We have been hearing that the world’s largest LCD panel producer, BOE (200725.CH) has begun planning a new Gen 10.5 LCD fab project.  While the company will not confirm anything official, BOE is already the global leader in large panel LCD production, and the addition of a 3rd Gen 10.5 LCD fab would certainly move them further ahead of their closest rival, fellow China countryman, Chinastar (pvt), who we expect will have a ~13.5% share of the global LCD market (capacity basis) against BOE’s 26.8% share at the end of this year.  While we have yet to see any documents that verify the new project, it would not be surprising to imagine BOE wanting to further capitalize on the large increases seen in LCD panel prices over the last year.
BOE’s last two Gen 10.5 fabs have a stated capacity of 120,000 sheets/month, and while speculation about the new fab is calling for 150,000 sheet/month capacity, we are, at least for now, adding 120k to our fab database, as duplicating an existing fab is always easier and less expensive than changing the configuration.  Timing is still a big question, as no announcements have been made, but we plan an announcement, if the project is approved by the board, in September, which would lead us to a phase 1 opening of June 2023 and a phase 2 opening of March 2024, with both ramping to full production over roughly 10 to 12 months. The ramp will begin to increase BOE’s global capacity share in 2023, and more so in 2024 and 2025, although much will depend on the timing and how quickly BOE can find customers to fill the new fab.  The table below compares our expectations for BOE’s gross LCD capacity share under current circumstances and with the inclusion of this new Gen 10.5 fb.

We note that BOE’s plans for its earlier Gen 10.5 LCD fabs were delayed from their original timeline, in part because of COVID-19, but also as a reflection of what were declining large panel LCD prices.  As those prices have increased, the expected profitability of new fabs, at current panel pricing, is very attractive and makes a strong case for government subsidies that now expect to see some sort of return, aside from job creation.  BOE’s management has stated that it believes the LCD display industry is exiting an era of cyclicality and becoming a more mature and more predictable industry.
While we agree with the fact that the LCD display space is maturing, we are not convinced that cyclicality is being washed out of the industry.  LCD displays will still have to compete with OLED and other potential display technologies, and while mini-LED backlighting and quantum dot enhancement can help to bring LCD specs closer to other display technologies, consumers want bigger and better TVs at lower prices and that means continued competition.  Unless the world has permanently changed due to COVID-19, there will always be periods of supply/demand imbalance, during which the profitability of LCD fabs will be marginal.  Building a Gen 10.5 fab does give BOE and others an efficiency advantage over older Gen 8.5 LCD fabs, especially with larger LCD panels (65” and over), but assuming that everything will remain the same as it is now years into the future is a dangerous game.

Yesterday we noted that El Salvador had become the first country to adopt bitcoin as a valid and accepted currency which adds bitcoin to the country’s official currency, which has been the US dollar since 2001.  As a primarily cash economy (only 30% of the population have bank accounts or access to financial services) this is an unusual place to start a bitcoin based financial system, especially one that is starting with relatively little regulation, but El Salvador’s President has run on a progressive ticket, and many say that bitcoin is about as progressive as you can get.
But it doesn’t end there, as we note that yesterday El Salvador’s President took things one step further and ordered the country’s state-owned geothermal power company LaGeo to develop a plan to provide facilities for bitcoin mining, which on the face of it seems a bit out of place from the companies that might normally build such facilities, which are essentially server farms.  That said, he also posted a short video of steam rising from a geothermal well, with the caption, “Our engineers have just dug a new geothermal well that will provide 95MW of 100% clean, zero-emission volcanic geothermal energy.  We start designing the entire bitcoin mining hub around that.”
As bitcoin mining consumes a great deal of energy, it has been criticized as a secondary polluter, increasing the need and use of fossil fuels to generate that power and adding to CO2 emissions.  Bitcoin mining supporters have been promoting the idea of using renewable energy for bitcoin mining to reduce its environmental impact, but this seems to set a new level of ‘wokeness’ to bitcoin.  China’s low-cost energy production has made it the bitcoin mining capital, but tiny El Salvador seems to be rising to the challenge and with a 930 mile active volcano subduction zone down much of the Western Central American isthmus, countries like Guatemala, Honduras, Nicaragua, Costa Rica, and Panama, might see this as a way to boost their economies and follow El Salvador’s lead.  

