All AMD does to make the FX9xxx line is to take their 8xxx series chips that are their best examples and overclock them, a lot. So they are basically already as fast as they are going to go. You are paying a lot of money for something you can nearly achieve with a much cheaper chip. Same holds true for the FX8320 and FX8350, just different grades of the same chip. That is why the TDP is so much higher (they used to sell them without coolers, now they bundle a liquid cooler)
Intel does the same thing. Celerons and Pentiums are basically rejects of fully functional i3 and i5 processors. You will notice that the high end chips always come out first, because they then take those 'broken' chips, disable the parts that don't work and sell them for less. They have to stockpile enough CPUs that are 'damaged' in similar ways to release a product line.
AMD has a few examples as well the Athlon 760k and 750k are A-Series APUs where the graphics chips failed to work, so they are disabled and released.
Same in the GPU world. Though Nvidia has been confusing things lately with their branding. Used to be that Ti cards were a crippled version of the next model up. AMD still sticks with it though. 280 is a crippled 280X etc.