some points why i don't think that you can really do a prediction just based on these few numbers.
to see what other factors exist over 1 year i look into last fy:
first half of last fy they were selling the old models which still cost them a lot more and were probably highly deficitable, then in autumn they introduced the newer models and a pricecut, with rumors of production cost beeing in 400$ range by that time. Selling at 400$ still doesn't make it profitable since a lot other costs have to be added on the pure production costs, but i expect loss should have been down to less than 100$ per console by that time.
But soon after the 100$ pricecut in october last year, dollar/yen exchange rates started dropping too and prices for sony actually dropped by another 9-10% for the holiday season, and another few % after holiday season. Which i guess had a quite a negative effect on sonys earnings since that also coresponds with the biggest selling period.
for this year things look a lot better: ps3 is still at about 400 euro in europe which is now an about 50% higher price than in the us, so its very likely that they are selling ps3 hardware for profits in europe, while they might still have a loss per sold ps3 in the us. Its expected that the new shrinked chips will be integrated later this year so production costs should definitely go down again, and if sony doesn't have to do a pricecut they might reach hardware profitability worldwide.
Also starting with this april it seems that the dollar is going up again a bit compared to the yen (or maybe yen is dropping a bit) so its at 106 yen for 1 dollar now (lowest was about 100 yen in march which also might have made sony losses of last fy look worse in dollars, than if you would have applied an yearly average exchange rate).
so all summarised your model is very simplified since you are just dividing lossed by the sold numbers and saying they are averagely loosing this much per unit. While the indicator loss per sold ps3 is an interesting thing for watching how the ps3 evolves over time, i don't think we have enough data with it to make a valid forecast how that number will evolve for next financial year.
we don't even know how losses per unit sold evolved over the span of last fy. As an example PS3 might have been highly negative for the first period of the fy with slowly decreasing losses over the next few months. then a big change when they introduced the 40/80gb models and from that point on slowly reaching profitabilty. Its allready possible that they might have allready reached a profit in ps3 hw close to the end of last fy or start of the new fy if you base it on this model, since very small profits would still not have much effect on the big overall loss last fy.
But without knowing details another model might also be valid which has nearly stagnant losses per unit sold for the times before the 40/80gb models (&pricecut) and after that another loss per unit number which is nearly stagnant too.
also based on only 2 estimates for the fy 07 and the fy06 it is nearly impossible to make a valid projection.
(and for those 2 numbers its still very hard to say how valid they are since we don't know what sony added all into these numbers. they even might have split hardware developement costs across a few financial years before and after release)
