Companies are required to keep books according to accepted accounting principles, under which sunk costs are in fact on your books, and that's what financial analysts look at, and those costs you need to amortize, both financial accounting and tax accounting. When you build a new factory, you charge what you need to to pay the price of the factory down (or assembly line). You don't spend lavishly on an expensive factory and say "sunk costs, let's call that money gone, what's our marginal cost, ok, that will be our price." Rather, you charge as high as the market will bear and hopefully make a profit. You don't control the market of course, and you have to respond to competition.
A competitor who has already paid down their investment has more flexibility than you do because they have more room to play with price and still report a profit.
The sunk cost fallacy refers to your decision making. What is the value of a project? its prospects, not what you have already spent. What has your spending done? It's bought you the option to make this decision. Do you call the option? Well, it doesn't matter how much you paid for it. Did you make money on the option? That does matter how much the option cost.
Some of it's psychological and perhaps not rational: if you have engaged in a big money losing project, if you pull the plug on it, your stock price will go up, because the market says "whew, at least we know they are not going spend money on that project any more" which I guess is them suspecting you will fall prey to the sunk cost fallacy. But these are complex decisions to make, especially in Europe where shutting down production entails a lot of costs that would be better put to something productive, and as you don't know the future, you try to figure out ways to make the project work.