The new EMI recording of Wagner's Tristan Und Isolde, starring Placido Domingo, is the subject of a very strange publicity blitz that seeks to portray the recording as the end of an era, the last big-star studio opera recording ever. Basically they're trying to make the recording into a success by calling attention to the fact that opera recordings are no longer successful enough to make. Or something.
I have to agree with Alex Ross that this is pure hype and that "What does seem to be dying out is the practice of spending a million or more dollars to make a studio recording of an opera that could just as easily have been taped live." Still, the fact that the big labels are mostly getting out of the studio opera game is nothing to be enthused about. A great studio opera recording, like the Solti "Ring," can be listened to over and over again, whereas a video recording, which is what we'll probably mostly be getting in the future, doesn't hold up as well on repetition (because visual mannerisms pall more on repetitions than musical mannerisms). Besides, it's not good that many important singers will not have a chance to record their best operatic roles.
On the other hand, it's not like big-label opera recording was ever a particularly accurate reflection of who the best singers were in particular roles; the star system, by which labels signed certain artists and recorded them in every standard-repertoire opera, meant that opera recording was more a commercial enterprise than an accurate record of opera history. The obvious example is Maria Callas: EMI signed her to a big contract, but wouldn't let her record most of her favorite roles (mostly bel canto operas, which didn't sell), and instead had her record a lot of operas that she didn't sing all that often on stage(Boheme, Butterfly, Pagliacci, and so on). Some of those recordings are very fine, and reflective of a time when record companies actually understood how to make money off classical recording -- but they are not reflective of opera history. We're always going to have to turn to bootlegs and broadcasts for an idea of who was singing what and when; studio recordings ar a separate form of entertainment, with different rules.
The upside of the collapse of the big labels and the big-label star system is that it gives non-superstar artists a chance to record complete operas, because the smaller labels can record a "Figaro" or an "Orfeo ed Euridice" without fear of competition from the big labels. Hence Rene Jacobs will be recording the Mozart chestnuts La Clemenza Di Tito and Don Giovanni with non-superstar casts, including excellent but under-recorded singers like Alexandrina Pendatchanska. Also coming soon, and also of note, is an English-language studio recording of Smetana's "The Bartered Bride", based on a popular live production and conducted by Czech music specialist Charles Mackerras.
So while I do miss big-label opera recordings -- at their best, they're very entertaining -- there are certain advantages to the collapse of the star system in classical recording.