will become increasingly more difficult to achieve if at all possible.
Only the flawed reviews (e.g. DPR) show the 50D has more noise than the 40D or doesn't have the expected resolution increase. This myth has been debunked a dozen times over on LL.
Is the APS-C going to become a relic in the near future?
My sources say no.
Will it be relegated to entry-level DSLRs?
Outlook not so good.
Considering these issues, are EF-S lenses a good investment?
EF-S is a good investment compared to Mortgage Backed Securities, certainly. But I suggest bonds for now, while considering international midcap.
Seriously, APS-C is not going anywhere in the next 10 years. The highest volume imaging sensor market is mobile phones, then P&S. But those are high-volume with very low margins. DSLR has higher margins, and over 90% of DSLR sales have been and will continue to be APS-C. Manufacturers will certainly continue to pour R&D into it.
Manufacturers will not be able to let go of the megapixel war. More pixels on APS-C is a dead-end due to lens limitations.
I disagree. The picture you get out of any lens will be related to the money you put in the glass. That's true for all DSLR sensor sizes. Full frame lenses have gotten so much more investment in the last 40 years that they are far ahead of APS-C lenses, so there are many inexpensive, high-quality lenses, especially when it comes to wide primes. This is beginning to change with things like Nikon's 35mm f/1.8.
Lens development will accelerate even further when the crippling optical viewfinders can be dropped for EVF, allowing APS-C to have wide primes with designs similar to FF35, thanks to unlimited flange focal distance (well, save a few mm for the filter stack, and no one wants too much cos^4 falloff anyway).
Innovating at the silicon level is the cheapest path to differentiation.
Innovating by increasing sensor size is hardly the "cheapest path", given that sensor cost goes up exponentially with the area of the sensor.
Innovations generally come in at the top of the product line, pushing older technology down market (FF for cameras or heated leather seats in cars).
I too think generally that's how it will go, despite the many exceptions.
4. As someone else has said, size and weight advantages will always be a reason for some serious photographers to choose APS-C (or FourThirds) gear, even those who also use 35mm format or medium format gear for other purposes. Our common interest in hiking photography is one example!
While size and weight of lenses tends to correspond with sensor size in today's products, there's no reason that it has to in the future. For example, lenses for medium format cameras aren't that much larger than 35mm. In fact, many of them are smaller. The reason is that the aperture is the same or smaller (longer focal length but much narrower f-number).
Someone that desires wide aperture lenses for thin DOF and light gathering power will have to carry around heavy lenses, of course, but everyone else may use narrow f-number. They will simply use slower shutter speeds or less ND filtration to get the benefit of the larger sensor without the weight of the larger lens. This is the same way that medium format is used now.
EVF and other advancements will lead to much smaller camera bodies in the future, which takes care of the non-lens part of the equation. The larger sensor will require more electronics for the same shooting speed, but reducing the frames-per-second will allow it to have a smaller ADC, image processing block, cooling requirements, etc. I think the Sigma compact APS-C is illustrative here.
The missing piece, of course, is large inexpensive sensors, which I don't think will happen for a long time. I do expect it to happen sooner for APS-C.
APS-C will continue to have the highest value and dominate sales of all DSLR formats for many years to come.