You are tired of all the extra charges imposed by your smartphone carrier, and you are ready to call and cancel the service.  Your new carrier, while they will port your mobile number over to a new account, is unable to give you a SIM card that fits your phone, so you will have to buy a new phone, which is a lot more expensive than absorbing the cost of the extra charges imposed by your existing carrier.  You hang up the phone and reconsider.  While this is not always the case, as some smartphones have multiple frequency availability, the problem also occurs when you have a good cell plan but a poor data plan.  In the US it is impossible to separate cell and data carriers as your smartphone would need to be able to accept two different SIM cards, and they don’t.
That said, dual SIM card phones are available and quite popular in China and India, and have appeared in some European countries, under the guise that one SIM is for personal use and the other for business use, but there is no rule that says they can’t be used separately for cell and data services. Or to allow you to switch carriers on a short-term basis when deals become available without buying new phones.  That said, current versions of dual SIM phones are unable to use both SIM cards for LTE service simultaneously, but that is expected to change with a series of new chips that will support the dual LTE function.
Alas, such smartphones are not available in the US, Korea, and Japan, and carriers are happy that such is the game, however Samsung Electronics (005930.KS) is expected to release a dual SIM version of the upcoming Galaxy S9 and Note 9, and rumors abound that the Apple (AAPL) iPhone 9 family will also have dual SIM capabilities and are expected to be available in most markets.  This will likely have the effect of giving customers more options and carriers will need to offer plans that make it easier for customers to pick services.  If the two smartphone giants take the plunge into dual SIM phones for these markets, it will also make it far easier to shift carriers and likely fees will be restructured to compensate.  We note that in the US the cost of running a mobile phone averages $35.62[1] (2015) and is the 6th highest worldwide with the Netherlands the highest at $41.47 and Bangladesh the lowest at $1.42, with China at $4.07 and the UK at $16.45.


[1] ITU "Measuring the Information Society Report 2015" Table 4.2

Shenzhen based Chinastar (pvt) has capped the roof of their new Gen 6 flexible OLED fab in Wuhan, China.  The construction on the fab began in March 2017, with the entire project expected to cost ~$5.4b, and is expected to be completed in 1H 2019, which is ahead of our forecast of 3Q 2019, although the company expects actual mass production to actually begin in 1H 2020.  The capacity of the fab is stated at 45,000 sheets/month, although we expect that represents the total LTPS backplane output, rather than OLED output, which we expect to be 15,000 sheets when the fab opens, with 30,000 sheets for LTPS backplanes.  We expect phase 2 to increase the LTPS backplane output to 45,000 sheets, and the OLED output to 22,500 sheets/month when completed, which we expect in late 2020 or early 2021.
