I don't even know if it's that Oracle has slowed, but rather their product has become too complex for them to sustainably innovate. They tried acquiring other options for the low end, but it hasn't worked. (Berkeley DB is theirs, for example.)
That said, there will always been a need for the product Oracle provides in the largest of companies. NoSQL isn't going to be running reports on the fortune 500 anytime in the next two decades, and neither is Postgres, as much as I like it.
Death by a thousand cuts will probably still be profitable for decades.
I work for an enterprise company (10K employees) and you are spot on.
Oracle will forever dominant HR, Payroll, SCM, Billing, FM etc type systems where more often than not they have custom software on top of it that has been built exclusively for Oracle. There is simply no good reason to switch. If anything the threat to Oracle will come from hosted apps e.g. Salesforce.
But what I've been seeing in our company is MongoDB dominating front end web apps and Datastax/Cassandra for more of the big data, analytics type workloads.
Oracle should be far more concerned with Amazon Web Services.
Big companies don't want to house large IT infrastructure anymore. It seems too costly in the long run and benefits are difficult to explain to the board.
Consequently, you already see companies moving what was before considered core infrastructure to the cloud and the trend is probably going to increase.
Deploying a VM on EC2 or Azure is easy, quick and cheap enough considering you don't have to house as many talents as before in the IT department. As companies move to cloud based services, they don't deal with databases anymore, the people providing the solution (we don't even talk about software anymore, that's the current level of abstraction) does. Problem is amongst the cloud based service providers, few are Oracle shop.
That's the reason Oracle is now trying to position itself in the cloud through acquisitions (like Nebula today). Unfortunately, they are a bit late to the game.
I'm not so sure. I've worked multiple places where they fully took advantage of big iron. Data from one department fuels three other departments, who in turn fuel parts of reports for six more. Could it be done with a series of ETL jobs to feed data around the organization? Possibly, but I wouldn't want to architect it.
That said, there will always been a need for the product Oracle provides in the largest of companies. NoSQL isn't going to be running reports on the fortune 500 anytime in the next two decades, and neither is Postgres, as much as I like it.
Death by a thousand cuts will probably still be profitable for decades.