Given the fact that most CE devices have a display of some sort, and as the device category becomes more sophisticated the display tends to follow suit, and the display itself tends to be the single highest cost component in many CE devices.  As an example, the cost breakdown of components in the Samsung (005930.KS)  Galaxy Note Ultra 5G, one of Samsung’s flagship models that was released in August of 2020 had a component BOM of ~$468, and while the cost of the display included the 6.9” OLED screen, touch overlay, driver IC and cover glass, it was 19.6% of the component BOM, with only the entire cellular system, encompassing the 5G modem (both Sub6 and mmWave), 2G,3G, 4G systems and antenna modules costing more (20.8%), although we believe the display itself was still the largest share of the component BOM.  As such the display market is a key indicator as to both short-term and long-term trends in the consumer electronics space and we have looked at and maintain data display sales and shipments on a monthly basis since 2004 to help in understanding the prospects for a wide variety of companies that feed the CE supply chain.
As the industry contracts after hitting a peak in July of last year, we watch for signs that those factors that caused or are causing the decline are gaining or losing ground in order to gain some insight into when a bottom in the display cycle might be reached and what the momentum on the up cycle might look like.  The data, particularly charts, paint a picture that can give hints as to where the display industry will be in 3 months, six months, a year, or two years but the psychology behind what goes on in the display space is equally important and we try to provide an unbiased perspective to balance the data.  When we speak with industry professionals we try to understand their way of thinking, which can be diametrically opposed to our perspective, but our goal is to read through the corporate speak, technical gobbledygook, and blind optimism that pervades the industry and figure out the root of the conversation and real answers to questions.  We have no positive or negative bias toward the display space or the CE space as we get paid to provide information and not promote sponsored products, which we hope our readers understand and value.  So much for shameless self-promotion…
As we have noted, panel; producers finally began to realize that panel pricing was not going to go up forever and that what they hoped was going to be the ‘new normal’ was just another wave in a cyclical industry.  As we noted yesterday panel producers have begun to lower fab utilization rates for a number of reasons.  The first, demand has declined, and second, many panel prices have fallen below cash costs.  As demand rose in 2020 and the 1st half of 2021, as COVID changed consumer habits, brands embraced that demand but held on to consumer momentum far longer than necessary, and coupled with an underlying fear of not having enough components to meet future demand, brands built inventories past normal levels.  Now, with lower demand from consumers as COVID wanes and inflation rears its ugly head, those inventory levels are the third leg of the stool forcing producers to lower utilization levels to work down that inventory.
Panel producers began this process in May and have since lowered utilization rates from the low to mid 90% level to the mid 70% level and while in yesterday’s China LCD Production Utilization Chart we left July industry large panel sales flat, we now have added actual July data, which indicates that large panel sales in China decreased by 8.5% in July, giving some indication as to the effectiveness of the most recent utilization cuts in China.  As China fabs generated 47.3% of total industry large panel LCD sales, the utilization cuts made on the Mainland have a significant effect on the global display supply situation, hence our focus on China during this part of the down cycle.

On a global basis, as panel producers in other regions have also cut utilization, July global large panel sales were down 4.3% m/m and down 35.7% y/y to the lowest point since May 2020.  July large panel shipments were down 12.1% m/m and down 16.9% y/y, and as shipments declined at a faster rate than sales, overall large panel ASP’s rose by 8.9% to $81.73, the first m/m increase since February.   All product categories saw shipment declines as shown below.  If we average the sales/month for this year, the reduction in sales between June and July would be the equivalent of 1.2 days of sales, a rather mediocre push toward reducing panel inventory, which we believe to be ~15+ days above normal, with some of that sales reduction from weaker overall demand.  While we expect the additional cuts made this month will impact sales further in August and help the industry to absorb more than a day or two’s worth of inventory, our expectations for how long such utilization cuts might last are now extended into 4Q and beyond.

