Micro-LED Redux
While we waxed optimistically about Micro-LED in yesterday’s note we feel obligated to put some of that optimism in perspective. Leyard (300296.CH), one of the two companies mentioned in reference to the new production level Micro-LED chip that was the basis for our note, has been developing production capabilities toward offering a number of chips and modules with progressively smaller pitch (the space between Micro-LEDs) implying the higher pixel densities needed for AR/VR applications. That said, the company’s public display business, its primary revenue source, has seen business deteriorate (net profit down between 5% and 9% in 1H) as lockdowns in China reduced tourism traffic and signage demand by ~50%, while the company’s VR related business increased by 14% y/y, although it remains a relatively small part of the company’s net profit.
Leyard did indicate that great progress has been made with the company’s Micro-LED products, as we indicated yesterday, and that the yield rate on Micro-LED mass transfers increased from 98.9% to 99.995%, a substantial increase, but to put that into perspective, if we were to look at the transfer of Micro-LEDs from wafers to substrates for a 4K display, a 98.9% yield would leave 91,238 pixels unusable, while a 99.995% yield would generate 415 bad pixels. If we look at the fact that each pixel is made up of three sub-pixels, those numbers increase to 273,715 bad sub-pixels at 98.9% and 1,244 bas sub-pixels at 99.995%.
What makes these numbers significant is that these bad pixels/sub-pixels need to be repaired, and that cannot be done using the same mass transfer techniques used to bring large quantities of Micro-LEDs from a wafer to a substrate. Each bad Micro-LED must be removed and replaced, a far more tedious and expensive process than the initial transfer, so step function improvements in transfer yield are directly related to process cost. There are techniques that can be used to offset the replacement of bad Micro-LEDs, such as creating non-operating duplicates for each sub-pixels during the initial transfer to act as spares that can be activated if needed, but the cost of same is equally prohibitive, leaving the ultimate goal to reduce the number of defective pixels/sub-pixels to near zero[1].
We are certainly not criticizing Leyard or other Micro-LED manufacturers about yield improvements, which have been significant over the last year, but investors should understand that while a defective segment of a memory IC can be ‘closed’ with the device being sold with lesser specs, in a visual device bad pixels are far more obvious and must be repaired, , keeping comparative costs high until yields approach 6 or 7 nines or other Micro-LED techniques are found. While Micro-LEDs an interesting and potentially lucrative display technology that will get increasing press coverage over the next few years, it will take some time before it is able to stand as the basis for a consumer oriented display product that is within the affordability of the average consumer. Progress yes, ready for Prime Time, not quite yet…
[1] A Micro-LED with only 1 bad sub-pixel would have a yield of 99.999996%
Leyard did indicate that great progress has been made with the company’s Micro-LED products, as we indicated yesterday, and that the yield rate on Micro-LED mass transfers increased from 98.9% to 99.995%, a substantial increase, but to put that into perspective, if we were to look at the transfer of Micro-LEDs from wafers to substrates for a 4K display, a 98.9% yield would leave 91,238 pixels unusable, while a 99.995% yield would generate 415 bad pixels. If we look at the fact that each pixel is made up of three sub-pixels, those numbers increase to 273,715 bad sub-pixels at 98.9% and 1,244 bas sub-pixels at 99.995%.
What makes these numbers significant is that these bad pixels/sub-pixels need to be repaired, and that cannot be done using the same mass transfer techniques used to bring large quantities of Micro-LEDs from a wafer to a substrate. Each bad Micro-LED must be removed and replaced, a far more tedious and expensive process than the initial transfer, so step function improvements in transfer yield are directly related to process cost. There are techniques that can be used to offset the replacement of bad Micro-LEDs, such as creating non-operating duplicates for each sub-pixels during the initial transfer to act as spares that can be activated if needed, but the cost of same is equally prohibitive, leaving the ultimate goal to reduce the number of defective pixels/sub-pixels to near zero[1].
We are certainly not criticizing Leyard or other Micro-LED manufacturers about yield improvements, which have been significant over the last year, but investors should understand that while a defective segment of a memory IC can be ‘closed’ with the device being sold with lesser specs, in a visual device bad pixels are far more obvious and must be repaired, , keeping comparative costs high until yields approach 6 or 7 nines or other Micro-LED techniques are found. While Micro-LEDs an interesting and potentially lucrative display technology that will get increasing press coverage over the next few years, it will take some time before it is able to stand as the basis for a consumer oriented display product that is within the affordability of the average consumer. Progress yes, ready for Prime Time, not quite yet…
[1] A Micro-LED with only 1 bad sub-pixel would have a yield of 99.999996%
RSS Feed