As a reminder, we look a bit more carefully at panel shipments, prices, and sales as they tend to be a relatively quick moving indicator of the CE space.  The bulk of CE items have some sort of display attached or as the mainstay of the device, so display panel shipments tend to be leading indicators.  Listing rules in Taiwan require companies to report monthly sales, which gives greater granularity as to how quarterly results are progressing, and as such, we track monthly results for the three Taiwanese panel producers, AU Optronics (AUOTY), Innolux (3481.TT), and Hannstar (6116.TT).  While both AUO and Innolux produce both small and large panel displays, Hannstar is primarily a small panel producer, and while they do produce a small number of large panels, the contribution to total sales is small.
Given the unusual circumstances seen in 2020, due to COVID-19, y/y comparisons will look a bit more out of line that might be normal for January, and will likely remain so through 1Q when the virus was severely limiting production.  As 2020 progressed, those limitations decreased and demand from a sequestered global population increased, and panel producers began to see high utilization rates as 1H progressed.  By 3Q capacity issues returned, but this time not because of staffing limitations, but due to a lack of capacity, which caused some panel producers who had previously decided to end large panel production, to postpone those plans.  While 4Q 2020 was a strong one, new issues faced panel producers as some components could not meet panel demand, and as such continues into 2021, panel producers have been able to pick and choose which customers they will supply and customers are therefore willing to pay higher panel prices for that guarantee.
January and February are typically the weakest months of the year for panel producers, as holiday devices have been built and the Chinese New Year holiday causes low display production output, however given the unusual nature of the past year, and the ongoing pandemic, we expect norms this year to be ‘un-normal’, with only AUO showing January m/m results close to the 5 year average.  The charts below give a visual understanding of how panel sales have progressed at the three Taiwanese panel producers indicated above and we summarize results in the table below.  We note that AU Optronics did not provide large and small panel shipment data for January.

​According to local Korean press, the last model releases from LG Electronics (066570.KS) are being sold through local carriers at substantial discounts as the fate of the company’s mobile division remains in the balance.  The LG Wing, released last October and one of LG’s more unusual smartphone configurations, sells for between $1,000 and $1,700 in the US, while it is said to be available in Korea for ~$360, and could see lower prices as additional subsidies are added to move inventory.  LG’s Velvet smartphone, released last May, is also said to be heavily discounted, with Korea’s most aggressive carrier offering the device for free if one signs up for their premium mobile plan and qualify for subsidies.  That phone originally sold for ~ $800.
As LG decides what they will do with the mobile division, which has been losing money for years, local carriers are assuming that current inventory will be the last batch of LG smartphones and are trying to move out whatever is left, but we expect that the final outcome will be the sale of LG’s manufacturing facilities, while it maintains ownership of core IP that will help its IoT and automotive businesses in the future.  Rumors concerning a number of potential buyers have been circulating for months and both a ‘calming’ note to employees from senior management and comments made during recent conference calls, seem to indicate that the company still sees value in the overall mobile assets but is considering all options.

LG Display (LPL) is expected to add an additional $750m to its $2.5b investment in its OLED display module assembly plant in Haiphong, Vietnam, making it the biggest direct foreign investor in the country’s 4th largest city.  LGD’s original investment in its module plant was made in 2016, with the new investment, which is expected to create 5,000 jobs (according to the state news agency) and generate ~$5m in fees and taxes.  The additional investment is expected to become income producing by May of this year.
While LGD is the largest investor in Haiphong, Samsung Electronics (005930.KS) has 6 factories in Vietnam, with two of those being their largest smartphone assembly plants.  Along with a $220m R&D center they are building currently, Samsung has invested over $17b in the country, making it the largest foreign investor in the country.  With the average gross salary in Vietnam ~$16,780/year and a similar cost of living, Vietnam represents a location that offers low-cost labor as an alternative to China, where labor and salary levels continue to rise.

​Kulicke & Soffa (KLIC), a provider of semiconductor process equipment and LED placement tools, has purchased Uniquarta (pvt) a small, private company for $26 in cash.  As we have noted previously KLIC has been making a name for itself with its Pixalux tool (See our note 01/06/2021) that is used to move LED die to mini-LED substrates.  KLIC had mentioned that they expected to come out with an even faster tool later this year, and while we originally thought they would possibly use similar technology to increase the die handling capabilities of the Pixalux tool, it seems they have taken the process one step further.
Uniqarta has developed a laser based process that can move much larger numbers of even smaller LEDs to a substrate, without damage.  The process would reduce the bottleneck caused by the need to transfer over 24 million LEDs in order to produce a 4K micro-LED display, and could be scaled down to adapt it to mini-LED transfer, a process that involves far fewer and larger LEDs for LED backlights.  While mini-LEDs are based on existing LED technology, micro-LEDs are self-emitting and are not LCD based, which is both a positive and a negative.  The positive is such devices would compete with other self-emissive displays, such as OLED, and should provide high quality displays, with the negative being that the display industry has no existing commercial micro-LED infrastructure.
The biggest problem for micro-LED displays is the need to move such vast numbers of LEDs in a cost effective way, while also being able to replace those LEDs that prove defective once transferred.  The LED industry has come up with a number of techniques for trying to make such a process financially viable, however the commercial application of same has yet to be accomplished.  Based on what we have seen of Uniqarta IP, their (now KLIC’s) IP is viable for mini-LED transfer and, could be adaptable to micro-LED applications as the scale of the process is developed.  KLIC expects to begin qualification of its ‘new’ transfer tool in September, so we should have a better idea then as to how far along the Uniqarta technology has moved and at what point KLIC will migrate from mini-LED applications to micro-LED.

