It is difficult to say for certain without their scientific thesis to study over.
With that said there is some validity to wanting more razors, and some there are reasons against it.
Consider this Shaving technology
This in no way proves that adding more razors is purely for marketing purposes, but it
helps raise the question understand that there is a point when there is too many razors and offers an idea to why there one is seeing more razors.
How stuff works
This describes that there is not a lot of information to go on, as this is a topic that scientists outside of Gillette rarely cover.
In addition only a minimum of two blades are required to cut and overcome nicking of the skin. The reason for adding more blades only further reduces that chance of nicking by not allowing the skin to 'hill'.
Apart from that it comes down to sampleing frequency. With two blades you can get a single cut from one stroke. One blunt blade to bunch the hairs, and second blade to cut them. With 4 blades you can cut twice with one stoke.
The final point to be made is that all designs have an optimal setting and razors are no different.
In a set distance we can place one razor.
[--------------------------] no razor
[------------/-------------] 1 razor
Certain qualities may be improved with two blades, such as reducing of nicks.
[------/-----------/-------] 2 razors
Again we can increase the # of razors and some key parameters will improve, such as sample rate.
[---/----/------/-----/----] 4 razors
When we reach the maximum possible number of razors. When the blades are paper thin, and there is no space left...
[//////////////////////////] Theoretical infinite razors.
The razor can not cut any more, it has become useless and the key parameters have decreased to zero.
There does exists a key point in the middle where blade spacing and size give the highest performance.

*Note this curve is over simplified.
So what the Gillette company is claiming is not entirely false. However, again without there thesis it is hard to reproduce their results and validate their experimental claims.
It might happen to be that any razors past 3 do nothing to increase the performance.