The RaspPi has three innovations: size and price and operating system.
It's size makes it possible to include it in projects that would not be possible if it were just twice the size. Its price makes it disposable. Its operating system makes it flexible.
If I want a micro-controller, then I buy an Arduino, because it is cheap. I can package it up into a project, and then buy another for the next project. I can even design it into a project where it might be lost. It is disposable. If it was $50 or $100 I would have to think twice.
A bigger board is more difficult to package. I can't build a wearable PC with a Eurocard-sized computer. I can't fit it into a R/C model or a meteorological balloon.
But a microcontroller has limitations. I have to write most of the software myself. That gets expensive and tedious. The software I write is single-purpose, whether it is controlling a quadcopter or a mobile phone. But if the computer runs Linux (or Windows, or QNX or any of a wide range of others,) then I get vast swaths of software as standard. I can add a web-server or diagnostics without a thought. My VOIP telephone is programmed in a week instead of three months and has more features.
There have been small cheap computers before (eg Arduino), there have been small Linux computers before (eg Cotton Candy). And any old PC can run Linux cheaply. But the Raspberry Pi is all three and that has never been done before. I don't believe it was even a design goal for the RPF; they wanted a $25 computer that could be programmed by children, and Linux and the size were ways of achieving that goal, not ends in themselves.
I can believe that there was no demand for something like the RaspPi. No one has ever made money selling only to hackers. Commercial companies think nothing of paying £100 for an SBC or writing their own software stacks.The hobbyists and the small developers did not have the ear of the manufacturers, nor the volume that made it worthwhile. The larger companies believed themselves well-served by the more expensive and more powerful products on offer, or wanted to package into a single-purpose device such as a mobile phone.
Even those hobbyists who did manage to get a project off the ground failed to understand the difference a few dollars makes. If a Beagle Bone costs $90 I might buy one and use it for a few projects, tearing the last one down to make the next. It might be a prototype platform, but it is not a production platform. But if a RaspPi costs $25 I'll buy one for each project.
Small, flexible and disposable. That's the key.