When you are trying something to do something truly new and different, trying to scale too fast can kill your business even before it's really gotten started. A truly revolutionary product is unpredictable because, by its very nature, it's not well understood. Charging boldly into the unknown is a sure way to run into unanticipated problems that are expensive to fix at scale.
A better strategy is to identify a "hair on fire" use case--someone who needs a problem fixed so badly that they are willing to overlook the inevitable glitches. They will help you identify shortcomings early and help you correct them. Once you get things ironed out, you can begin to scale for more ordinary use cases.
For example, developing a self-driving car is a risky proposition with a dizzying amount of variables you can't account for. However, one company found a remote mine in western Australia where drivers are scarce and traffic nonexistent. It was an ideal place to test and improve the technology. In a similar vein, Google Glass failed utterly as a mass product, but is getting a second life as an industrial tool. Sometimes it's better to build for the few than the many.