China Star is also building a Gen 11 mixed use fab in Shenzhen which is expected to be able to produce both large panel LCD and OLED displays.  That fab was capped last month and is expected to be in operation toward the end of this year, with commercial production levels in 1Q 2019, although we would expect that much of the initial output of that fab would be for LCD large panel production.  The total capacity of this fab is expected to be 90,000 sheets/month, although we would expect the split between LCD and OLED to be ~70%/30%.  Using those percentages, we would expect the initial production capacity for phase 1 of the project to be ~7,500 OLED sheets/month.  This fab has been constructed to be most efficient for large size TV panels, with six 75” panels being able to be cut from a single sheet at a 94% efficiency ratio.  This means only 6% of the glass is wasted, as opposed to a Gen 8.5 fab, where only two 75” TV panels could be cut from a single sheet, yielding a very poor 56% efficiency.
As we have noted in earlier notes, Chinese panel producers have latched on to the concept of mixed mode large panel production and China Star has designed this fab to be able to mix two panel sizes on the same sheet, in this case 43” and 65” panels, which greatly increases efficiency.  Since this has not been done on Gen 11 by any other panel producer, we did the calculation, which yielded five 65” and eight 43” panels on one sheet of Gen 11 glass, with a nearly 98% efficiency, which would seem a profitable trade off against a slower process time for multi-mode fabs.
We note that the completion of construction of a fab, and the fab’s official opening do not necessarily mean that the fab is in actual production, other than being able to move a substrate across the line from start to finish.  In fact, the time needed to go from fab opening to actual mass production levels varies considerably between producers, with Samsung Display (pvt), having the most OLED experience, likely the shortest, and those, like China Star, with little or no commercial flexible OLED experience, the longest.  In this case, China Star seems to have made a realistic timeline when estimating full mass production, which is unusual in the industry.
We have included three charts below, the first of which indicates Chinastar’s increasing raw OLED capacity, but it gives little indication as to whether that capacity will be used for small panel production, such as smartphone panels or tablets, or large panel production, for OLED TV and signage.  Figure 3 breaks that capacity down into its size components, which indicates Chinastar’s desire to become a supplier of OLED TV panels, competing with LG Display (LPL), currently the only OLED TV panel supplier.  Figure 4 however, and this is across the industry rather than for Chinastar only, indicates what we believe is a more realistic way to look at actual capacity, that of ‘available’ capacity.  As we have noted in the past, most capacity estimates in the display space are theoretical and represent a situation rarely seen in commercial production, that of 100% factory efficiency.  In order to meet the ‘stated’ capacity of a fab, it would have to be utilizing every piece of equipment at 100% efficiency (that means no maintenance downtime), operating 24/7 and if we include unit volumes, running at 100% yield.   All fabs have equipment that needs to be cleaned (particularly OLED deposition tools) and maintained, and all fabs take time to reach even much lower equipment utilization rates when they first start production, sometimes as long as 18 months.  Chart 4 estimates our ‘available’ capacity expectations for OLED production, compared to the stated capacity, taking into consideration a variety of factors that we believe make the available capacity numbers considerably closer to what the industry can actually produce.

