Tesla is definitely using DBE in their 4680 production lines. Below is a plan I've posted before of the main 4680 production floor for the battery cell building at Giga Berlin. This is one of over 2000 pages in one of the files I've looked through that was submitted for the recent planning application. When zoomed in to the anode cal/lam production area it states that it is using a "dry powder". The production floor includes tesla silicon production, anode and cathode mixing and anode and cathode cal/lam, foil coating and can stamping. There is no inclusion or space for any wet slurry or drying ovens in the 4680 production proposed so it must be DBE.4680 production will absolutely use a DBE process. That decision needs to be made before any factory is produced given the alternative is a wet electrode process that requires football field length drying ovens. It sounds like Tesla is not getting the yield from the DBE process they need for commercial production just yet.
As a purely theoretical example - let's say they need 80 out of every 100 cells they make to meet their performance specifications before they scale up production (otherwise the cost per good cell would be too expensive when defective cell costs are included), the current process using DBE might only be producing 75 cells meeting these specifications. The DBE process is not quite refined enough yet due to some variability in output from the process. Tesla "just" needs to tweak the DBE process a little to get more consistent results to the point that Tesla is confident that 80 out of every 100 cells meeting their specifications.
The latest update from Elon was a couple of months ago stating that they will need somewhere between 12 and 18 months to get the process refined enough for commercial production.
I also recall seeing a twitter thread this week talking about Kato Rd being at iteration 5 or 6 of the production line and basically ready for volume production (I haven't found this again, so if anyone knows it then please link). I take Elon's twitter comments as being about continuous improvement in the 4680 DBE process and not suggesting any significant delay in initial production.