Last November we noted that a fire at a Unimicron (3037.TT) PCB factory in Taoyuan, Taiwan had caused enough damage as to limit the company’s supply.  As the 3rd largest producer of PDBs, the limitations on capacity, which was already tight, were heightened, particularly for more advanced FCCSP (Flip Chip – Chip Scale Packaging) products, which have an even more limited supply base.  Ironically, the Taoyuan Fire Department was called back to the same plant yesterday after an acetylene torch being used to dismantle equipment damaged in the November fire, caused another fire that brought in 42 vehicles and more than 100 FD personnel.
After the first fire, expectations were that the plant would resume full production sometime in the first quarter of this year, although that timeline seems to be in jeopardy currently, although the company has stated that they do not expect a change to the previous restoration timeline.  This Unimicron plant produces boards for Qualcomm’s (QCOM) 5G Application Processor, and is also part of Apple’s (AAPL) iPhone supply chain, so production constraints are a sensitive issue.  There are other FCCSP sources, such as Samsung Electro-Mechanics (009150.KS), Daeduck Electronics (353200.KS), and Korea Circuit (007810.KS) that can also produce same, but given how tight the market was going it, we expect buyers will have to pay up to garner supply to offset the Unimicron issue

​In early December we noted that Sony (SNE), the leader in the CIS (CMOS Image Sensors) market, lowered their fiscal 2021 expectations from -6.5% to -11.8% for the year ending this March.  While we believe some of the reduction was due to a general slowdown in the smartphone market, the bulk of the cut was due to the expected severe reduction in orders from the company’s 2nd largest customer, Huawei, as the US extended trade ban limited Huawei’s ability to source necessary smartphone components.  The #2 producer is Samsung (005930.KS) who has pushed hard against Sony for the last 18 months, especially in the smartphone market, where its ISOCELL image sensor line currently has a top end sensor with 108m pixels and 12 bit color depth, giving it an effective 12,000 x 9,000 resolution (company info.) and has been used in a number of high-end smartphones, primarily those by Samsung Electronics, Xiaomi (1810.HK), and Motorola (MSI), with models from RealMe (pvt) and Nokia (NOK) expected soon, with a multitude of smaller sensors sizes.  The two companies make up ~85% of the image sensor market.
To make matters worse, Sony reported a ~10% drop in CIS sales for its December quarter, citing slower smartphone sales and a lower-priced mix, but they also revised their full-year sales guidance (March 2021) up by ~6% for the year.  So what changed recently? Sony says it is a number of things, pointing to higher-than-expected image sensor sales to the smartphone market, which was a bit too general, so looked deeper and discovered that Sony’s sales to Huawei, which were suspended in September leading to the downward revision, have been resumed.  Whether the resumption was due to legal issues surrounding the US Department of Commerce ban on sales to Huawei (even by foreign companies that use US products during production), or Huawei’s resumption of ordering for its 2021 smartphone line (even at reduced volumes), it seems that when Huawei sneezes, Sony gets a cold, despite being the leader in the space.
We expect that Sony will be able to make up the expected lower volumes from Huawei in their 2022 fiscal (March 2022), but with Samsung constantly nipping at their heels, we expect Sony will have to offer some price incentives to new or growing customers, who certainly understand that Sony, while they might still be in a leadership position, needs to fill in some sales gaps, and as such will likely make it more of a negotiation rather than Sony setting hard prices.  If Sony continues to cite mix as an issue for weaker results, it would be a sign to us that such pricing pressure has already begun.

