There is an open source Delta that is a very good library. This is not the same as Databricks' implementation and there are at times compatability issues. For example, by default if you write a delta table using Databricks' dbr runtimes, that table is not readable by the open source Delta because due to the "deletion vectors" optimization that is only accessible within Databricks.

That aside, I was more pointing out that Delta, particularly via a commercial offering, is a data format biased towards Spark in terms of performance, since it is being developed primarily by Databricks as a part of the spark ecosystem. If you are plan to use Delta regardless of your compute engine, it makes perfect sense as a benchmark. However, for certain circumstances, the performance wins could be (in my case was) worth it to switch data formats.