Json - is an ubiquitous data format, which supported in almost any popular databases. PostgreSQL was the first relational database, which received support of native textual json and very efficient binary jsonb data types. Recently published SQL 2016 standard describes the JSON data type and specifies the functions and operators to work with json in SQL, which greatly simplifies the further development of json support in PostgreSQL. We compare existing features of json/jsonb data types with proposed SQL standard and discuss the ways how we could improve json/jsonb support in PostgreSQL. PostgreSQL offers to application developers a rich support of json data type, providing known advantages of the json data model with traditional benefits of relational databases, such as declarative query language, rich query processing, transaction management providing ACID safety guarantees. However, current support of json is far from ideal, for example, json is still "foreign" data type to SQL - existed jsquery extension tries to implement their own query language, which is being powerfull, is opaque to Postgres planner and optimizer and not extendable. Extending SQL to support json, without commonly accepted standard, is difficult and perspectiveless task. Recently published SQL 2016 standard describes the JSON data type and specifies the functions and operators to work with json in SQL, which makes clear the direction of future development of json support in PostgreSQL. We present our ideas and prototype of future json data type in PostgreSQL with some further non-standard extensions and improvements in storage requirement and index support.