For me Oracle has gone into the arena of companies like IBM and CA, where they pretty much only deal in enterprise IT scenarios and have no interest in small/end-user computing. Those companies can continue to make money on support and upgrades for a really long time as the pace of IT change in core enterprise systems is glacial.

Personally I think this kind of high-end only strategy is a really bad idea, as it cuts off your supply of new people who know about your technology and are developing new software for your platforms. I always thought it was a big reason why Linux did so well against proprietary Unix systems like HPUX, Solaris, AIX etc, and in turn why they did well against even more closed-off systems like mainframe products.

When it's hard for people to learn to use your product, due to expensive hardware and licensing fees and complex install processes, fewer people do it, which directly leads to increased costs to hire people in that arena. Compare that to Windows/Linux admins who can easily learn their craft on a cheap/free software and cheap hardware.

Microsoft is a good example of a company who generally gets that problem despite playing in the enterprise space. You can get 180-day licenses and VM images for most of their products for free (completely fine for training), and there's a load of cheap/free training materials.