The simple answer to the question, “Is Apple (AAPL) getting ready for mini-LEDs?” is yes, but the real question is when.  Rumors that Apple was to release one or two mainstream mini-LED products at the end of last year proved to be untrue and additional rumors that micro-LED based AR glasses would be released with last year’s iPhone 12 family (produced in conjunction with Taiwan Semi (TSM)) also proved incorrect.  New speculation that Apple is reading a mini-LED iPad Pro for release this quarter have resurfaced and underlying some of this speculation is Apple’s ‘secret’ mini/micro-LED project that sits in a little marked facility in Taoyuan, Taiwan.
The ‘secret laboratory’ as it is called, houses the mini/micro development efforts of Apple, Ennostar (3714.TT), and AU Optronics (AUOTY), all of whom are working toward the commercialization of mini/micro-LED products for Apple (and themselves), particularly Ennostar, a holding company recently formed by the combination of Epistar (2448.TT) and Lextar (3698.TT), two of Taiwan’s LED producers and packagers, for the express purpose of building a mini-/micro-LED business that would supplement the compound semiconductor business.  Lextar had already been producing mini-LED products for a number of applications before the merger, so, aside from the direct development project with Apple, the combined entity was already up and running in the mini-LED space.
While speculation continues about Apple’s mini/micro-LED plans, we believe Apple’s decision to migrate at least some products to mini-LEDs fall into three categories.  First, “Does it enhance the product as it stands now?”  That’s the easy one, as in any LCD based display having more granular control over the backlight system would improve the contrast of the entire system, reducing halos and adjacent segment light leakage that causes blacks to become grays.  Second, “Does the cost of the enhancement justify the improvement?  This is a bit trickier because estimates as to whether the improvements will justify a higher price from the consumer perspective have to be made, and we find consumer surveys and similar test marketing to be far more biased toward the product than actual consumers whose money is on the line.
Third, ”If the current cost of the enhancement is higher than justified, is there a path to a lower cost?”   Here is where Apple shines, as the company has the vision and resources to look down the road to see if that justification is financially feasible in the future.  In fact, Apple has been working on the development of mini/micro LEDs for years and while they will rarely be the first out of the gate with new technology, when they see the development process progressing, they will adopt the technology, slowly at first, but with increasing speed as the benefits to the Apple base are recognized.
As an example, Apple first tested commercial OLED displays in 2015, two years before it released the first iPhone using an OLED display.  The 2015 OLED product was the Apple watch 1st Edition series, taking Apple until the November 2017 release of the iPhone X to move the technology into the iPhone line, and even then leaving the iPhone 8, released at the same time, with LCD screens.  In 2018 two of the three new iPhone models were OLED and it took until last year for Apple to transition all models of the iPhone (exception was the iPhone SE which remained an LCD display) to OLED.
There is one more question that Apple has to answer before it can really set a true timeline for mini/micro-LED products, “Is micro-LED a viable technology?”  Note we say ‘micro-LED’ here because mini-LED technology is based on existing LCD displays while micro-LEDs are not, which makes mini-LEDs an enhancement to LCD technology, giving it a more finite lifetime, while micro-LEDs are self-emissive and would represent a new technology (and infrastructure) that would compete with OLED, which is also a self-emissive technology.  Of course, Apple will spend money on the development of mini-LEDs, micro-LED, and OLED in order to make sure it does not mis an opportunity, but at some point Apple will have to decide whether micro-LEDs offer a real challenge to OLED (in a cost effective manner) and if so, channel more specific product development down that road. 
We do expect Apple to release a mini-LED product this year, whether it is a new iMac or a monitor (XDR series) is immaterial, as the question to ask is really how Apple views micro-LEDs.  Mini-LEDs are a logical step, essentially taking existing LCD, LED-based backlight technology and continuing the path of shrinking and adding LEDs for better contrast, but micro-LEDs could represent another path for the display space as LCD technology improvement become progressively harder to make.  IF Apple sees a path for micro-LEDs it will go a long way toward legitimizing the technology, even in these early stages, and as noted Apple has been spending money to that end, but Apple’s mini-LED products have little to do with Apple’s eventual decision toward micro-LEDs, so it is important for investors not to lump both together or assume that mini-LED adoption has anything to do with micro-LED adoption.  They are as opposite as the sun and the moon.