While we do not have the breakdown of new 5G devices by band yet, we do have the growth data, which we present below.  The primary indicators we use to track 5G growth were all up, with vendor count up by 4.6%, device count up by 5.2%, and 5G smartphone count up by 5.8%.  In particular the number of laptops with 5G availability increased from 9 in December to 11 in January against 2 in January 2020, along with a 4.6% increase in CPE (Customer Premise Equipment), which now has reached 113 models.  While mobile devices are the mainstay of 5G, at least sub6 5G, FWA (Fixed Wireless Access) is where the growth in mmWave will be focused. 
In order for mobile 5G to become seamless and available to a household without the problems associated with interference and building materials such as plaster lath, users should have some sort of CPE device, in the same way that optical cable has a connection device (ONT) that connects the fiber to a distribution box or internal wireless network, so the growth of CPE devices helps to understand how 5G is being used, or at least, the possibility that it can be used in other than mobile situations, even with sub6 5G.  All in while we expect the ROC for 5G smartphones to continue to contract as the number of available 5G smartphones begins to approach the total number of new smartphones.  In 4Q in China, approximately 50.9% of new smartphone releases were 5G smartphones, up from 42.9% in 3Q.

China’s largest particle physics lab has issued a statement that denies it is working on the development of a time machine after a document that claimed the lab was working with Ruitai Technology Development (pvt), a subsidiary of Ruitai  Corp (002066.CH) a refractory company, began to be circulated late last week.  The lab, which is state-owned denied its involvement with any such project or the company.  The document, which was said to be a PowerPoint presentation ‘mistakenly’ created by a financing information platform and indicated that the project was able to prolong life by distorting space-time, and had allegedly been ‘blessed’ by a number of prominent Chinese physicists.  The plan was for Ruitai Development, which had been formed last December, to raise ~$200m for the project, which was valued at $838b in order to set up an experimental lab on a 16 acre plot within 7 months.  The project plan was said to be ‘conservatively valued’.
It seems that Ruitai also denies any involvement in the project and has severed ties with the financial institution that created the document, although one wonders what their relationship was intended to be before the document became circulated, but it seems that there will be no Chinese time machine in the near future, at least not one that we will have access to, although the US has already developed one as seen in Hot Tub Time Machine 1 & 2.

As far back as July 2020, we noted that various parties were in discussion with Samsung Display concerning the purchase of their Gen 8 LCD fab in Suzhou, China.  The potential sale of the fab by SDC was part of its program to divest itself of large panel production assets, as the competition from Chinese panel producers had rendered that part of the display business for SDC unprofitable or marginal at best.  At the time TCL (000100.CH) was doing due diligence on the fab and the controlling entity, but at times other panel producers, particularly BOE (200725.CH) were also mentioned as possible suitors.  As the end of 2020 neared, it seemed that TCL, and its Chinastar (pvt) subsidiary were the winners in the Suzhou fab sale, however the end-of-year closing was delayed by South Korean government red tape, as a more detailed review of the technology to be transferred was requested and the closing was pushed into 1Q ‘21.
As of today, the South Korean government has given the sale its seal of approval, which will allow the transaction to continue, still likely before the end of 1Q, and while the profitability of the Suzhou fab has likely increased markedly over the past few months as large panel prices have increased, a deal is a deal.  The original deal structure was Chinastar would acquire a 60% stake in the holding company and a 100% stake in the fab for $1.08b US.   Of the remaining 40%, the Suchou government would retain its holdings (30% of total) and parent TCL would keep its 10%.  Samsung Display was said to be using $739m from the sale to purchase a 12% stake in Chinastar, although those parameters might have changed a bit.

