These cost figures come from a number of old sources, Combat Fleets of the World and old Proceedings magazines, some issues as old as 1977. I repeat, cost figures are very misleading, some refer to price per submarine, other cost figures reflect R&D costs.
The Seawolf class of SSNs for example including R&D was from the current issue of Combat Fleets of the World. There was a large R&D effort and expenditure for the Seawolfs, whereas the Virginia class of SSNs R&D expenditure wasn't as much.
Frankly, I don't like to answer costs questions because its difficult to narrow down the price per ships because it depends on programme R&D, the number of ships bought per class overall, the number of ships bought per year, and which year the ship was bought, ships built early in a long programme cost less in the beginning and cost more at the end of a programme.
There is a huge difference in FY dollars as well. Obviously ships cost less in FY 77 dollars as compared to FY 97 dollars.
As I recall New Zealand purchased two Anzac class frigates in FY 88 for $NZ 470 million each, or $NZ 940 million dollars. New Zealand passed on purchasing one or two more for the same FY 88 price during 1998-1999, ten years later. One of the reasons why Australia and Tenix offered these extra ships at a price ten years prior more than likely reflected a 12 ship programme versus a 10 ship programme. Acquiring ships in numbers usually result in a lower price.
Another example of this confusion is the price of Australia's upcoming LHDs. Combat Fleets says the Mistral cost $235 million in US dollars back during FY 2000, while Australia is willing to spend $2 billion Australian dollars for two, or $1 billion Australian each during FY 2007. The last I looked the US dollar wasn't worth two Australian dollars. It appears if the Australians purchase this LHD they are also paying for some of France's R&D costs. Not to mention seven years of inflation have inflated the price. Australia is probably paying for a complete support package too, which the French more than likely didn't include.