​When Dustin Hoffman, as Ratso Rizzo, slammed his hand on a taxi and yelled, “I’m Walkin Here!” while crossing against the light, he exemplified the attitude of New Yorkers in the late 1960’s, who felt the streets of the city were just as much theirs as for cars and trucks.  Crossing the street was an anywhere affair and even a split second space between moving vehicles was an opportunity to save a few seconds by crossing in the middle of the block.  While that attitude has never fully changed, the issue has transformed into a more rational one, where crossing at the corner seems to have finally caught on, but the smartphone has made even that more rational concept as or more deadly than before.  It is not uncommon to see those socially self-absorbed wanderers step off the curb into ongoing traffic without lifting their eyes off of their smartphone, often with AirPods or other devices stuck in their ears, lessening the impact of horns or traffic noises that might alert them to an oncoming vehicle. 
Trying to change the habits of those addicted to Twitter (TWTR) or Facebook (FB) is a moot point, so LG Electronics (066570.KS) has come up with a solution, an app called ‘Soft V2X’ that has now been modified from its original terminal based application to a smartphone app, and once the app is loaded, those city ramblers will be protected from the hordes of vehicles whose job is to get across an intersection just as the light changes.  The application transmits the users current location, direction, and speed to the real-time cloud, where the server side application selects information about vehicle movement, traffic signals, and other relevant metrics, and evaluates the risk of collision.  If that risk level reaches a set point, it warns the user with an overlay warning message, sound, and vibration.
The app is also able to detect the potential for collision between vehicles or pedestrians who do not have the app by accessing CCTV information and using AI to predict possible issues by making its own estimates of speed, direction, etc., and the system is also able to access RSU (Intelligent Roadside Units) information that is being added to traffic lights, light poles, and highway markers, in ‘smart’ cities, and combines camera and LIDAR information, making it an effective solution for traffic and pedestrian monitoring during adverse weather conditions or at night.  The LGE system also has a ‘child’ mode, which alerts the user when there are children pedestrians nearby or when the user is entering a school zone or near a stopped school bus.
While V2X was originally designed as a control system for vehicles, the adaptation to mobile devices opens that up to billions of potential users, and while LG Electronics does not have an actual commercialization date yet, the thought is that they will give the app to users for free in order to generate as much data as possible, with some tie-in to the proliferation of RSUs in cars and roadside that can communicate with V2X systems.  Expectations for V2X adoptions in vehicles differ greatly, but estimates for 2026 seem to be for ~40m vehicles equipped with some form of V2X, and with 5G slowly becoming available, the ability of those systems to communicate with little latency are becoming a reality.  No longer will you have to say, “I’m walkin’ here!” as your smartphone will say it for you and slow the car just enough to allow you to not miss that post on last night’s “The Bachelor” wardrobe mishap by being knocked over by a fast moving side view mirror.
We note that in order for a V2X-like system to operate effectively it must be in widespread use, which means it needs to be funded by both the government and private sources, and there needs to be an international strategy toward spectrum assigned to such systems.    The FCC has recently began to open the 5.9GHz band to V2X, which is a first step, but that needs to be globalized or the cost of implementation will remain burdensome.  As to the need for widespread use being a limiting factor, the example is, if 10% of cars are using the system, the chance of two equipped with V2X meeting would be ~1%.  At 25% the chance would be 6.25%, and if 50% of cars were equipped, the chance would be 25%.  Once the rate was above 70%, the chance of two cars with the system meeting would be over 50% (RATE2), so as is typical of many new CE product technologies, we have a chicken and egg scenario, with standards, cost, and funding the mitigating factors, although here, what could be more important than saving a life or getting the latest info on those 4” Aquazzura heels that just went on sale on ShopStyle (4755.JP)?

​After Taking what is expected to be a rather large jump in prices, DRAM and NAND Flash price increases are expected to be a bit more modest in 3Q and the 2nd half.  With DRAM prices up between 18% and 23% and NAND Flash prices up between 5% and 10% in 2Q, 3Q expectations look far less aggressive, but seem to be carrying a bit more risk.  With component shortages on the minds of memory buyers during 1H, buyers aggressively bid for memory to build inventory safe stocks in case the shortages seen in a number of IC products shifted to memory. 
