The claim of saving "up to 6W" is probably legit - important part is "up to" of course as chain losses related to load (in a non-linear way) and expressed in %%.
So it is possible that given
new chain at the certain load will loose those 6W vs same chain after it was
properly broken-in, cleaned & lubricated.
Unfortunately that website does not disclose any details and "test data" they have is a joke really
Supporting Data UltraFast Optimization
Looking at the link above they show on graph 1 new chain losses vs graph 2 broken-in, cleaned and waxed chain losses - right there difference is already almost 4W.
It is well known fact that broken-in chain is more efficient, waxing and cleaning has noting to do with that.
Besides you can clean your chain at home using ultrasonic cleaner (they around $50 or so on sale) and wax/lubricate it as you wish.
Clearly even from their own test data that "PTFE/MoS2 lubricating layer" only added tiny bit of gains so why bother and pay for it in the first place?