​[1] Includes Notebooks, Monitors & Tablets

Before we go further we need to quantify what we mean by ‘better’.  For the first time in many months actual panel pricing (in some categories) came in better than our forecasts, hence the ‘better’ in the title, but we also have to qualify that in noting that the aggregate large panel price is currently at its lowest point in the last three years and you would have to go back as far as February 2014 to find an aggregate price lower than current.  With that said, we look at panel pricing for August against that backdrop and while TV panel pricing came in within our forecasted range, IT panels, particularly notebooks and monitors, did a bit better than our expectations.  Again, qualifications are necessary as monitor and notebook panel pricing in August was still down 5.5% and down 4.6% respectively, but given that those are the smallest monthly declines since March/April, the glass is half full.
We expect that the utilization reductions have tightened the panel market to a degree, and panel producers are a bit more picky about the profitability of projects that they take on for customers as they have less capacity to fill, which pushes us to look at least a bit more favorably at pricing in September, and while we still expect the aggregate large panel price to decline between 2.9% and 3.5%, the momentum to the downside has lessened.  The question to us is whether there is a seasonal improvement to demand driving the ‘improvement’ or is this more of a drive by panel producers to avoid unprofitable jobs that they might have taken on in the past in order to maintain a good relationship with large customers.  We expect the power issues in China might also have a bit of psychological impact on buyers who still need to meet quotas, albeit lower ones, and are willing to negotiate a bit more to ensure volumes are met, but we expect things will cool down in China by the end of the month and buyers will remain in the driver’s seat.
One metric we watch, which can be seen in Figure 7, is the m/m rate of change, and while still negative the ROC has begun to improve, a sign that at least some pricing pressure is being alleviated.  That said, if we look  at Figure 8, it can be seen that pre-COVID, the m/m rate of change was considerably smaller than what has been seen since late 2019, and while the direction has certainly improved, stability is what we look for to establish a basis for a sustained recovery.  We expect such stability will not come during periods of low utilization and will take some time even after utilization rates return to more normalized levels as those changes will destabilize panel pricing for a short period until a more realistic balance between brands and panel producers is achieved.  While we are not want to give timelines for the industry, we would expect the first point at which a stable balance between both sides of the equation could occur would be late 1Q or early 2Q 2023.  Lots of things would still have to fall in place to achieve that timeline, but we see that as the first point at which it could occur.

Of the three major panel producers in Taiwan, Hannstar Display (6116.TT) is the exception, as the company’s focal point is small panel production rather than large panel, as is the case with both AU Optronics (2409.TT) and Innolux (3481.TT), the other major producers in Taiwan.  As seen in Figure 1 Hannstar’s monthly sales have historically tracked reasonably close to industry small panel pricing up until this year when small panel pricing continued the decline it began in 3Q 2021 while Hannstar’s sales increased and have maintained a similar high level with only April as the anomaly. 
What makes this unusual is that a Hannstar VP noted late last week that given the steady order pull-ins that the company had been seeing, Hannstar expected to see higher utilization rates in the 2nd half than in the 1st, essentially the opposite of the rest of the industry.  The VP went on to say that the company has been less affected by the slowdown in overall panel demand as the company has diversified its product portfolio and its customers.  He cited a company utilization rate of ~80% for the first half of the year and set expectations for a 90% to 95% utilization rate for the 2nd half of this year. 
While sales for the 1st half of 2022 for Hannstar are down 40.78% y/y, the implication is that the 2nd half (assuming small panel prices remain roughly flat), would be up 11.7%, the 5 year average, a far cry from the expectations for large panel producers in the 2nd half, especially with lower utilization rates now being assumed to remain for much, if not all of the 2nd half.  If Hannstar is able to do as the higher utilization rate implies, it would put them in a category by themselves for the remainder of the year, and while there are still many factors that could influence Hannstar’s full year and 2H results, we would certainly give credit to management for the improvement in product and customer base.  Hopefully the customer list has broadened, making the company less reliant on a particular customer, but it is a bit too early to see if the data proves out. 
All in, Hannstar presented one of the most optimistic views of the next two quarters among panel producers, but at the same time we note that the additional capacity that Hannstar announced last July (see our note 7/9/21), the first time the company has added capacity since it began production in 2005, has been ‘paused’, which seems to contradict the outlook above, but we expect Hannstar will not be the only one to slow or stall capacity increases, especially if the utilization rate reductions being seen currently remain in effect for the remainder of the year.  While we have to assume that at the first sign of either stable small panel display pricing or any improvement in consumer demand for mobile devices, those construction projects will be resumed under the thought that there will be a return to the insatiable demand for smartphones and other mobile products that existed in 2020 and early 2021.  Any optimism we might have for the small or large panel LCD display business would be predicated on at least some of those capacity projects being permanently cancelled.