​On 12/15/20 we noted that a 5 hour power blackout in Takatsuki, Japan has placed Nippon Electric Glass (5214.JP), aka NEG in a potentially serious position.  The 5 hour blackout is said to have caused damage to between three and six of the company’s melters (six total at this site), which would reduce the company’s glass capacity by between 13% and 27% of the company’s capacity, depending on the extent of the damage, and while this represents less than 7% of worldwide display glass capacity, it would have an unusually large effect on LG Display (LPL), NEG’s largest customer.
At the time, we expected the production lines at NEG to be up within 90 days, or around the middle of March, although the company made no statements as to the extent of the damage or repair time after the event.  NEG has now stated that they expect the plant to be operational by the end of March, and will take a special charge of $61.7m to account for the repair cost.  That said, the company also stated that they will be adding glass capacity in China this year, with a project spend of ~$285m that will increase the company’s capacity in China by 50% and will add ~5% to the company’s overall capacity. 
No precise timetable was given for the project, but given the expansion of large panel LCD fabs in China, the benefits of such additional capacity would be most felt if available by the end of August, which would be a very aggressive plan.  Given that there are only a small, number of glass producers that can guarantee sufficient material volumes for large panel fabs, NEG would be competing with Corning (GLW) and Asahi (5201.JP), with Corning just having recently announced expansion plans for its lines in Wuhan and Guangzhou.  NEG expects to see a 7% increase in sales and a net profit increase of 1.6% this year and a 2% q/q net profit increase in 1Q as a result of increasing demand.  The loss of the NEG production capacity in 1Q and an explosion at Ashai’s Gumi glass plant (Gen 10.5 line) has already squeezed what was a tight display glass market and, as we noted last week, Corning is expecting to see flat glass pricing in 1Q, a very unusual circumstance, with the increased demand for IT products and large format TVs driving demand.

​We have criticized a number of Chinese suppliers and panel producers for aggrandizing their expectations for unit volumes, production capacity, or share, most recently Tianma’s (000050.CH) statements that they will expand volumes by 5x this year.  In most cases as the year progresses there are the inevitable issues that reduce volumes or slow capacity expansion, with early targets rarely reaching those highly optimistic early predictions.
While it is still early in the year, and this being year two in the brave new world of COVID-19, it would be incumbent on most to step back a bit and perhaps pepper that optimism with a bit more reality.  That said, we note that Chinese OLED panel producer Visionox (002387.CH), still a relatively small player in the small panel OLED world when compared to leader Samsung Display, seems to have landed a bit of volume confirmation from some of its customers.  Not only did the company begin tweaking production at its new Gen 6 fab in Hefei just over a month ago, but it seems that smartphone brand Huawei has committed to purchasing over $850m in product this year (~5m units), more than double what it received (unaudited) from Huawei in 2020, with the estimates coming from Huawei rather than Visionox.
But it doesn’t end there, as reports that commitments from Honor (pvt), a former Huawei brand that was recently sold to a consortium of dealers and financiers, has also committed to purchasing up to 10m units from Visionox, despite it separation from Huawei.  Visionox, while small compared to SDC, has a reputation for being an innovator, with high refresh rate displays, the first commercial under-display solution, rollable displays, and high pixel density displays, a number of which are used in Chinese high-end smartphones.  Of course things could change radically at any point during the year and those numbers could easily be reduced, but it is refreshing to hear high expectations coming from customers rather than from producers for a change.  Reliability is still a question when we live in a rapidly changing environment, but at least there is something behind expectations rather than marketing and hyperbole.

​Despite the devastating pressure put on China’s Huawei (pvt) by the US government, Huawei is still releasing new smartphones and is expected to release its 3rd iteration of its Mate X foldable this month, which was first released in November 2019 and followed up with the Mate Xs in March of last year.  The new device is expected to have the same size foldable display (8” open), with a slightly smaller (6.45”, second screen, down from 6’6”) but at a slightly higher resolution.  As per the US restrictions, Google (GOOG) services are not available (nor were they on the Mate Xs), but the chipset has been upgraded from the previous version to the Kirin 9000, designed by Huawei subsidiary HiSilicon (pvt) and produced by Taiwan Semi (TSM) until more stringent trade restrictions were imposed last October.  We believe Huawei stockpiled Kirin 9000 inventory before the ban took effect, while the GPU is an upgraded (Mali G78) version of the previous GPU.
Huawei upped the camera complement, increasing the capacity of the main (wide angle) camera, but removed the TOF sensor seen in the previous models and added a 4th and 5th (selfie) camera in its place.  The battery is roughly the same, with a slightly larger charger, but the overall device is expected to be quite similar to the previous version other than the fact that the foldable displays, which had been previously supplied by both Samsung Display (pvt) and BOE (200725.CH), are now thought to be supplied singly by BOE.  We believe that this was not a choice based on quality or other technical factors but because Samsung Display was not granted a license to supply such displays to Huawei under the US Commerce Department’s updated rules regarding Huawei.
Lastly, the ‘big’ change between the previous versions of Huawei’s Mate X foldable series and the Mate X2 is the fold.  Not the fold itself but how the device folds.  Huawei had taken the path of having the two previous versions fold outward, while most other tablet-sized foldables folded inward.  The tradeoff between having access to the large screen when the device was closed had to be balanced against the increased chance that the outward folding display would be more easily scratched or damaged, with Huawei finally joining the ”In Crowd” (Dobie Gray – 1964).  The price of the original Mate X foldable was ~$2,600, With the Mate Xs a bit more expensive at $2,750 and while the price of the Mate X2 has not been announced, it is expected to be about the same as the previous version.  All in, not much to get excited about.

