Start with high-quality items of the absolute minimum needed and add more/change/adjust from there. ‘Blank check’ type scenarios never work well, because centrally-planning something such as a game development or pipeline is impossible to do effectively. It is much better to stay lean and flexible, and iterative, just like a game design, instead of trying to plan everything up front.
Try to find the sweet-spot for system specs- get good RAM you can upgrade and add to artist machines as needed, don’t just get 4GB for everyone off the bat. Procs, again, find that sweet spot. Make sure mobo’s and PSU’s are high quality.
Choose video cards depending on your target specs. If you are working on a game to release in 4 years, maybe start with 8/9 series and upgrade to 280’s in a year when they are cheaper.
Don’t add render farm machines until you start your RTT and lightmap baking.
Artists should be able to choose a Wacom when they are hired.
If given a blank check to do something, you will always get the most value from it when you don’t treat it as a blank check, treat it as the minimum you need. Living within one’s means always yields the greatest prosperity and enjoyment; probably the words of Ben Franklin would be fitting here: http://www.pickthebrain.com/blog/ben-franklins-way-to-wealth-part-ii-frugality/ .
It is easy- and fun- to ‘plan out’ what to do with a blank check, but one should never make decisions based on that.
Sorry for the slightly roundabout answer but I felt I needed to say it. I see people plan things way too much and it never works out.