As you've picked up, there are several aerodynamic and aero-acoustic advantages that can be achieved from serrated trailing edges, however they are many issues that make it difficult to implement. A by no means exhaustive list would include: patents; ergonomics/safety of those edges in consumer hands; increased tool complexity and manufacturing costs etc etc
Perhaps most importantly, aero-acoustics is a relatively young field of research. Our ability to accurately model and correlate the noise generation is still fairly limited even in the most advanced cfd codes. For the (relatively small, low volume) fan companies, a huge amount more modelling and research would be required to tune the blades to the desire levels, for a potentially tiny gain.
The other major factor to consider is that the airflow around the fan blades is often not the dominant source of the noise (although a large contributor), a lot of what you can hear is driven by the fluid-structure interaction with the surrounding geometry. The manufacturer could tune the blades to be quiet whilst hanging in its shroud in free space, then you could put it behind a grill in your case, or stick it onto a radiator, or you could have it acting against a positive pressure gradient in the case. Essentially any number of small things you did in the installation could render their work useless as you alter the details of the flow around the blades and hence the emitted noise spectrum.
