I went without signing contracts a long time, just direct sales. In the beginning it is probably smart to only develop on processor at a time.
There are 4 brackets, top end, mid range, low end, industrial.
In the beginning it is probably the easiest to go with top end.
For lower end processors the money gets made in volume and fab lines are very expensive.
Compare your processor to the ones on sale from the other companies.
Always keep researching.
Processor speeds not so important. MIPS is more important. I had late game unexpected success with with a very low watt processor.
The formula behind "demand" is taking into accound the prices of the other processors in your bracket and if they are better or not.
But demand is also calculated from the demand yesterday. So if you automate your manufacturing (I did almost always, much more efficient) it can be, that you produce 0 and you keep decreasing your price but demand doesn't take up. Retire old processors to not compete with yourself. To ramp fast a new product, keep with the recommended price start raising prices when the production and demand picks up.