Nothing personal DJ, but I don't learn much from these AI summaries, whoever posts them. They seem to find a couple of facts and then re-state them in different ways to make them sound smart.
The tests were run at active lengths of 250, 500 and 1000 microns, much smaller than the cited values of 5mm, 10mm, and 20mm. Why did they test those specific lengths? Are they only relevant to CPO applications, or are they applicable to transceivers, too? That's probably the question that's most important to me.
I initially thought the testing was directed at the Stage 3 partner working on CPO, but then I noticed the different active lengths on the graph. Also, the post refers to "reference devices", not a specific device, so, yes, the testing is directed at a broader audience. I'm trying to get a feel for who that audience is based on the specific active lengths they tested.
Market Data 
Markets 