As we have noted previously, panel producers have finally begun to accept the fact that the demand levels seen in 1H ’21 are not continuing this year, and that inventory levels have been increasing faster than sales.  This has led to utilization rate reductions as a number of display products become unprofitable and maintaining full equipment readiness is no longer cost effective.  While these cuts should have started earlier, panel producers were hoping that demand would return in 2Q, and then in 3Q, finally facing the fact that neither was the case.  As China has become the regional share leader in the LCD display space,it behooves Chinese panel producers to bring down utilization I  order to reduce inventory and tighten supply.
When necessary, we track those utilization levels to see if utilization cuts among Chinese panel producers are lip service or are meaningful enough to make a difference, and while the cuts are now reaching levels that have some impact on production, it was not until the end of May that such cuts began to any large degree.  While we do not have hard data for July large panel industry sales (we extend June’s sales through July until the data becomes available), but the initial cuts to utilization quickly brought done sales but have made little difference through the end of June, particularly the cuts at ultra-large fabs (Gen 10+), and while Gen 6 OLED utilization rates were low in China to begin with, they have not changed materially over the last few months.  Hopefully the additional utilization rate cuts made in July and thus far this month will have more of an impact, as inventory reductions need to be more rapidly effected if there is to be any hope for improvement at the panel level in 3Q or 4Q, so we will republish the data as soon as July hard data becomes available.  Until then the data seems to indicate that utilization rates at Chinese display fabs need to be dropped further.

While the undercurrents in the tech space over the Metaverse have been around for quite some time, but when Facebook (FB) changed its name to Meta and began its proselytizing of the Metaverse as the ‘wave of the future’, the media began to churn out article after article on how we will all be spending much of our lives wandering through endless worlds, allowed to do all of the things we cannot in the real world, and, of course, buy lots of stuff, much of which does not exist in reality.
Along with the media hype, company executives called emergency meetings to brainstorm with staff over how they could capitalize on this brave new world where folks might be willing to pay for things that don’t cost anything to make, a concept that puts a smile on the face of every manufacturing VP.  In reality (sorry), there are costs associated with the Metaverse, particularly advertising and content creation, although the cost of content creation for the Metaverse is a fraction of the cost of almost any physical product.
Once those meetings occurred the word went out that corporations large and small needed to create new ‘teams’ that can devise and execute a strategy that would move the company onto this new platform known as the Metaverse.  Job postings with the word ‘VR’ began to creep up after the Facebook rebranding, but things really took off at the beginning of this year, doubling the 2021 high point in January and reaching almost 4x the January number by April.

Since then however the bloom is off the rose so to speak, and the demand for VR related jobs seems to have slowed back to January levels, with IT consulting firm Accenture (CAN) posting far more VR related job postings than Meta.  Meta VR jobs focused are on R&D, likely ways in which to improve and reduce the cost of the VR experience, given their desire to seed these new virtual worlds that will provide an additional source of user data that can be sold to customers looking to tailor their Metaverse ‘worlds’ to specific buyer groups.  Accenture seems to be headed more toward content creation, a market that does not require an investment in the manufacturing of physical hardware.

​While content creation in the VR world is more complex than what we are used to seeing in 2D mode, content creators from the gaming industry tend to be well-versed in visual 2D development and have a leg up on those making the transition from print on static media creation to VR.  Tools like Unity (U), Sumerian (AMZN), and Unreal Engine (pvt) are familiar to game developers who can relatively easily make the transition to VR once they master the nuances of 360⁰ cameras, but now they will have to contend with a bit of competition for VR jobs, as hiring has slowed considerably as consumer spending in the CE space has slowed.  With medical school costing between $130k and $230k in tuition alone but an average salary of ~$210k for physicians, the potential for a profitable return from a medical education certainly exists, but with the average income for a software engineer being ~$120k, with a relatively low cost up front, one can see the attraction.  Now someone just has to figure out what the Metaverse is, other than a concept, and businesses can get back to hiring and get those developers out of their parent’s basements and into the workforce…