With customer DRAM inventory levels higher than normal, expectations for 3Q DRAM pricing is for an increase of between 3% and 8%, a far cry from the previous quarter, but supplier inventories still remain low and demand for server memory typically increases in 3Q,, giving  producers the ability to maintain or increase DRAM prices.  NAND Flash is in a similar situation, with customer inventories higher and supplier inventories lower than normal, with wafer shortages and enterprise storage demand also increasing in 3Q.  This has pushed up 3Q NAND Flash price expectations from earlier expectations of a 3% to 5% increase to 5% to 10% in the September quarter.
That said, all is not set for smooth sailing in the memory space as the smartphone space is now seeing a slowdown as Chinese brands lower their full year unit volume targets, lowering overall smartphone growth from 9.4% earlier to 8.5% currently.  COVID-19 outbreaks in Southeast Asia and an overall slower end to the pandemic in China and India, are taking their toll on what we have noted previously as aggressive smartphone unit volume estimates and inventory levels at OEMs will have to adjust to those lower targets.  Despite the higher inventory levels at smartphone and PC OEMs, demand from other memory categories (see above) should offset the demand declines from a weaker smartphone outlook in 3Q, and the balance between negative and positive forces here still lean a bit toward the positive side.
That said, it would not take much to tip the scales in another direction, should the COVID-19 outbreaks prove more difficult to control, and this is true for many other parts of the CE supply chain as we also have noted previously.  While CE component sales remain high on a relative basis, as comparisons become harder, enthusiasm over the 1st half’s easy comparisons will wane a bit and the potential for declining q/q sales comparisons along with y/y comps, paints a higher risk scenario than was apparent in 1H, both for memory and the CE space in general.  We don’t want to seem overly negative but we certainly feel a higher risk scenario for the 2nd half is possible, and while OEMs in Taiwan are still showing reasonably strong results, the last two months have seen m/m sales turn negative for all but two assemblers and y/y sales comps for all but one OEM turn negative in May.  Short-term data but likely beginning to reflect Chinese smartphone target reductions.  More to follow.
 

​El Salvador is a tiny country, bit smaller than New Jersey and about the same population as Maryland, but with a GDP/capita of $8,388 (NY State GDP/capita is $87,866 for reference) it is not a rich country and ~70% of the country’s population does not have access to traditional financial services, such as a bank account or credit card.  It does have one defining characteristic however, it is now the first country to approve Bitcoin as a legal currency, with the parliament passing a bill (64 for, 24 against) yesterday that gives Bitcoin “…currency with free powers, unlimited in any transaction, and freely available to anyone, public or private”, according to the country’s President.  The official currency of El Salvador has been the US dollar since 2001.  The price of Bitcoin spiked on the news.
The country recently announced a partnership with Strike (), which went live in the country in April, is there to help set-up a fintech banking infrastructure, and given the lack of financial regulations, it is open to a wide variety of efforts to help the local community, or so the press reads, and is pioneering, not the conversion of bitcoin into local currency, but the conversion of all transactions into bitcoin.  This means paychecks can be issued in bitcoin and bills can be paid in bitcoin, with local businesses posting item pricing in bitcoin, and ATMs converting bitcoin into cash.
While this is a novel idea, albeit a bit unusual for an economy that receives 20% of its GDP from payments to citizens from relatives outside of the country, there are few regulations that both safeguard citizenry and help to regulate transactions.  While bitcoin is said to be the ultimate in transparency, given that all bitcoin transactions are permanently stored and but not identified by name, there are a vast number of ways in which bitcoin purchases, sales and transactions can be made ‘less transparent’ and disassociated from those who wish to use them for nefarious purposes.  Will El Salvador become a hub for those who have accumulated Bitcoin wealth ‘illegally’?  Likely not, but it will be interesting to see how this plays out in such a microcosm.  We are not sure how the population of El Salvador would have viewed the rise of Bitcoin from 29,004 on 12/31/20 to its peak of 65,587 in April, and its fall to 34,877 currently on their bank accounts, while the US Dollar index moved from 89.97 to 90.86 over the same period, but with 70% of the population without bank accounts, maybe it is the right place to try it out..