​The OLED Patent License Agreement between Samsung Display and Universal Display (OLED) expired on December 31, 2017, as well as the material purchase agreement signed at the same time (August 22, 2011).  We had hoped that a new agreement would be signed before the expiration, however thus far that has not been the case, with the company issuing a carefully worded press release stating, “The Registrant and SDC are engaged in on-going discussions regarding a formal long-term extension of these agreements.  The Registrant expects the parties to continue to conduct business during these discussions. The Registrant intends to provide an update once a resolution to these discussions is completed. No time frame has been established for the completion of these discussions.”
The previous contracts, which were originally signed back in 2005, were updated in 2011 to reflect the needs of both parties, particularly as Samsung Display was actively producing OLED displays using UDC materials.  The license agreement called for a yearly license fee (paid bi-annually) in lieu of a per unit royalty, and escalated each year through 2017, allowing SDC to manufacture licensed products using UDC’s patented methods and processes, and giving access to all future patents during the contract period, with certain exceptions.  It did not give SDC, or its affiliates the right to manufacture the phosphorescent materials needed for such devices and prohibits SDC or its affiliates from participating in any challenge to UDC patents.  The license fees (below) were set based on both parties expectations for the growth of the OLED display segment and Samsung Display’s growth within it, and while some question whether such a non-unit based fee was correct, we believe that at the time, it was, and provided UDC with a guaranteed income, regardless of the ebb and flow of Samsung Display’s demand, and gave SDC the ability to expand its small panel OLED business without a per unit royalty to burden cost.
The supply agreement, which covered UDC red and green phosphorescent organometallic emitters, included certain minimum purchase requirements, but did not cover any other UDC materials, including a blue phosphorescent emitter, should it have been offered during the contract period.  We believe SDC satisfied the minimum amounts every year that the contract was in effect, although we believe that UDC would have worked with SDC to allow a push of those minimums into the following year if necessary.  What was most important about the material supply agreements was how the price of each material was determined. 
Given a particular formulation of an emitter, UDC set a base price per kilogram and shipped to SDC on a purchase order basis.  A cumulative total of purchase value of that material was kept, and the price per kilo dropped as each successive cumulative volume point was reached until a ‘terminal’ value at which new purchases would not affect the price.  This process was applied to each material, not just separately to red or green phosphorescent materials, but to each successive new formulation that reset the discount bar back up to the initial price when SDC began its use.  This encouraged UDC to continue to improve materials to maintain its already high margins, and while one might assume it kept SDC from adopting those new materials, as the price per kilo would rise, most of those material usage decisions made by SDC were based on improving the performance of the displays, which also included a number of other factors that were more important to timing than the cost of the materials.
Moving forward, UDC and SDC must now work through not only the details of a new long-term contract that faces some of the same issues indicated above, but a number of new factors are now present, some of which we present here. 
BLUE – A blue phosphorescent emitter is now closer to reality than when the updated contracts were signed, and the inclusion of a third[1] primary phosphorescent material in the contract would need to be accounted for in the license fee arrangement.
GROWTH – When the updated contracts were signed, the OLED space was in its infancy, and both parties had to rely on estimates and forecasts that included considerable speculation as to the growth of the industry.  Today there are far more indicators that can help to guide such estimates and should the new contract be based on that growth, would change the yearly license fee escalators.
MATERIAL USAGE – Samsung Display continues to grow its OLED business by adding capacity and consequently their material usage requirements continue to grow.  UDC must determine new ‘break points’ for cumulative material usage discounts based on current and planned future capacity expansion without pushing SDC to explore alternative materials.
PATENT EXPIRATION – Some of UDC’s basic patents are expiring this year and while we expect it will not change the overall IP protection that UDC garners, it has to be a point of negotiation for SDC.   We believe there will be some attempts by other suppliers to produce phosphorescent emitter materials over the next few years as some of the early material IP patents expire, but we expect all will be challenged in various courts worldwide by UDC.  While the outcome of such litigation is extremely complex, subject to a variety of nationalistic biases, and likely to continue for years, it is a point that Samsung must consider given its vast resources in specialty chemicals.  The previous contract was quite specific about UDC being the only supplier of phosphorescent materials, but if a supplier can levy a successful challenge to UDC’s expiring IP, Samsung will not want to be prohibited from using an alternative producer.  This would not extent to the ‘device’ patents that specify the use of phosphorescent materials in a device, which would still be covered under UDC’s portfolio, but could give Samsung some leeway in the negotiations.
ALTERNATIVE MATERIALS – While some of this is covered in the IP section, most of the emitter materials produced by UDC are based on organic molecules attached to a heavy metal, primarily Iridium.  In most instances the UDC patents also cover platinum and Osmium based organic emitters, but much research is being done to find emitters based on other metals or polymers, which would not be covered by UDC material IP.  TADF’s[2] have been a focus of a number of well-funded companies as an alternative to UDC’s phosphorescent emitters and SDC would have to consider the potential for these materials to be substituted for UDC phosphorescent emitters in some products.
We have barely scratched the surface of the potential points of negotiations between UDC and SDC, and given the complexity and constantly changing OLED landscape, we had relatively low expectations for a new contract before the end of 2017.  That said, until negotiations are finalized formally, both parties are working under limited contractual protections, and if the format of the new agreement remains the same, a license payment from SDC would be due in 2Q.  Hopefully these issues will be resolved before then, but we expect little change in SDC’s material buying patterns in the interim.


