I haven’t looked into this specific design and I’m no CFD expert, but I have played with designing and built several small empirical test setups to compare and contrast various designs for shrouds.

Don’t put trust in CFD. These little fans are turbulent flow regime and nowhere near laminar. These two flow regimes are completely unrelated and laminar flow simulation is not relevant in the slightest. You can get turbulent CFD simulation software but they all cost many thousands of dollars as far as I am aware. The last time I went down this rabbit hole was a year or more back.

The boundary flow effect is a major factor in airflow, as is the total length of the system. The rough printed surface area of this design is likely hurting the total flow. Any sharp face angle transitions will cause eddies to form.

The deeper you get into real airflow design, the more you will find that almost everything in the 3d printing space is wrong. The best source of intuitive information I found was looking for info about HVAC duct systems in large buildings. You will find a good bit of info in that space from qualified engineers and academics, and HVAC duct flow is all in the turbulent regime.

Long before I got into 3d printing I was into hotrods and also worked for a bit at some machine shops building engines and porting heads. I was into trying to understand airflow in intakes, heads, and manifolds. I’m still no expert, but I can look at a design and see, intuitively it doesn’t align with all of the super engineered stuff I’ve come across. I know that is super subjective, but like I said, the HVAC space can produce good info and rules of thumb equations about duct length, transition losses, and the impact of surface roughness on the boundary layer. Any basic info in these areas will make almost all designs in the 3d printing space seem like nonsense.