Taiwan based panel producer AU Optronics (AUOTY) has purchased a 7.7% stake in Sintrones Technology (6680.TT), a company that specializes in vehicle computing systems. Such customized systems can provide a variety of applications ranging from self-driving to vehicle repair systems.  These on-board systems manage a number of processes including wireless (LTE), ignition power management, backup battery monitoring, and are based on Intel (INTC) CPUs and graphics, and are also used in railroad and fleet applications.
AUO is expecting that the stake will allow the two companies to develop a tight link between the displays that AUO provides for vehicles and internal networking  for AIoV (Automotive Internet of Vehicles), with applications already developed for fleet condition management and networking from “…cloud to ground, software to hardware, and AI to vertical applications”.  While LG Display (LPL) and Japan Display (6740.JP) are typically the leaders in the automotive display market, AUO has made a name for itself by producing more specialized panels generally, and has found customers in similar niches that have become high margin drivers for the company.  Given expectations for a doubling of the automotive display market size between 2018 ($15b) and 2025 ($30b) and an increase in units from 165m to 350m during the same period, AUO is looking to develop a display product line that is a solution rather than just a display.
Ona more general basis, the automotive display market is a bit less like the consumer electronics market as the lead times for product evaluation and adoption are longer than the half or full year cycles seen in smartphones or similar CE products, but given the more custom nature of automotive displays, the margins are higher than generic panel applications.  While LCD displays have been more commonplace for most automotive applications, the ability to develop OLED displays that can be flexible enough to mold to unusual shapes presents a challenge to those who have less emphasis on OLED and more on LCD.  That said, one of the most important characteristics of in-vehicle displays is brightness, where LCD displays stand out and the use of mini-LEDs increases contrast, so we expect there will be no clear winner between the two technologies in the automotive space for a few years, giving both time to develop more customized products that fit the ‘connected car’ theme.

​Samsung has a micro-LED TV called “The Wall”, aptly named for its 110” diagonal size, or 8’ wide by 4.5’ high or over 5,170 in2 of surface area.  At ~$150,000, it is out of the price range for most, so Samsung is taking steps to reduce the price by changing the way the sets are manufactured.  When producing these sets, Samsung moved each red, green, and blue, micro-LED individually from a singulated wafer to the TV substrate.  For a 4K display, that would entail 8.29m pixels, each with three micro-LEDs (red, green, and blue), a slow process regardless of the technology used, making it costly, especially when such a large number of very small LEDs will result in some chips being damaged, further slowing the production process as they are tested and replaced individually.
Samsung is working toward combining the red, green, and blue micro-LEDs into a tri-color micro-LED chip that will be packaged as a single pixel, and would, in theory, reduce the number of die to be transferred from almost 25m to ~8.3m.  We assume that the tri-color packages have been tested before transfer and are KGD (Known Good Die), which would reduce the possibility of transferring non-functioning die.  There would still be non-working die when the transfer is completed but we expect the yield should be higher overall if there is not a substantial number of additional steps to combine and package the three die.
As we have noted previously, there are many ways in which micro-LED die can be transferred to a substrate, and there is no particular process that has been universally adopted by any micro-LED producers (see our 02/12/21 note for details), and it seems that Samsung is still trying various techniques to reduce the transfer bottleneck, increase yield, and lower overall production costs.  We believe Samsung has explored pick & place, array stamping, and magnetic self-assembly, but continues to explore a number of potential transfer and repair processes in order to make micro-LEDs a commercially viable technology in the future. 
 

