Like all technologies that follow a hype cycle, ORMs have reached the "plateau of productivity". People have gotten over the vietnam of computer science, accepting the fact that ORMs are good at what they do best: Loading and persisting complex object-graphs from and into a relational database.
But not everyone sees ORMs as what they are. Many people still have hopes towards a better integration of Java with relational databases. Some may even think that escaping to "NoSQL" would solve their integration problems. This was recently illustrated by this popular InfoQ article, which contains a survey about how people prefer accessing their relational data from Java. Interestingly, jOOQ, an increasingly popular newcomer on the scene, already competes very well with more established Hibernate alternatives, such as MyBatis or EclipseLink.
jOOQ offers a very novel approach that appears to be similar to .NET's popular LINQ-to-SQL. However, unlike LINQ and various Java or Scala mimicks thereof, jOOQ focuses on implementing only SQL as an internal domain-specific language in Java. Many Java developers know SQL very well and are missing appropriate tooling to have their data access code properly interact with SQL. Using jOOQ's source code generator and expressive fluent API, embedded SQL can now finally be compiled by the Java compiler directly, even for very complex statements.
With Data Geekery's recently announced dual-licensing strategy for jOOQ, things can only get better in the ecosystem of Java and SQL integration, where SQL as a powerful and very expressive language really matters.