Deepnote internally supports .ipynb format and you can always export the Deepnote notebook to .ipynb similarly as you'd in Colab.
In general the main selling points are live collaboration (you can work on a notebook with you team as you'd do on a google doc), and integrations (you can plug-in your snowflake db, or s3 bucket or whatever, and have it connected for any further analysis, or a long-term training, etc.
For many non-software-developer data scientists, it's also easier to work in a cloud environment compared to installing stuff locally, and to version their notebooks in Deepnote instead of git. But this really depends on the particular workflow that one has.
In general the main selling points are live collaboration (you can work on a notebook with you team as you'd do on a google doc), and integrations (you can plug-in your snowflake db, or s3 bucket or whatever, and have it connected for any further analysis, or a long-term training, etc.
For many non-software-developer data scientists, it's also easier to work in a cloud environment compared to installing stuff locally, and to version their notebooks in Deepnote instead of git. But this really depends on the particular workflow that one has.