My outsider perspective is parts of the Media Lab had an ivory-tower, idealized view that Thinkpads running Linux are the proper way to compute (no judgement, I still believe that), thus the OLPC should resemble that, but childlike. One route to productionization would have been to move the platform to existing expensive Windows(/Intel) laptops after the first production run proving that a $100 BOM was quite possible. Moving to Windows laptops would have been a very large compromise to the original project's high-minded, ambitions goals, but one that would have preserved the project. The Windows kernel and Explorer isn't open but that wouldn't have prevented porting the Sugar interface to Windows and kept the main UI software goals of the project open source and editable/"compilable" directly on the laptop.
It feels like they didn't have enough Plan B's for when their Plan A's failed.
:-(
At that point in time, the OLPC was perceived as an existential threat to Microsoft and, to a lesser degree, to Intel. Intel because it had nothing that could be used in them (which was solved to some extent by netbooks and Atom processors). For Microsoft, it meant a whole generation of kids could grow up having never used Office or Windows, something that Microsoft would not be able to counter unless OLPC failed.
Brazil was pretty much the same. They made an incredible effort to push Atom-based computers running Windows that had none of the more interesting aspects of the OLPC - not the sunlight-friendly screen, the mesh network, the software for group activities, the openness to tinkering...
I don't think the OLPC team realized its mere announcement would be a declaration of war. They'd need Intel's buy-in (at least to cut Microsoft lose) and they'd need to quickly mass-market the tech through someone like V-Tech in developed countries to get funded through hardware licensing and financial commitment to get software and content developed.
Then they could properly scale up to government level and demonstrate the Microsoft argument that the kids wouldn't be prepared to join the workforce (and enjoy the Microsoft lock-in) was void.
Now, imagine the colossal power it would be to be able to help shape the content kids all over the world would use to learn. I would expect the US government to get interested in that, but it didn't.