I run a large tower. This requires 28800 blocks of fuel per month, which is 144000.00m3. I live in moderately deep wspace: I cannot always find a highsec exit. Also, the path out goes via random wspace systems, which may have hostiles present. So the bandwidth of goods in and out is a problem worth worrying about.
Worse, I am planning to do significant PI to make money. If each of 10 characters makes 5000m3 of PI goods per planet per 9 days, as I discussed in the previous post, then that is 833333m3 that have to be carried out per month. This can certainly be done; at 35000m3 per Iteron V load, it's 24 loads. But it would certainly be nice to reduce it.
One way to reduce it would be to create P3 goods. In the step from P2 to P3, some shrinkage is attained: the factor is either 0.6 (for P3 goods requiring 2 P2 inputs), or 0.4 (for those requiring three P2 inputs). So, one way to proceed is to produce P3s, particularly the ones which require three inputs. We have one of those we can make, so that's one thing I will do.
Another thing we might do to decrease bandwidth is to make POS fuel. (Certainly, that one should make one's own fuel is a popular idea among wspace dwellers.) What is the size reduction of POS fuel of its inputs? And what is the input volume of ice products? I don't know, but I am about to figure it out.
... OK, I just punched numbers into my spreadsheet (a copy of the one that I discussed before). The result is pretty clear. Here are the facts:
- POS fuel blocks (5m3) are a little bit smaller than their inputs (5.75m3).
- The ice-derived inputs to POS fuel (4.5m3) are just slightly smaller than POS fuel itself.
- The PI-derived inputs to POS fuel (1.25m3) are a small fraction of its volume.
Given these findings, as well as the low profitability of POS fuel, and the fuss of having to set up PI to create its inputs, as well as the opportunity costs, I find no value in making POS fuel myself.