Here's a scenario. My company, who you'be never heard of before, has just created a revolutionary new car. It will run almost as well as a new "normal" car on just peddle power, it's amazingly cheap, and is a paradigm shift in transit. Even better, you can maintain it yourself, and they won't be stolen because they have a distinctive look to them (kinda like mail trucks).
Trust me? Good! Because you can't just go to a car dealer and buy one of these things, your whole community has to go in together and everyone has to get one, regardless of whether everyone is really interested in them or not. You have to buy at least a million of them, so your community has to be reasonably big. And you can't take any for a test drive to see if you actually even want to use it.
Now, city governments are interested in buying these, so regardless of whether their citizens are interested, they've been sold on all the promises given by this new company, and are going to take out loans or raise your taxes to pay for the cars.
Does this sound like a good idea? Would you want your city doing this? Without anyone giving it a test drive, seeing if it even works on the city streets, or if people are interested in driving them?
This is how I see the OLPC implementation plan currently. I'd love to see the OLPC project succeed beyond its wildest dreams. A paradigm shift in computing? Global unfiltered Internet access regardless of if you were born in Ghana instead of the Netherlands? Can you imagine the boon that would be to education, human rights, and marginalized groups?
The risk is really high, however, for an untested technology that has not been piloted in any country with a high amount of "trust me" in it (the kids will teach the teachers, the kids will maintain the units, no one will try to steal/re-sell the units...).
These things may still happen, and that alone wouldn't be a reason to not go forward, but requiring countries to purchase at least 1 million units, most likely through World Bank loans, makes the consequences of failure very high, and with little reason to be so high.
If the laptops are as great as they seem to be, and all the "trust mes" come true, and the educational theories are solid, then pilot programs will reveal their power, and everyone will be streaming in orders. Sure, OLPC needs to leverage scale economies - big shipments - to get the unit price down to $100, but this is not some impossible obstacle - sign up 10,000 here, 100,000 there and let countries do their own pilot projects with the initial split-up orders of 1M globally. They're so simple to use, so we shouldn't need to worry about support from OLPC to all these various sites, so I don't see any reason not to let this hapen. The laptops will sink or sail on their actual merits and in-country utility, not hopes and debts.