The main Databricks experience is essentially the same on both Azure and GCP. The difference is in the cloud infrastructure that supports them.
Azure Databricks is a bit more integrated with Azure services like Azure Data Lake, Synapse Analytics, and the Microsoft ecosystem in general if you are already using them. This can be a great advantage... if you know...
Besdes that, it also makes security better for business customers in some aspects. If you use GCP's BigQuery or other Google-native services, GCP Databricks, on the contrary, might be the one that fits your needs better.
It could be just a minor thing that you may notice that GCPโs networking and latency have the potential to be quicker for certain workloads, depending on the location of your clusters.
But at the same time, it is very specific to a particular workload.
Therefore, if you are concerned about performance, then you should consider where your data is stored and the entire cloud infrastructure so that you can be sure that you are using Databricks with the best cloud for your needs.
Just so you know, latency changes are typicaly small unless you have many heavy, real-time streaming workloads.
There are people who say that GCP is more convenient to scale, particularly when it comes to ML pipelines. However, this is dependent on the use case.
So, don't just think about Databricks; think about the whole stack.
More information about Microsoft Azure can be found here