Microsoft’s (MSFT) Windows is the dominant operating system for computers, at least at the PC level, but the Apple (AAPL) MacOS has been the ruler of 2nd place, albeit a fraction of Window’s leading share, but 2020 was an unusual year in the CE space, to say the least, and it seems that Google’s (GOOG) Chrome OS took on the Apple challenge last year and won.  Normally the share of the Chrome OS would be directly tied to the sale of Chromebooks, devices that are essentially gateways to the internet and the cloud, with relatively small storage and applications, and that was still the case last year, however because of the many programs that helped students (and others) to work remotely during the COVID-19 pandemic, the popularity of the less expensive Chromebooks increased.
Of course, that change in direction is touted by some as a warning shot to both Apple and Microsoft that their days of dominating the OS market are over and to quote Fatal Attraction, “I won’t be ignored”.  Speculation that application developers are taking note and suggestions that anyone who provides a product or service over the internet should allocate the same resources to Chrome as for MacOS and Windows seems a bit premature.  Without sounding condescending, while we would expect Chrome OS to continue to grow over time as generational changes filter through the CE space, last year was a very unusual year and while 2021 might continue that trend, at least for part of the year, the rate of change for Chrome OS will likely slow when the pandemic slows.  It might be a bit early to declare a major victory…JOHO

Well, not exactly a folding pen, but talk from the Samsung (005930.KS) supply chain seems to indicate that Samsung display has solved the problems associated with adding a pen feature to the next iteration of its foldable Galaxy series due out later this year.  The challenges associated with a foldable pen input device are considerable, as the embedded pen sensing component has to be able to flex in the same way as the screen without creases that could cause false input, while at the same time avoiding scratches and deformation from the pen itself.  The use of UTG (Ultra-thin glass) is likely the most logical solution that provides protection while maintaining encapsulation while the foldability is a more difficult fix.
That said, while the production of such a display will be difficult and Samsung will not confirm or deny the addition of the pen to the next foldable, it seems they have confirmed production of the components (display) for May and full device assembly beginning in July, which would infer a September or October release for the Fold 3, which is consistent with the Galaxy Fold (2019) and the Galaxy Fold 2 (2020), both of which were released during September.  The Galaxy Z Flip, Samsung’s other foldable device was released in February.  If such is true, it will increase the chance that the Galaxy Note series will see its end in the near future, as the pen was one of its differentiating features over the standard Galaxy smartphone series.

OLED taillights are a ‘thing’, and while they might be limited to high-end models currently, their ability to be shaped and programmable makes them an ideal candidate for such vehicles.  OLED taillights are very thin, especially when compared to bulb taillights, and will not need individual bulb replacements ever, since they have no bulbs.  That said, if a problem should arise the entire unit would likely be replaced, although display manufacturers tout their 20,000 hour lifetimes (54 years @1 hr./day average) and high brightness.  What makes OLED taillights even more attractive, despite their higher cost, is that each pixel is addressable, meaning that turn indicator and other indicator patterns can be programmed into the display system, making them individually customizable.
But there’s a new kid in taillight town and Hyundai Mobis (012330.KS) is using micro-LEDs to produce a combination brake light and rear indicator package for an unnamed European car manufacturer.  The micro-LEDs are packaged in a 5.5mm film that can be conformed to any shape or curve that according to Hyundai Mobis, ‘lowered the volume by 40%”, which we take to mean the package thickness when compared to high-brightness LEDs.  What makes this unique, aside from the use of micro-LEDs, is that typically brake lights were separate from indicator lights as the former need a higher brightness.  Based on the fact that each micro-LED is addressable and programmable in the Mobis display, the brightness of each function is controllable by turning on more LEDs, along with the same pixel-by-pixel functionality that is seen in OLED taillights.
Since full specs are not available yet, we cannot confirm the actual size of the LEDs used, which would confirm whether this was a true ‘micro-LED’ product, but at least the lights have passed reliability tests in the US and Europe thus far.  The lamps took two years to design and develop manufacturing and the Hyundai Mobis lamp segment head says the company will continue to accelerate the development of the technology.  There have been some taillight designs using small (mini?) LEDs, but none that have been able to combine the brake and turn functions in one product.