Last November we took a survey of 4K TVs that were being offered on Best Buy’s (BBY) website.  We collected data concerning brand, model year, screen size, and price, and have updated those prices a number of times since that point, usually at key points in the 2020 and now 2021 ‘holiday’ cycle.  Originally (11/17/2020), there were 68 4K TV models being shown on the site, which we use as our base.  Here’s what we found:

• While non-OLED models have come and gone, all six OLED models remain available and have throughout the survey period.
• Of the 68 sets advertised, 36 were unavailable, either because they were sold out or they were no available within 500 miles of our location.  In the current survey 32 are unavailable or sold out, with 18 of those models remaining unavailable since day one.
• In the original survey, 5 models were from 2018, 20 models were from 2019, and 36 models were from 2020, with 7 models having no indication of model year.  No change currently.
• LG (066570.KS) was the brand most represented in the survey, with 47.1% of units, followed by Sony (SNE), Hisense (600060.CH), and Toshiba (6502.JP), all with 10.3% share.  Samsung saw a mere 7.4% share, with Sharp (6753.JP) the lowest at 1.5%.
• LG and Sony represented all but one of the highest priced sets, with Westinghouse (VIAC), Toshiba, and HiSense much of the bottom tier.
• Of those sets that were unavailable, 30% were LG, 22% Toshiba, 11% Sony, HiSense, and Samsung, with 8% TCL and 4% Vizio (pvt), however almost all of the Toshiba sets that were unavailable at the survey onset remained so throughout.
• Currently, of the 36 models available, 19 (52.8%) were equal to or at their lowest price point since the first survey, while 14 (38.9%) were equal to or at their highest price point, with 3 (8.3%) somewhere between.

Here’s where the data needs to be looked at more carefully, as statistically, as noted above those sets at their lowest price and those at their highest are not so far apart as to being able to draw a definitive pricing conclusion.  However when the data is looked at more carefully, the nuance is revealed.  Looking at the 5 highest priced models, 4 out of 5 at their lowest survey price point, while all 5 of the lowest priced models are at their highest survey price point.  One could draw the conclusion that the lowest tier sets are smaller (lowest tier average size 50.2”), and therefore less desired during a sports venue like the Super Bowl, while the top tier (highest tier average size 74.6”) would be the most desirable, and therefore in demand, pointing to the fact that the ideal purchase time for buying large 4K TVs is right before the Super Bowl, rather than during the holidays, while the ideal time to buy smaller sets would be on or near Black Friday, especially given that the average spread between high and low price points in smaller sets is 16.5%, while the average spread between high and low price points for top tier sets is 11.3%.
While events, such as the Super Bowl, are good buy points for larger TVs, we note that new model announcements made at CES a few weeks back, will also have an effect on TV set prices.  While only a few of the ‘new’ sets are available (usually start showing up in March) and even less have been priced, as they appear next month or in April, those prices will set the tone for 2020 model discounting.  Next Friday begins the Chinese New Year holiday, which could have a small effect on prices of Chinese brands in the US, but we expect the largest influence to be the relative prices and feature sets of the new models from major brands sold in the US going forward.