[1] UDC actually provides a number of phosphorescent emitter colors, including a yellow/green used for OLED TV, but blue material was outside of that group.

[2] Thermally Activated Delayed Fluorescence

​LG Display (LPL) set the tone for the New Year, announcing the world’s first 88” 8K OLED TV, which will be shown at the Consumer Electronics Show next week.  The company expects to offer the new set ‘soon’, adding to the OLED product portfolio.  The price of the unit has not been announced but LG Electronics (066570.KS) largest OLED TV, the 77” (OLED77G6P), which comes with all the bells and whistles and sound by Harman Kardon, sells for a mere $19,999, although you can steal either the 77” (OLED77G7P) less tricked out model or the 77” wallpaper OLED TV (OLED77W7P) for the low, low price of $14,999, which would imply that the 88” 8K model might be a bit more expensive.
Other than for the express purpose of showing such a large, high resolution TV at a show, there is little reason to produce more than a few such TVs as 8K TV is barely a ‘thing’.  Actually the Japanese Public Broadcasting Network NHK has been sending out a bit of 8K material over the airwaves since August 2016, when it covered a few portions of the 2016 Rio Olympics in 8K, but as there were almost no 8K sets available for Japanese viewers, the broadcaster had to set up a number of ‘viewing stations’ at public locations for the public to see the 8K broadcast.  Sharp (6753.JP) was the first to actually produce an 8K TV (85”) which it showed at the 2012 CES show, and a number of other TV panel producers and brands have shown similar 8K TVs at more recent events, including the massive 110” Samsung 8K 3D TV shown at CES in 2015, but given the lack of 8K content, they are not big sellers, particularly as the world of consumer electronics is just getting used to 4K, a resolution that broadcasters are still struggling with.
We assume the idea here, other than attracting crowds at the show, is that they can.  Panel producers need to show that they are at the bleeding edge of technology to attract those who are early supporters and buyers of technology, regardless of relative cost or other factors.  However, 4K TVs have 8.3m pixels, each with one red, green, and blue sub-pixel, which makes for a lot of pixels, which need to be refreshed between 120 times and 240 times each second.  That means a lot of data has to be processed, and for a broadcaster it means a lot of data has to be sent to viewers, which means it takes up a lot of broadcasting bandwidth, and while these are questions that we have discussed many times for 4K TV, when you step up to 8K, the problems become far worse.
8K resolution is 7680 x 4320, which boosts the number of pixels on the screen from 8.3m to 33.2m (times 3 for subpixels), which means that instead of processing ~25m datapoints every 120th of a second, the set now has to process ~100m datapoints/second, and the broadcasters has to send 4 times the information to subscribers, which means he will need 4 times the bandwidth to do so.  As bandwidth costs money, and every bit of programming takes bandwidth, this means the broadcaster has to buy more bandwidth or reduce the number of simultaneous streams he can send at one time.  Not something that makes broadcasters happy (See table 1 for details).
Much of our discussion here is simplified and there are a number of compression techniques that can reduce the data rates needed for 4K and 8K, but if the idea is to improve the picture, using compression defeats the purpose as it degrades the quality of the signal.  Many in the industry say that the average consumer will not be able to see the difference, and this is probably correct, but if that is so, why go to a higher resolution in the first place?  This is an ongoing issue that the industry will contend with for years, but panel producers and TV brands are more than happy to sell consumers sets that are ‘ready for the future’, a bit before the future is visible.  There are standards for 4K and some for 8K, but they evolve as the industry struggles with the points above.  Regardless, you can buy the new LG 88” 8K OLED TV very soon and watch reruns of the Rio Olympics and a few movies, although we believe at least one TV brand will show an 8K TV that will allow you to watch four 4K sources at one time on the same screen…maybe there is something to the idea after all….