It’s time to predict the future, in an environment that might be slightly less politically volatile but seems to be changing on an almost weekly basis.  That said, as we head into February and the Chinese New Year holiday, we note that while we expect a decline in panel shipments in 1Q ’21, there are a number of factors that must also be considered in understanding the CE picture for 1Q, especially in the display space.
Due to the particular situation that the world finds itself in due to COVID-19, particularly in the US, demand for notebooks, tablets, and monitors has been unusually strong since 2Q of 2020.  We expect that to continue through 1Q, and likely into 2Q, until restrictions such as limited indoor dining are reduced (Hopefully by May or June based on current expectations).
While we expect quarterly y/y comparisons to moderate as the year progresses, we see continued upward pressure on panel and CE device prices generally, less from demand and more from capacity limitations that lead to shortages.  We see February panel prices up again as foundries allocate resources toward large, profitable devices, and turn their focus away from less profitable but industry ‘essential’ products.  This lack of silicon capacity not only limits quantities, which in turn, limit end product production, but also creates situations where foundry customers move from contract level negotiations to bidding wars that exacerbate the price situation.
This follows a similar pattern in the display space, where large customers can ensure display quotas by increasing order sizes while smaller customers must pay higher prices and raise the price bar further.  As the shortages get worse, which we believe will be the case in February, we expect panel producers to face some limitations as to production levels.  We don’t expect this to mean that utilization rates will decrease substantially, as we expect that most panel producers will use what we expect could be lower shipments to large customers (shortage and seasonally related) to bring smaller customers, whose orders might have been backlogged due to large customer demand, up to quota.  We expect this ‘gap filling’ will keep panel producers active during February, which is usually the slowest month of the year.  In fact we know of a number of producers in China that are offering unusually high incentives and bonuses if workers stay on-site during the holidays.
All in, while shipments might decline, we expect panel prices to rise again in February, but more due to shortages than from demand.  We expect notebook panel prices to increase between 2.0% and 3.1%, while we expect monitor panel prices to rise between 2.6% and 3.0%.  TV panel pricing is a bit more convoluted, as we expect component shortages to continue to affect production, particularly on the glass substrate and driver side, while demand should weaken a bit this month.  Samsung Display’s (pvt) decision to keep its remaining large panel LCD fabs open for the remainder of the year helps capacity a bit, adding ~1m 55” TV equivalents back into the system this year, but that is less than 1% of total TV panel yearly production, so it is not a game changer, while substrate glass could see some allocation due to issues at NEG (5214.JP) and Asahi (5201.JP). 
This pushes us to expect TV panel prices to increase between 3.7% and 4.5% this month, despite the slower shipment growth, which would put aggregate TV panel prices 74.4% above the 2020 TV panel price low point, 75.6% above 2019 TV panel year-end price, and 8.3% above 2020 year-end TV panel prices.  While TV set pricing does not track panel pricing on a monthly or even quarterly basis, there is only so much component cost that set manufacturers can absorb, and adding in the rising cost of base metals and other components, TV set manufacturers will have little alternative as to raising set prices.  While many new TV sets have been announced thus far this year, most of those sets have not been released, nor have prices.  It should be relatively easy for set manufacturers to build higher component costs into new sets by justifying higher prices with new features or new technology, but if the panel and component price trends continue to increase, we would expect TV set prices to increase throughout the year.  There is a point at which this becomes visible to the average consumer, as it did with smartphones, and demand will weaken and TV panel prices will moderate, but at least for now (February), that is not the case.

​As we have previously noted, Samsung Electronics (005930.KS) is considering building a new semiconductor plant in the US.  The company has been shopping for potential sites in Arizona and New York, however a recent filings in Austin, Texas, where the company already has a fab, indicates that Samsung is looking for ~$805.5m in tax abatements over a 20 year period ($40.275m average) from Travis County and the City of Austin to develop a 640 acre plot that the company already owns.  Late last year Samsung purchased a 258 acre plot near its existing Austin location and filed development plans for the site.
As expectations are that construction is expected to begin in 2Q, Samsung will have to make a decision relatively soon if they want to keep to a timeline that will allow the fab to begin semiconductor production in 3Q 2023, so we expect similar negotiations are happening with local and state governments in the other possible locations.  The documents indicate that Samsung expects to hire 1,800 employees over the next ten years for the plant, with the necessity for being able to draw qualified engineers from higher education institutions nearby.  With the possibility of still building the fab in South Korea, we expect Samsung has a great deal of leverage in the negotiations with US cities, but if recent history has proven one thing, it is that ‘plans change’ as they did in Wisconsin where Foxconn (2354.TT) was planning to build a Gen 10 LCD fab, now severely scaled down to almost nothing.

​Composite consumer electronics data is hard to find, that is, not hard to find but harder to understand, as most do not define in detail what is included in such numbers.  This makes apples-to-apples comparisons difficult to say the least, but numbers for 2020 are beginning to appear so we present them individually, rather than as averages, which we can do with individual device data.  Strategy Analytics has calculated 2020 global consumer electronics revenue at $385.5b US, up 7.3% y/yy, inclusive of PCs, tablets, Set-top boxes, wireless speakers, TV sound bars, game consoles, digital camera, video streaming devices and, of course, ‘other’, which could include anything.
Shipments of ‘home computers’ (hard to define) and tablets rose ~11% to 396m units, while revenue for the same category increased by ~17% to $199b, while game console revenue increased 18% to $11.9b.  Wireless speakers (smart speakers) saw units decline by 3% to 240m units, and TV sales saw a decrease of 2%, so not every category benefited from the new stay-at-home paradigm, but at least from the perspective of expectations earlier in 2020, even the negative TV growth looked better than might have been expected.  Again, while we can collect and average unit volumes and sales for certain CE categories, composite data varies considerably between sources, so we cannot verify all of the specifics.  That said, the composite data does not reflect how unusual 2020 was from a historic perspective, which is why we qualify the data when it is presented.  As we aggregate or generate more specific segment data we should be able to present a better picture of each segment going forward.