​It is a bit unusual when the subsidiary of a large corporation goes in a direction not supported by the parent, but such has been the case, at least to a degree, between Samsung Electronics (005930.KS), the parent, and Samsung Display (pvt), the subsidiary.  As Samsung Electronics was the biggest customer of Samsung Display’s large panel LCD business, SDC’s desire to end all large panel LCD production pushed Samsung Electronics to expand both their OEM development programs and find large panel LCD capacity elsewhere, and the closing and/or refitting some of SDC’s large panel fabs helped to tighten the large panel market, even before COVID-19.  SDC’s program to develop a combination OLED/quantum dot technology as a progression from generic LCD also did not sit well with Samsung, with rumors that Samsung Electronics was not interested in using the technology.
That all seems to have changed as an agreement between Samsung Electronics and Samsung Display is said to have recently been made where Samsung Electronics will adopt the OLED/QD technology next year and SDC will continue to supply large panel LCD capacity to its parent through the end of this year.  There has been some speculation as to such when SDC decided to halt its large panel fab closure plans with no timeframe given, as panel prices continue to rise the odds have been leaning to a longer extension rather than a shorter one.
 SDC has agreed to sell its Suzhou, China large panel LCD fab to TCL (000100.CH), which was to close on January 1, but was extended to Match 31 as technology reviews by both governments have taken longer than expected, but will continue production at the remainder of its large panel fabs in Korea, with the exception of the L8-1 fab in Asan, which it has converted to OLED/QD production.  That project, which has been sampling product to its parent and outside customers since the end of last year, will be the SDC’s first production line for either OLED/QD or a variant using nanorods made of GaN and quantum dots.  While production would be relatively limited this year at a maximum capacity of 30,000 sheets/month when fully built-out, SDC has the option of converting other large panel capacity if needed, although much will depend on the cost of the process.
While there had been little indication that Samsung Electronics had changed its mind about using SDC’s new technology, we expect SDC’s alternative (nanorod) process might have made enough of a difference to convince parent Samsung to adopt the technology, although that could also be a function of the agreement to continue large panel LCD supply.  Either way, it stabilizes the next few quarters for Samsung Display, as they will continue to maintain a profitable large panel LCD production business while developing new technology that will likely provide little financial support.  Samsung Electronics will see less need for outside large panel sourcing, which would likely be a disappointment for other panel producers who might have moved up in the Samsung supply chain, and the overall industry’s capacity will be higher this year than originally expected, assuming SDC’s closings, which could help to slow upward large panel pricing pressure as the year develops.

We have mentioned rollable displays a number of times over the last year as smartphone brands look to both differentiate themselves and solve some of the problems faced by foldable devices.  By rolling the display on a spool, the way film (a material that formerly was used to record images, for those that are under 30) was wound in a camera, the pressure on an OLED display can be distributed across the display, rather than at the fold point, which has a tendency to cause a crease in the material and potentially break it down over time.  Springs, gears, and motors are usually cited as the methodology for scrolling and un-scrolling these devices, but since there are no such devices commercially available, there are no rules as to how this might be accomplished.
There were a number of such devices and displays ‘unveiled’ at CES this year, although most have either an indefinite timeframe for release, ranging from ‘soon’ to ‘?’, despite some specs that at least give some idea as to how large the display/device will be.  The Oppo (pvt) X2021 has a 6.7” display that extends to 7.4” and comes in 5 colors (whenever it comes out), and is powered by a Mediatek (2454.TT) processor.  LG introduced the “LG Rollable”, again with a screen size of 6.8” that extends to 7.4”, and while Samsung did not actually show a rollable smartphone to the public at CES, they have been engaged in a rollable development program for years and have shown mock-ups to customers.  The most recent Samsung iteration, which was detailed in IP filings, is a slider that pulls out in order to enlarge the careen.  Huawei (pvt) has also teased a number of times and many mock-ups have been ‘suggested’ by designers, although no real devices have surfaced.
Samsung is the only smartphone brand to have promised to have a smartphone sixed rollable device out this year, and is likely a step ahead of most OLED panel suppliers in developing the necessary production techniques given their experience in producing flexible OLED displays, but despite the fact that there is not market for small rollable displays yet, the competition has already emerged.  Aside from the names mentioned above, some of whom are both smartphone brands and display producers, China’s BOE (200725.CH) has been working to develop rollable OLED displays, and is rumored to be a partner with LG in development.  TCL, subsidiary Chinastar has also been developing same as has Tianma (000050.CH), who has been working on flexible OLED displays for a number of years.  While we don’t see much in the way of product this year, at least at a reasonable price, by mid- 2022 we expect to see actual product from a number of smartphone brands, and maybe one day we will get the fully rollable display that pulls out of a cylinder that we have been hoping will appear for many years.