MmWave represents 5G at its finest.  High speed, low latency, and lots of bandwidth, however as 5G frequencies increase from low to mid to high, transmission becomes and increasing problem.  Higher frequencies are easily blocked by many things including trees, walls, and even some weather conditions, and the distance broadcast nodes are able to transmit also decreases as the frequency increases, increasing the number of base stations needed for coverage. That said, mmWave is where 5G will eventually go, as businesses and similar commercial users push the technology for higher speeds and more bandwidth.
Globally, mmWave is a mixed bag, with many countries focused on 5G low and mid bands to enable rapid coverage, but countries are beginning to make mmWave frequencies available to carriers, the first step in opening up mmWave to the general public.  While the US has been a leader in mmWave, primarily due to Verizon’s (VZ) early focus on 5G for business customers, some countries have also taken the mmWave initiative seriously and have begun that allocation process.  Based on current Data from GSA, these are the countries that have already allocated all or part of the mmWave 5G spectrum.  If a country is not on the list, they are either in the evaluation stage or have not yet begun planning.
 

• Australia – Ongoing Assignments
• Bhutan – Temporary Assignments
• Chile – Assigned
• Finland – Assigned
• France – Temporary Assignments
• Greece – Ongoing Assignments
• Hong Kong – Assigned
• Italy – Assigned
• Japan – Assigned
• Norway – Assigned
• Puerto Rico – Assigned
• South Korea – Assigned
• Taiwan – Assigned
• Thailand – Assigned
• Uruguay – Assigned
• US – Assigned

Universal Display (OLED) reported 4Q and full year results that were considerably above consensus.  4Q sales were $141.54m, up 20.9% q/q and up 37.8% y/y, against a consensus estimate of $109.3m, net income was $53.9m up 33.1% q/q and up 104% y/y, while EPS was $1.14, up 33.0% q/q and up 102.8% y/y, against a consensus estimate of $0.64.  Material sales, which drive the company’s business, were $62.54m, down 9.0% q/q but up 2.9% y/y, while license & royalty sales were $75.05m, up 68.5% q/q and up 98.5% y/y.  Guidance for 2021 (full-year) was for sales between $530m and $560m or a mid-point of $545m, which is right on the consensus estimates for 2021. 
Before we go further, we believe it is worth deconstructing the UDC numbers a bit to get a better idea of how the company worked through 4Q and what was an unusual year.  Since accounting rule ASC 606 was put into effect in 2018, rather than book license and royalty revenue when it is received, companies must recognize license and royalty on a pro rata basis, based on material sales and estimates of the total value of the contract over its life.  While this rule goes toward smoothing out license and royalty revenue that varied significantly on a quarterly basis under previous rules, it also requires companies to estimate the total contract value regularly and how price changes and sell-through progress agree with previous ASC 606 estimates.  If there are material changes to contract estimate variables, the ratio between material sales and license and royalty revenue gets ‘trued up’, which was the case with UDC this quarter.  Approximately $17m of additional revenue was recognized in 4Q as part of the ‘true-up’, which the company stated was based on how COVID-19 affected their business over the 2020 year.
A more visual way to see the result of this true-up is to look at the ratio between material sales and license and royalty income (Fig. 1).  Before ASC 606, License & Royalty revenue was sporadic, primarily a result of Samsung Display’s (pvt) bi-annual payments, different from all other L&R agreements.  Since ASC 606 (blue), the ratio has become far less volatile, averaging 1.925, 1.645, and 1.357 in 2018, 2019, and 2020, with the previous four quarters averaging 1.53.  The 4Q true-up caused the rationto drop to .83, meaning the 4Q examination indicated that the L&R revenue, as measured against expected material sales over the life of the contract, was too low, and hence the $17m of additional L&R revenue booked in the quarter.  If there were no true-up and the 1.53 ratio was held into 4Q, the 4Q sales would have been ~107m, slightly below consensus.  That said, UDC guided to a return to a 1:1.5 ratio for this year, barring any unusual circumstances that would cause another true-up.  Fig. 2 shows the actual revenue segments, along with the ratio trend line and the ratio itself (gray).
 