Quick Note: As far as indicators go in the CE space, LCD/OLED panel production metrics are among the best.  There are few CE items that do not have at least some form of display attached, and the main categories for panel production, large panel (10” and above), and small panel (smartphones, AR/VR, etc.) cover just about every device that has a display.  Smartphones alone shipped ~1.4b units last year and adding feature phones, the total comes to over 2b, and along with large panel displays, the industry generated over $150b in panel sales last year.  As we collect data from a myriad of sources, we are able to generate a picture of panel production and pricing, which tends to be a precursor, in a broad sense, for the CE space, given that displays tend to be the most costly item in CE product’s BOM. 
That said, panel production and pricing is not a perfect indicator but more incremental data in our quest for understanding the current and future state of the CE space.  On a long-term basis, panel production is far better at predicting or anticipating CE trends, but many of us have the pressure of making shorter term judgements, and so we present both short and long-term data whenever possible.  Panel producers in Taiwan have an exceptional characteristic in that they are required to report monthly sales, which helps to spot trends that can be muted when reporting sales on a quarterly basis, which is why we present such data.
All three Taiwanese panel producers saw a m/m increase in panel sales, and given the weakness seen across much of the CE space in 1H last year, all three were up substantially on a y/y basis, as they have been for since roughly a year ago.  AU Optronics (AUOTY) saw the largest m/m increase of 8.5%, followed by Innolux (3481.TT), up 3.4%, and Hannstar (6116.TT), which is really a small panel producer, up 2.4%.  With large panel prices up between 3.5% and 5.8% in May, one could draw the conclusion that much of those sales increases were driven by panel price increases, but more likely they were driven by mix shifts and available capacity, which have a greater influence on monthly sales than panel prices, given that short-term production contracts tend to lock in a price throughout the production run.
We do note that, while AUO no longer gives small panel and large panel shipment data, both Innolux and Hannstar saw a decline in small panel shipments, which was likely affected also by the flat small panel pricing seen in May.  The panel pricing trend for small panels has been far less volatile this year than large panel prices, up only 5% YTD, so it reflects a more accurate picture of demand for that segment, and we note that for Hannstar, whose primary business is small panel production, small panel shipments are down 4.4% YTD on a y/y basis, singling out the difference between stronger performance for large panel products and weak performance for small panel products.
Again, this is short-term data and limited to Taiwan based panel producers, but the performance of small panel production and pricing is far more normal and becoming predictable once again, while large panel shipments, pricing, and sales are still reflecting the changes to consumer habits as a result of the COVID-19 pandemic.  While large and small panel categories have different characteristics, those who say that the current trends in large panel shipments, pricing, and sales are the new ‘normal’, should look to small panel metrics as to what they might look like once the global effects of the pandemic are assuaged. 

​OLED materials are drivers for companies like Universal Display (OLED), DS Neolux (213420.KS), Material Science (pvt), Jilin (688378.CH), and Idemitsu (5019.JP), along with divisions of a number of well-known global chemical companies, but there are many materials used in the OLED stack, and every OLED producer has its own recipe as to what and how much of these materials it uses, with those recipe’s changing regularly.  While all the materials in an OLED stack are important, the emissive materials are the key to the actual operation of the device, with all of the other materials acting as ‘supporters’ of the emitters, giving them what might be considered ‘better working conditions’, allowing them to operate at their best.
There is some question about the materials that are included in OLED material sales data, and without the dataset, it is difficult to make sure they are what they say they are, especially when translations are concerned, but we present the data below as a reference point for better understanding the companies involved and the growth of the OLED industry itself.
UBI expects the OLED emissive materials market to expand from $1.75b this year to $2.25b in 2025, or 29.1%, with y/y growth in 2022 through 2025 of 2.7%, 9.1%, 7.9%, and 6.8% respectively.  While China will certainly increase their share of these materials, with BOE (200725.CH) leading the way, Samsung Display (pvt) and LG Display (LPL) combined will still be the largest purchasers, and as a country will purchase over 70% of total sales..  UBI goes further in saying that RGB OLED materials, essentially those used in all OLED displays other than OLED TVs will make up 78% of total OLED emission material purchases in 2025, down from 81.4% this year, while WRGB OLED emission materials, those used in OLED TVs, will increase from 18.3% currently, to 20.4% in 2025.