As we noted on 12/23/20, Chinese panel producer Tianma (000050.CH) has set a very aggressive goal of producing 40m OLED panels this year, after shipping an estimated 1m panels in 2020.  It is not to say that Tianma is not expanding its production capabilities, it is, with two major construction projects, a phase 2 fab extension at its Gen 6 fab in Wuhan, and a new Gen 6 OLED fab in Xiamen, and a line conversion from rigid to flexible.  That said, according to a site in China, the Xiamen fab that will house the new Gen 6 line, a 120,000 ft2 structure, has been capped 50 days earlier than scheduled.  The construction project manager indicated that despite a number of both physical and material based obstacles, the job was completed ahead of schedule.
This is somewhat typical in the construction of the physical plant for display projects in China, and we give credit where due, as they tend to finish such construction projects on or before schedule, but rarely are the bottlenecks that cause problems at display fabs caused by the physical nature of the structure, but more so the ‘tuning’ of the line, which requires considerable expertise in the operation and coordination of a vast array of equipment, which can hold back a fab from profitability for many months, which can easily offset any timeline gains made during construction.  Again, we give credit to Tianma’s sub-contractor for staying ahead of schedule, but it means little to the aggressive production goals the company has set for itself this year.
Despite the early Xiamen fab capping, the fab is not expected to be in production until mid to late 2022.  Based on our calculations, if Tianma were to convert all of its rigid OLED capacity to flexible as of January 1 2021 (certainly not realistic), and produced at a 75% yield, they would just barely be able to hit 40m units.  Slightly more realistically, if they were able to convert the Wuhan fab to all flexible production on Jan. 1 of this year, they could just barely hit 20m flexible panels.  Since we expect the conversion to take far longer, we find it almost impossible for Tianma to reach such aggressive goals.  Of course, there is one way a high unit count could be produced and that is by producing smaller displays, such as watch size, as roughly 15 times the number of watch displays can be cut from the same size panel as standard 6” smartphone displays.  As we stated in our earlier note, we expect Tianma to produce between 7m and 8m units this year.  Given the potential for a slightly earlier production start, we could see an additional .5m units, but much would depend on equipment delivery schedules and yield.  We look at devices shipped to customers, not un-yielded production.

We have mentioned a number of times that prices for various materials and components related to the display space and CE products in general have been rising.  This ranges from basic materials such as copper or more esoteric but equally important metals such as Iridium (used in phosphorescent OLED materials) (see note 1/15/21), to components that are essential to many CE products, such as camera modules and lenses (see note 1/21/21).  As we move further into the New Year, we find more of such shortages, as the industry tries to deal with the increased demand generated by our COVID-19 lifestyle changes.
Polarizers are an absolutely essential part of LCD displays.  There is a polarizing film in front of the LED backlight which sets the orientation of the light to allow waves of only one direction (vertical or horizontal) to pass into the liquid crystal.  Since the liquid crystal can orient the light itself by ‘twisting’ its orientation, it can allow the light to pass through (pixel on) by aligning the direction the same way as the polarizer and can turn off the light by ‘twisting’ the light 90⁰.  A second polarizer in front of the liquid crystal, that is aligned the same way as the first, keeps any stray light from escaping when the liquid crystal is ‘off’.  Such a system would work without the polarizers but very poorly, with low contrast and a host of other problems.  The polarizers themselves are really composite films that also perform other functions but the structure of LCD displays is based on polarized light and therefore polarizers are essential.
The optical compensation film market is estimated at ~30m m2 per month, with FujiFilm (4901.JP), Konica Minolta (4902.JP), and Zeon (4205.JP) being the primary suppliers.  Konica produces about 11m m2/m while Zeon produces about 10m m2, with FujiFilm at about 6m m2/m.  This falls about 10% short of current demand, however FujiFilm has taken part of their production off line for maintenance, reducing their monthly output from 6m m2 to 3m m2, creating a 20% shortfall.  That line is expected to be restored in March but will not be up to full capacity until June, leaving the increased shortfall for much of the 1st half of 2021.  Since building capacity in such a short period is not an option, polarizer film producers have taken another approach and decided to raise polarizer prices by between 5% and 10% this quarter, adding a bit more to overall display BOM.  As we have noted previously, large panel module assemblers have high volume deals with such component suppliers, but smaller assemblers will have to accept the price increases or even bid higher to secure enough inventory to meet orders, which will give component producers the confidence to initiate another round of increases in 2Q if demand continues to outstrip supply.  Sort of the snake that eats its tail…