Material sales, however give far more insight into how UDC is performing, after taking into consideration the variability of customer ordering patterns, the red and green emitter trend lines seen in Figs. 3 & 4 are indicative of the expansion of the OLED display space, and 4Q results are close to both trend lines.  Based on the sales guidance mentioned above and the return to the 1.5 material/L&R mix, the implication would be for material sales to grow ~38% and L&R sales to grow by 14.6%. Again this is based on ASC 606 and a 1.5 ratio, but in terms of actual L&R payments, rather than recognized L&R, we would expect L&R to grow more quickly, due to the expansion of LG Display’s (LPL) OLED TV production schedule, which is expected to rise from ~4.5m in 2020 to between 7m and 8m this year.  As LG Display pays a unit price based royalty (Samsung pays a flat fee), the dollar value of the increasing number of OLED TV sets sales will rise quickly relative to material usage, as we expect LGD to continue to drive up its material usage efficiency, and with the increasing volumes, see a lower per kilogram material cost.
 

One point that should be noted is that while the differential between 1H and 2H material sales is normally a function of the 2H holiday season (2H share average is 57.7% over 5 years), the split in 2020 was affected by the COVID-19 outbreak, which caused 2Q, normally a sequentially better quarter than 1Q, to be one where companies were so unsure as to how the pandemic would affect consumers, that they were unusually cautious about production and material purchases.  While final TV set sales data is not complete, we expect the 1H/2H ratio to be above the norm, possibly a record for 2H share, so we note that while dealing with the pandemic has become the norm for many people and companies, UDC is expecting 2H to be better than 1H, although no ratio was given. 
While this seems quite logical, we caution investors that if the pandemic subsides, either due to a more careful society (not likely) or increasing immunity from vaccines, there is the possibility that a return to a more ‘external’ lifestyle could reduce the demand for TV set purchases.  This caveat is less of an influence for TVs however, as while North America saw overall strength in TV sales last year, less was seen in other regions, leading to the possibility that while North American TV set sales last year could have included pull-ins of this year’s TV set sales and could see slow growth under a ‘COVID-19 recovery’ scenario, China could offset that with stronger growth.  In either case, we expect UDC material sales, especially those to OLED TV manufacturers, will be greater in 2H again, but there are some scenarios where the 2nd half could see more moderate growth.
UDC’s regional and key customer breakdown also gives some insight into 4Q and full year results.  The regional breakdown shows that in 4Q sales to the company’s Korean customers increased markedly, with both Samsung Display and LG Display seeing increased sales.  We note that the $17m true-up mentioned above is embedded in the overall sales number and likely weighted toward at least one if not both South Korean customers, so 4Q in Fig. 7 might be a bit exaggerated and less reflective of absolute material sales to those customers.  China sales on an overall basis were down a bit in 4Q, but the drop in sales to BOE (200725.CH) (down 63% q/q and 67.8% y/y), while not totally out of the ordinary, was a bit disconcerting, although the $6.4m q/q drop in sales to BOE was offset by an increase in sales to other Chinese customers of ~$3m, the net effect, a $3.5m drop in overall China sales in 4Q.  While it is always concerning to see negative q/q growth in a top customer, variations in ordering patterns, especially from Chinese customers are commonplace and would only become a real concern if they extend for two or more quarters.  That said, as a supplier to Huawei (pvt), we expect it could take BOE a bit to fill the lack of Huawei volume with long-term customers.