Again, without knowing exactly what materials and categories of materials are contained in the numbers, it is difficult to match them up with actual company results in order to derive the amount of actual emissive material sold and the portion of the total that are other stack materials.  Universal Display, who through control of a considerable amount of OLED IP, owns 2/3 of the direct emission material RGB market (red and green) and half of the direct emission WOLED market yellow/green), is expected to generate over $300m in material sales this year.  Based on our estimate of UDC’s share and UBI’s total sales estimate, UDC should see growth of at least ~29% in true emissive material sales during the period between 2021 and 2025, which we believe understates the potential growth for true emissive materials under the UBI scenario. 
These are back-of-the-envelope estimates, but we note that all the way back in 2014 UBI made some similar forecasts, projecting OLED emission materials to grow from $540m in 2015 to $2.5b in 2020, and while we almost always look at longer-term estimates with skepticism, we expect they might have understated their case currently after overstating it years ago.
As Chinese panel producers did with LCD, expanding capacity to the point at which they are the dominant supplier and low cost producer, they will likely continue to do with small panel OLED, and the fact that BOE is currently ramping two Gen 6 OLED fabs and building out a 3rd, points to that conclusion.  Korean producers, particularly Samsung Display, have the advantage of production experience, which gives them a lead in advanced OLED technologies, but they did have the same lead in LCD, and their exit from the large panel LCD space is a matter of record.  Nothing happens overnight and there are always bumps along the way, but the investment already made in OLED panel production will keep it as a viable and growing display technology for years to come.

Zhihui Jun is a smart guy.  He graduated from China’s University of Electronic Science and Technology in 2018 , did a stint at Oppo’s (pvt) AI lab, and then joined Huawei (pvt) as an AI algorithm engineer through their “Genius Youth Project”, and is working on “self-developed deep learning inference frameworks and Heterogeneous acceleration related technologies for mobile computing platforms.” He hopes to create an iron suit like the one worn by Robert Downy Jr. in the Marvel series, hence his personal title “Wild Iron Man.”  But like most folk, life is not perfect and a few months back, while riding his bicycle in a heavy rain, he fell and injured himself.
After the accident, he wondered how hard it might be to improve the stability of his bicycle and over a four month period, working on it on weekends, he developed a system that not only improved the stability of his bicycle, but turned it into an autonomous vehicle.  He developed a drive system, making plastic parts on a friend’s 3D printer (we all have those, right?), built a sensory network, including an RGBD depth camera, accelerometer, LIDAR, and gyroscope to detect surrounding conditions, and designed and built both the circuit board and the perception and control algorithms.  Using a CAD modeling tool, he added details to his simple bicycle, and had the metal parts tooled.  There is a small display on the control box to show various parameters to the user and the system runs (bicycle and electronics) on a model airplane battery that lasts for 2 – 3 hours. 
Once he installed the system on his bicycle, he used a gaming engine to make modifications and improvements by creating simulations, and set the bicycle loose.  The video of him describing the entire development process (in detail and in Chinese) has over 1.6m hits and he has opened the project up to students that have an interest through open source.  The only thing the bicycle cannot do is carry a passenger (yet), but its only a demo project.  The GIF of the bicycle on the move is here: https://www.ednchina.com/d/file/news/2021-06-09/0008471fdeae7e8a855e2e5d7d6ebb1b.gif
​

Foldable smartphones are a novelty, but as we noted yesterday, Samsung Display (pvt) and parent Samsung Electronics, wants them to be more.  Samsung has already stated that they expect to sell between 6m and 7m foldable units this year, up from last year’s 2.5m units, which was considerably below Samsung’s 2020 target of 4.5m units, and expectations are that at least one or two new models will be released in August, likely the Galaxy Fold 3 and/or the Galaxy Flip 3.  That said, with the Galaxy Fold 2 (released last September) originally selling for $2,000 (subsequently lowered to $1,800), and the Z Flip at $1,380, they remain among the highest priced smartphones available. 
Recent rumors seem to see that changing, and while the new Samsung Galaxy Foldable models were expected to have a number of new features, they have been expected to be priced the same as the previous models.  However recent ‘unidentified sources’ have indicated that the new models will be cheaper, despite the added features, with up to a 20% decrease in price from previous models.  While that would bring the price of the Z Fold 3 down to $1,600 and the Z Flip to $1,104, actual pricing will vary by country, along with a number of incentives for those that pre-order, when that option becomes available.
