Performance Boost repetitive Measures

I have a data model that consists Dim_Item, a Dim_FAC_Storage and a fact table called ItemLedger. These tables are well-structured.
One special feature in the Itemtable is that each article has 9 levels, which represent a pricing assessment.

The report is designed to always show values as of a specific reference date. All values are calculated cumulatively up to that date.

There is currently a matrix in the report that displays both stock quantity and stock value. The stock value should be dynamically changeable via a filter, allowing users to switch between Level 10 and Level 90.

The stock quantity is calculated as sum of ItemLedger[Qty].
The stock value is calculated by multiplying this Quantity by the selected level value from the D_Item table.

Right now, I have created one measure for each level — so 9 measures in total — all identical except for the level used.
Since I also need to calculate the same values for the Start of Month, I now have 18 measures.

To save on additional measures, I used a visual calculation to calculate the delta field.

I believe there must be a much simpler way — ideally, I could combine the base logic for all levels into a single measure instead of maintaining 9.
I already tried using a SWITCH function, and it worked, but performance was extremely poor — the table took around 40 seconds to load when switching the level.

Please note: this test was done on a small data subset. The full dataset includes:

• ~ 1 million articles
• 20 storage locations
• ~ 20 million entries in the ItemLedger table

What would be the best practice for this requirement to achieve high performance?

All relevant files are attached here: https://drive.google.com/drive/folders/1U8yNJY9dl8OzXUTu7AgPfNUThCh12GgS?usp=drive_link