The casing is about 2 hours 3d-printing a piece. That is 20'000 printer hours for 10k units, giving a machine-time cost of 0.025 USD (including plastic). That i

The casing is about 2 hours 3d-printing a piece. That is 20'000 printer hours for 10k units, giving a machine-time cost of 0.025 USD (including plastic). That is off by a factor of 10-100, at least. For injection molding, one would need to make the molds. That would cost probably more than 1000 USD, excluding engineering costs (the design is not injection molding ready). Amortizing the mold costs over 10k units alone would be 0.1 USD per unit. Then comes actually making the parts. Maybe 0.5 USD is doable. I checked an off-the-shelf casing (no custom molds needed) it was quoted at 0.20 USD at 1K units - but this then needs to be CNC machined on 2 sides to add the necessary openings. Probably 1 USD a piece, at the least. The costs in the scientific article does make any sense. It is really sad to see that it gets through peer review so easily... I mean OK that the scientist does not really know what things cost to produce, but then they need to consult someone who does! Ask a manufacturing or even a design house for a cost estimate. Quality check your numbers at least a little bit.