The Economist’s 2006 Annual has an article by Nicholas Negroponte:
Heres how well do it. To start, at least 50% of a modern laptops purchase price is swallowed by the cost of sales, marketing, distribution and profit. OLPC has none of those costs. Our device will not be available in retail channels, although to discourage a grey market we will authorise production of a commercial version, where a share of profits will be dedicated to further lowering the cost of the OLPC machine. Distribution in most cases will simply piggyback on well-established textbook channels.
The remaining 50% of the cost of a laptop can be divided into two roughly equal parts: the display, and every-thing else. Without question the display is the technical challenge. The Media Lab has short-term ways to bring the cost of a laptop display close to $30 and longer-term solutions like E Ink (which we invented) that could eventually be as inexpensive as 10 cents per square inch for a full-colour, sunlight-readable screen with better-than-textbook resolution in print mode.
The interesting part of laptop economics is the every-thing else: the processor, memory and power management. Todays laptops use about 75% of their own processing capacity to support bloated software applications and the operating system itself. I am reminded of a suspension cable where after the span reaches a certain length most of the strength required is to support the weight of the cable itself. Likewise, software has become self-serving, obese, more complicated and less reliable. The solution is a strict diet. In the case of the $100 laptop, this also means open-source software for kids all around the world to participate. For us, the answer is thus a skinny Linux operating system.