​There were two other points of significance mentioned during the 4Q call, both worth mentioning.  First, UDC expects to see spending toward the development of the company’s OVJP (Organic Vapor Jet Printing) system increase this year as the spin-off created last July works toward developing an ‘alpha’ version of the tool by 2022. While we are less concerned about the spending associated with the development of this project, we share some of the concerns expressed on the call about whether the timeline for a commercialized OVJP tool could allow for alternative display process technologies (such as micro-LED) to gain a foothold in the industry and lessen the need for the OVJP tool.  We understand that such a tool will take time to develop, but we also note that technological development in the display space are moving quickly and while we expect micro-LED onr a commercialized mass production level will not be available before 2023, the resources behind this self-emissive technology are large and well financed.
The other point that is of concern are margins, specifically material margins.  UDC has indicated that it expects material margins to be in the range of 65% to 70% for the 2021 year.  This is lower than the previous year’s expectations of 70% to 75% for material margins, and while overall company GM’s will still be ~80%, we believe the cost of some of the materials used in the formulation of UDC’s products have been rising.  The price of Iridium, the metal that binds UDC’s organics together, has gone from $1,670/troy oz. on December 1, 2020 to $4,400 currently, and while the amounts are small that PPG (PPG), UDC’s material supplier would use, that is a very large increase. 
UDC also stated that as their customer base continues to grow, customers demand customized materials that have to be developed by UDC in relatively small quantities until qualified and accepted by the customer.  These ‘developmental’ materials are expensive to produce and as the expanding customer base requires more such projects, lower material margins are a result.  While this is an acceptable explanation for lower material margins, along with higher raw material costs, even with UDC’s IP lock on phosphorescent OLED emitters OLED panel producers are pressing all of their suppliers for lower prices, and we expect UDC is no exception.  UDC’s customers know there are many OLED material development projects underway, with some trying to develop alternative materials that do not have license or royalty fees attached, so we suspect there is also some material pricing pressure across the industry that could also lower UDC’s material margins in 2021.  We also note that based on what we know of UDC’s pricing scale, as their customers grow their OLED material capacity and consequently their OLED material consumption, it takes less time for customers to reach price reduction goals, and while the volume is certainly the most important part of that equation, it also means less time at the higher price points.
While we always have our concerns about some of the blind sighted optimism the display industry and some investors have toward OLED (the segment and the company), our concerns are, at least currently, overridden by the overall growth in the OLED space, which is fueled by greater penetration in existing display markets and expanding capacity, both of which tend to override the nuances we mention each quarter.  All in it was a good quarter for UDC and we expect the overall picture for OLED to continue to improve throughout the year, but we also note that technologies are shuffled around mare aggressively each year and every investor should keep their eyes wide open.
 

​In our note on 1/26/21, we mentioned that one possible reason that the Samsung Galaxy S21 Ultra’s display was being touted by Samsung as a new display that can emit ‘brighter light with less energy, which can improve the usage time of smartphones…’, was the potential use of new material developed by Universal Display (OLED) that could enhance the stability and brightness of OLED displays, while reducing their susceptibility to ‘burn-in’.  We made this supposition based on data in UDC IP filings and the language used by Samsung when describing the few details they gave as to the technology used in this new display.  While this was, and is, speculation on our part, we note that it might not be the only reason that the display in the S21 Ultra is ‘better’ than other Samsung smartphone displays, but we stick to the concept until we are proven wrong.
Solus Advanced Materials (336370.KS), a recent spin-off of South Korea’s Doosan Group (000150.KS), has been producing OLED materials, particularly Hole and Electron Blocking layer materials that keep electrons from passing through key emissive layers too quickly, reducing their output.  The Solus material called aETL has been supplied to Samsung Display (pvt) for the last eight years for use in its small panel OLED displays and is one of the materials in its M11 OLED stack, the most recent iteration of materials used in Galaxy smartphone displays.  Improvements in this material, which is exclusive to Samsung Display, could also been part of the improvement in the Galaxy S21 Ultra, but we expect that the Solus material while a contributor, is not the only factor in the S21 Ultra’s improved display.
That said, Solus also mentioned that they are expanding their supply of materials for OLED TVs, which some take to mean for Samsung Display’s QD/OLED project that is under development.   As the technology for Samsung Display’s QD/OLED project has now been approved by parent Samsung Electronics (005930.KS), with tacit timelines for commercial product slated for later this year or early 2022, we expect suppliers to begin to ramp production capabilities for new materials relatively soon.  How different these materials will have to be relative to what is being used currently for SDC’s small panel OLED displays or even LG Display’s (LPL) WOLED remains to be seen, but the approval of Samsung, as the likely largest buyer of QD/OLED panels, was necessary for the potential infrastructure to develop.

