There are always reasons why you shouldn't do something... whatif's can be more of a hinderence than a help in the wrong hands...
---
>> Any server on your network is open to attack by a malicious individual - and not all of these individuals are on the outside of your company walls.
>> Any traffic over the network will either reduce performance of the network (increase latency) or will increase your network infrastructure costs to support the required bandwidth.
>> Any system will require physical real-estate, hardware, software, licensing, support & maintenance and hosting/FM. This will add to the cost of delivering an intranet.
>> Anything that contains information will need management to keep that information maintained and up to date - this again will add cost and risk of inaccuracies.
>> In many companies (regardless of size) the transition to centralised electronic information vs. distributed or paper based information (think telephone extension lists, notice boards etc) is a difficult one to make - this could lead to a perception of failure of the centralised system if this is not managed properly - training and PR is essential, aswell as good requirements analysis.
>> It may also be a 'distraction' to users if it contains 'interesting' information not necessary for their job - newsletters etc (although useful) can get out of hand.
>> Also, single points of failure may be in question if the system is not HA'd in some way - load balancing, HA clusters would help, but you have to consider all the angles, such as power loss to the datacentre, or network failure on the Server LAN/VLAN - or worse, complete WAN failure. Without the information on the intranet site, could the users do their job ?
----
The thing is... this applies to almost any system, and is just part of the due diligence process all new systems should be put through (and much more) to ensure that it will help the business, not hinder it.
Further more, there should always be a business case for doing anything - what is the value proposition of having an intranet ? Why spend thousands or millions on a system that nobody needs?
Having said that... you only asked about why NOT... there are just as many (if not more) why you SHOULD have an intranet - in my opinion these usually far outweigh the potential risks (that can be easily mitigated). Businesses succeed or fail based on the quality of the people and their information - make peoples' jobs easier by making the information more easily available.
First things first - understand the requirements of the business, and then see how you can meet those needs with an intranet type technology - do you need a wiki, forums, notice board, contact lists, document store etc etc..
There are lots of other reasons in both directions.. search for Risk Assesment or Risk Management or look at standards such as ITIL, BS15000 etc to see what is important when providing Information systems, this will take you in the right direction.
Hope that helps
A smile is worth a thousand kind words. So smile, it's easy! 