Agile Journal is a spinn-off of CM Crossroads (Configuration Management). In their first issue they published results from a agile survey. The survey was answered by the readers of CM Crossroad, so not specifically people interested or believing in agile development. Never the less, only 12% though that there is no benefit in applying these methods in their business. Still the most interesting point for me was that 32% answered that their organization is "first learning about Agile processes". This means that there is going to be huge leap in number of experience reports in near future. I'm still waiting for reports on long lasting affect of using agile methods. Most of the studies are on agile intervention level, where we can expect some bias from hype and enthusiasm. Some of the future research may even finaly close the gap between academic research and "agile research" (=story telling).
The survey had data from large organizations as well. As many of these were just learning agile methods, they are soon going to begin scaling the agile methods on enterprise level. Today in large organizations we usually have islands of agile knowledge, maybe just a agile team, but the positive experience fosters the information sharing.
The data from Agile Journal survey further showed that 17% had agile methods in use in two or more projects. This supports the fact that we are moving from extreme to mainstream. The study does not say anything about embedded system development, but I do not have any reason to doubt that our domain would not have been represented.