Manufacturers of components of PCs of all kinds are in a tough spot. OEMs have placed orders for parts for tablet PCs and there is not capacity to produce it all: parts for desktops, notebooks, netbooks, tablets, smart-thingies. To expand capacity to adapt to the new demands is risky as total shipments may or may not rise in the market. To be caught short in capacity is risky as customers may be lost. see “Upstream supply chain facing challenge from strong tablet PC orders”
It is the same question faced by system managers who see increased demand for storage and services but who are not certain that past performance will be repeated in the future. Do we add one more server or a bank of them?
No one knows for certain that tablets will be popular indefinitely. Building capacity to produce for them is a risk. It could be that tablets will replace notebooks in many cases because tablets are more portable and cost less to produce, or it could be that tablets will be an accessory and add to the pie.
I would bet that tablets will be part of the diversity of IT for a long time and I would ramp up production if it were not my money being invested. What manufacturers will do will likely be more cautious. They will shift capacity until the trend is clear. That could hold back adoption of tablets a bit and keep prices high for a while. The new kids on the block will have an opportunity to pick up the slack.
No one knows for sure whether M$’s entry in this market will succeed in a few years. They seem to have missed the boat on smart-phones. I don’t see the attractions of that other OS being large enought to command the market. I do see that enough suppliers of software will port software to ARM in expectation of M$ being there that GNU/Linux, already being on ARM, will get some of the action. The cost of porting software is not trivial nor free of risk. Suppliers of software will want to hedge their bets. One way to do that is to write platform-independent stuff (ie. Java) so the cost of porting becomes less. Java applications will likely see lots of action in 2011. Suppliers who have their applications written in Java will be exposed to much less risk as things sort themselves out. A year or two is sufficient time to port many applications to Java as well as to ARM.
Of one thing I am certain. M$ is not in a position to dictate to the market by announcing vapourware on ARM as they used to be a few years ago. We saw with Vista M$’s inability to dictate. Their share of the market has been dropping steadily since even while their take has increased because the market is growing. The best growth will be with ARM and GNU/Linux in the future.