The move, seems to me, is to slow down the momentum ARM is getting. If, indeed, sub 10W is achieved in 2012, using ARM in micro servers will be less appealing.
Price might be another important factor for the popularity of microserver. Unlike large corporation, SMB is looking for saving as much as they can. It's actually one of the value prepositions of cloud service provider to SMB. There is no initiate huge investment. There is only monthly cost. Cloud services provide flexibility and scalarability.
Like the article says, with low power requirements kicking in soon, Intel would do well to introduce Atom quickly. Otherwise, the ARM chips wouldn't take much time to lap up the above quoted market share!
A good try by Intel for the micro servers.Each generation of computers has been replaced with new computers which are smaller, more energy efficient and less expensive. Similarly the servers and now the micro servers. Even though prdicted 10% of market this will raise upto 30%.
In-case of low end servers, there is not much need of the processing, still most of the server proving hosting are running on pentium-iv, so what the article is talking about is the reduction in power consumption, again at low end servers users will not be bothering it as expected by Intel.
Cloud will not be mainstream for quite some time because of security and busniess models needs to get establised.
With just 10% market forecast for micro servers I wonder why Intel is interested in this market apart from the market driven initiative to provide an alternative to ARM in the low power end.