For handling large JSON files using read_json() is not an efficient way. You will always get an memory error. Because this will take complete data in a memory and process it further.
You can try ijson module that will work with JSON as a stream, rather than as a block file.
Also worth a look - Python & JSON: Working with large datasets using Pandas