Out of the Textbooks and into Real Life: Deadweight Loss

I've enjoyed Matt Yglesias' reincarnation as a microeconomics professor:

What about the case of unauthorized downloading of music files. Consider Katy Perry’s “California Gurls”. This tune costs $1.29 on iTunes. At that price, some people will buy it. Others will refuse. You might refuse because you hate Katy Perry and hate the idea of owning one of her songs. But say you’re not a hater. You’re just a skeptic and a cheapskate. You’d gladly pay a dime for the song were that an option, and since the marginal cost of distribution is basically zero it would be profitable to sell you the song for a dime. But it’s not an option, since the overal profit-maximizing price is $1.29. And say there are a million people like you out there. That adds up to $100,000 in deadweight loss—the value of the transactions blocked by copyright protection.

I think he gets this wrong, though. While price should always equal marginal cost in a competitive industry, I'm not sure iTunes would be able to stay in business if it was charging only 10 cents per song, because presumably the fixed costs would be too great. Therefore, when it comes to digital music vendors, we might have to choose between paying far more than the marginal cost per song or vendors not staying in the market.