​Once again, the Korean technology press has changed their mind about whether LG Display will be supplying small panel OLED displays to Apple (AAPL) in 2018.  According to ‘industry sources’, ETNews is stating that LG Display ‘temporarily confirmed’ that it will be supplying Apple with flexible OLED displays for the new iPhone that is expected to be released later this year.  The details are quite thin although 15m to 16m units have been quoted and a size of 6.5” for the panel has also been rumored, but negotiations are said to be ongoing (quite likely as LG Display is still building out its E6 fab in Paju).
Based on a June 2018 opening of the new E6 fab, we calculate that if LGD dedicated the entire line to Apple, they could produce ~1,000 sheets in 1H and 31,500 sheets in 2H.  This would translate to 2.3m units in 1H and 7.3m in 2H at 100% yield.  While LG Display has been involved in producing small panel flexible OLED displays for a number of years, a new fab will likely take a bit of time to both get into full production and bring yields up to acceptable levels.  Based on a 70% average 1H yield, they would be able to produce 1.6m yielded units in 1H, and using an average 80% yield in 2H, they could produce 5.8m units in 2H, for a total of 7.4m units in 2018.  In these numbers we make quite a few assumptions, particularly those concerning LG Display’s ability to meet timelines, ramp production, and improve yield.  As we take a conservative view here, we struggle to see how LG Display will produce the aforementioned 15m to 16m units in 2018 if they are being produced only at the Paju facility.  If we forget all ramp up assumptions and start LGD’s production on June 1 at a full 15,000 sheets/month and 100% yield, they could produce a total of ~24m units in 2018, but that is only a calculation for perspective.
The same ‘industry sources’ mentioned above have indicated that Samsung Display (pvt) is expected to supply Apple with 180m to 200m small panel OLED displays this year as their primary supplier, and rightly so given few other suppliers are able to meet both Apple’s high volumes and technical specifications.  As has been discussed, Apple is certainly looking to broaden their display supply chain, but at the moment must rely on Samsung for its immediate production.  Other small panel OLED producers are obviously wooing Apple with demos, samples, and promises of high volume availability, but in 2017 we believe Samsung Display’s share of the small panel flexible OLED display market is still in the mid 90% range on a unit volume basis, and despite the constant press releases extolling the virtues of small panel flexible OLED displays from other vendors, we expect only a modest change of unit share in 2018.  Apple is inherently a difficult customer, as they usually push the limits of at least one technology with every new model year, which gives Samsung Display a considerable leg up in staying on Apple’s good side toward technology changes on the display front, while others are trying to gain production expertise and yield, before they have to deal with Apple’s exacting specifications.  2019 will likely be a far more fertile ground for other small panel flexible OLED display suppliers.

​Back in late December 2016, Foxconn (2354.TT) announced it had decided to build an $8.8b LCD production fab in Guangzhou, China.  The fab, which would be built under the auspices of Sharp, a Foxconn  affiliate, was to be completed in 2019 and was to focus on producing 8K ultra large TV panels.  Little has been said about the progress of the fab, and we noted that we had heard that negotiations with the local Guangzhou government had bogged down last September, likely delaying the start-up date.
While not directly related to the Foxconn fab, we note that Corning (GLW) Display Technology Guangzhou recently completed its registration process for active company status in the Zengcheng District of Guangzhou.  The Corning project, which we estimate will cost ~$300m will likely supply the new Foxconn/Sharp fab when it is completed, and will be constructed simultaneously with the new display fab.  While we don’t have fresh information as to the status of the Foxconn/Sharp fab, the notion that primary suppliers are registering local corporations as suppliers is a good sign.
Our October update on the fab plans, despite the site manager’s disagreement, stated that we had pushed back our expectations for the fabs start-up to the end of 2019, and we still believe that is a realistic expectation, given the size of the project shell.  If construction were to begin tomorrow, we would expect 14 to 16 months to completion and another 6 months for infrastructure, equipment delivery, and production tuning.  Chinese construction crews do take schedules quite seriously and if they are able to meet more aggressive goals, the original September target could be met, but for now, even with the encouraging signs that major suppliers are moving forward, we remain convinced that year-end 2019 is most likely.

Last week we noted that the Beijing Securities Regulatory Bureau had sent a request to Jia Yueting, the beleaguered founder of LeEco aka Leshi (300104.CH), to return to China to repay the debts he piled on the company before he moved to the US last March.  The request asked that this be done before the end of 2017.  Jia replied this morning via his WeChat account, stating “I have paid full attention to the notice issued by the regulator...but there is a lot of work that needs done (in the States), to guarantee volume production and to ensure the delivery of the FF91 (his luxury electric car which is to be released this year).  To better solve the debt problem, I have entrusted Gan Wei (his wife) and Jia Yuemin (his younger brother) to act as my sole agent to exercise shareholders’ rights and responsibilities, including asset disposals and other related work.,”
While the Beijing Securities authorities have no power over Jia when he is in the US, given the lack of extradition treaties between the two countries, they have little recourse in forcing his return.  Sunic China (pvt) paid $2.3b to gain a stake in LeEco’s subsidiaries, but the main trading vehicle of the LeEco empire, Leshi Internet Information and Technology has been under trading suspension since April pending a restructuring.  The China Securities Regulatory Commission has yet to start a full investigation of the issues behind LeEco’s problems due to the lack of recourse despite the paper losses sustained by the company’s 200,000 shareholders.