Component shortages are a big topic in the CE space currently, and we have noted a few examples of what we call ‘real-world’ examples as trade press has a tendency to embellish the story with eye-catching lead-ins or mis-leading headlines.  On area that has been cited a number of times as a point at which shortages are enough of a factor as to cause display producers to rein in their demand enthusiasm a bit is display drivers.  These ICs act as a point between the source of data to be shown on a display and the display itself, translating the data and commands into a format that the display understands.  There are display driver ICs for all kinds of displays, and while there is a certain amount of commonality between drivers for each type of display, display drivers tend to be customized according to both the specifications of the display and the desires of the CE application designer.
One of the largest DDIC producers is Magnachip (MX), a South Korean company listed on the NYSE.  To be clear, we are not promoting Magnachip[1], but because of where they sit in the CE supply chain, their results offer some insights into the current CE shortages.  Magnachip is the leader in OLED display drivers with a share close to 35%, and while overall smartphone shipments declined ~7% last year, OLED smartphone shipments and the number of new models using OLED increased as adoption of OLED displays continues in the smartphone market.


[1] We have no position in Magnachip, or any other public or private equities.  We receive no compensation from any companies for promoting their products, the companies themselves, or their particular industry.

​Magnachip’s 4Q OLED display driver revenue was up 18.9% q/q and up 9.6% y/y, reaching the highest sales in the company’s history, with overall 2020 OLED display sales up 6.5%.  Driving their growth was OLED growth and 5G growth, as 5G makes the possibility of high-frame displays something smartphone designers want and Magnachip can provide, with ~70% of MX’s sales coming from 5G drivers in 4Q (up from 40% in 3Q and 20% in 1H).  But while results were good in 4Q there were two parts of commentary that we saw as significant and the reason we mentioned MX in the first place.
First, Magnachip, which is a fabless producer for its OLED drivers, indicated that it expected to be unable to supply ~$10m of customer demand due to supply constraints at their outside foundry (12” 28nm).  They expect some of that excess demand (likely less than 50%) will carry over into 2Q, but given the short cycle nature of the smartphone business, more than half will disappear.  While it is hard to pin down how much of this shortfall at the foundry level is due to component shortages and how much is due to competition for fab production time, the result is the same.
To add a bit of color, there are many fabless companies that have been double ordering in an attempt to meet customer demand, which has caused both foundry prices to rise and further tightening of foundry resources, especially at the 28nm level.  This is particularly prevalent with generic parts, as opposed to the more customized ICs that are used in OLED display drivers, and Magnachip was quick to point out that their OLED driver customers, who are primarily South Korean panel producers, have not been doing so with them, giving them an accurate picture of hard demand.  If this is the case, it makes what would be a ~15% OLED display driver ‘shortfall’ in 1Q a real measure of what the impact of IC foundry shortages can be.  On the positive side, the company also indicated that it expects the foundry limitations will begin to ease toward the end of 1Q.
One unrelated point that also came to light during the Magnachip call had to do with foundry blackouts, a topic that seems to be quite visible recently.  While Magnachip is fabless for their OLED driver products, they do have two internal fabs in Korea that are used for the production of power semiconductors.  Last August, a crane operated by a city contractor damaged power lines, causing a blackout at one of the MX fabs that lasted for 6 minutes.  While outages such as this tend to be focused on the cost in equipment damage, the real cost is time, given that much of the silicon in process when the outage occurs has to be scrapped and each process tool has to be restarted and run without commercial production until they meet pre-blackout specs. 
There is much speculation about how long such a process can take, with company’s stating anything from ‘days to months’, but Magnachip was quite open about their the fact that the fab that experienced the 6 minute blackout ‘was not completely healed until October’, what we believe was a realistic assessment of recovery time.  Based on that comment and using the middle of October as a target date, the fab took ~70 days to regain commercial production status.  Every fab would be different, however this does put some parameters around such issues and we thank Magnachip for their transparency.