Even with the possible price reductions foldable smartphones are still expensive, although they do carry significant ‘look-at-me’ value, but in our note yesterday we indicated how Samsung Display was opening up its foldable production to other smartphone brands, which will certainly help to lower the production costs for foldable products, at least Samsung’s foldable products.  Because Samsung Display is becoming so dependent on its small panel OLED business, they must walk a fine line concerning margins so as not to imping on profitability, and we expect, while yields for foldables are likely lower than for flexible OLED displays, margins are comparable or higher, even with the lower yields.
If Samsung is going to eliminate the Galaxy Note line this year or next, which sold about 10m units/year at a rough average price of $1,000, they need to reach the top of their expected range for foldables to make up the difference this year, which is not an impossibility, but what we would consider a reach.  From Samsung Display’s perspective, selling foldables to outside customers is gravy since Samsung Electronics foldable demand will make up for the overall loss of Galaxy Note displays, and by next year, if Samsung and others are successful in maintaining foldable momentum, they will have built a new business around foldables that would continue to grow. 
That said, all is not without risk, as expectations grow for foldable price reductions.  If Samsung does not meet those expectations this year, we could see another unit volume shortfall, which could tarnish the foldable market and slow adoption.  As SDC has no choice but to forge ahead, we expect to see more foldable smartphone adaptations later this year and as we mentioned yesterday a number of initial models from brands just entering the foldable market.  That said, we do expect it to take at least another two years for foldable smartphones to reach parity with high-end flagship smartphones, as all brands are looking for the segment to boost margins, and SDC needs the same to balance its flexible OLED business, so what will push down foldable prices?  Chinese competition, which to date has been modest, more from a delivery standpoint than a quality one, but that will continue to improve over time with BOE (200725.CH), Visionox (002387.CH) and Tianma (000050.CH) all vying to develop viable foldables.  Samsung Display has the lead but China is certainly a viable contender.

​11/11 is a well-known ‘shopping holiday’ in China, while 6/18 is less known outside of China.  That said, while the ‘holiday’ does not officially start until the 18th, promotions, particularly on-line promotions, have already begun and will continue through June 20.  Alibaba’s (BABA) T-Mall, which has 250,000 brands participating in this year’s holiday promotions is already offering a 20 yuan ($3.13) discount on every 300 yuan purchase ($46.90), with a total of 10b yuan ($1.56b US) in coupons and subsidies so far this holiday.  While last year’s holiday followed a 76 day lockdown in China it was a path to growth for many brands after a dismal 1Q, so much focus has been placed on this year’s holiday as a way to gauge the sustainability of the Post COVID consumer recovery in China.
One issue that has already become a focus for the Chinese government’s anti-trust watchdog organizations are practices that damage the rights of consumers, terms not usually regarded as high priorities for the Chinese government.  Two major e-tailers have already been censured for listing counterfeit products and pressuring partner companies, and has been said to be part of the general crackdown in China on tech companies.   Because these shopping holidays are so competitive, many retailers place very high quotas on their managers, which tends to push them toward abusive practices, which the government says it will not tolerate this year.  Shares in the retailers that have already been called to task over their recent practices are down more than 40% from their recent peak.
One point that has been made leading up to the holiday is that TV sales in China, which are usually stimulated by discounts during such holidays, are not expected to be as strong as planned earlier in the year.  As the price of LCD TV panels has increased, set prices have also increased and some brands, particularly smaller ones, are reducing their expectations for the 4Q holiday season, as they are having trouble selling current product, even with discounts.  This seems to be a bigger problem in China than in the US and Western Europe, where sales seem to have been sustained by government stimulus and the lessening effects of the pandemic, while China’s CE companies have insisted in building inventory more aggressively.  Whether this was based on a false sense of optimism earlier in the year or just a desire not to be caught off-guard as component shortages began to appear, if those inventories are not sold through during the holidays, procurement for the 4Q holiday period would be at risk.  There is still time for a better CE holiday in China and North America seems to still be holding up, but the risk level keeps increasing and flat TV unit volumes in China for the month and for the year leave a bit to be desired.