​According to the Taiwan Stock Exchange, shares of Micro-LED producer PlayNitride (6854.TT) began trading today.  PlayNitride is the first company to be trading on the new Taiwan Innovation Board, similar to that of the Chinese Innovation Board that has less onerous requirements than standard trading platforms on those exchanges.  The original deal was for a maximum of 6.305m shares at a maximum price of 138 NT$, and while final trading results are not in quite yet, it seems the deal came at $NT$ 127 and closed 9.9% higher at NT$139.5, after trading 597k shares.  We will note deal value when we have actual shares completed, although preliminary value was ~.5b US$..
The company has posted 3 months of monthly data (May – July), with July sales of NT$37.419m (~1.24m US), up 93% y/y, and full year sales of NT$122m ($4.05m US) in 2020 and NT$205m ($6.8m US) in 2021, with operating income of NT$-792m ($-26.3m US) and NT$-1.24b ($-41.13m US) respectively, with capex in those years of NT$327m ($10.85m US) and NT$232m ($7.7m US).  PlayNitride’s chairman stated that the company is to become (a/the?) major micro-LED chip producer over the next two to three years.  As the first listed pure play in micro-LED PlayNitride will be closely watched. 

​Vietnam is the new kid in town, making a bid to replace China as the mecca for CE and technology manufacturing in Asia.  Samsung has invested over $17b in Vietnam over the last 10 years and Intel (INTC) opened a $1b semiconductor assembly and test facility, while Foxconn and Pegatron (4938.TT) have built out significant assembly facilities in the country, due to its low wages, young population, and numerous trade agreements, but there is a catch.
Vietnam is run by the Communist Party and earlier this week the Vietnamese government issued a decree stating that technology companies that operate in Vietnam must store their user data locally beginning October 1, which would force those without an office in the country to set up a local office.  That’s the easy part as the decree states that “Data of all internet users ranging from financial records and biometric data to information on peoples' ethnicity and political views, or any data created by users while surfing the internet must be to stored domestically”, and authorities have the right to request that data during any investigation.  Further, if any content is thought to be a violation of government guidelines, the government has the right to ask content providers to remove it.  Social media firms such as Meta (FB) and data collectors such as Google (GOOG) will have 12 months to set up local offices and data storage and will be required to keep all data locally for at least two years.
Taking a nod from China, Vietnam began working up to such a rule starting in 2019 when it implemented new cyber-security rules, which were followed by social media content rules last year, so the upside to Vietnam’s manufacturing infrastructure and lucrative trade agreements is government censorship and a free look into data if so asked.  These are the same issues that pushed companies out of China, and while their implementation might not be as rigorous as those on the Mainland, they are going to cause companies to think a bit more about the implications of adding production in Vietnam, especially if those rules are enforced or requests for data are actually made.  There is always a downside.
 
 

Innolux (3481.TT), a subsidiary of Foxconn (2354.TT) is the 4th largest global large panel LCD producer by revenue, having been producing display panels since 2003, and as one of two remaining large panel producers in Taiwan is a supplier to a wide variety of CE brands.   Last month the company predicted that it would see shipments for 3Q remain flat, despite panel price weakness and high inventory levels, but it seems they have changed their outlook since then and are now expecting shipments in 3Q to drop by 12% and expect to lower their utilization rate from 90% last quarter to between 50% and 70%.
The company’s chairman stated “It is meaningless to manufacture that many panels when selling prices have dipped below the cost level and inventory digestion is slower than expected”, which seems to be stating the obvious when looking at Figure 7 & Figure 8,  and is directing the company to allocate more capacity toward high-value added products in order to ‘optimize’ the product portfolio, something other panel producers have been doing since the middle of last year.  The company is now looking for only modest sales gains in 4th quarter, despite the holiday season after the company slipped into a loss position for the first time in almost two years.
Other than the closing of the barn door after the horse has gone, what gives us concern is that the company is pinning its hopes on the weather in China, with it’s best case scenario being one where the power cuts in China cut panel production between 3.5% and 5%, or 2m to 3m 32” panel equivalents, and stated that the ~70% reduction in power to factories in Sichuan would not be enough to keep production lines running.  Under such a scenario Innolux management sees a bottom for the industry by year-end and a good start to 2023.  The company also believes that global notebook shipments will be maintained at the higher levels reached as the COVID-19 pandemic forced people to remain sheltered, with the idea that remote learning and a hybrid lifestyle will maintain notebook demand for the next few years, although based on Figure 9 we expect notebook shipments will fall back to pre-pandemic levels, as they seem to already have. 
Unfortunately, this ‘head-in-the-sand’ approach does little but exacerbate the negative display market situation that has been developing since the middle of last year, but we keep our expectations low.  While Samsung Display (pvt) and LG Display LPL) made the decision to step out of the large panel display business a few years back, others seem to be willing to look at the world through rose colored glasses, which only works well when prices are rising.  The glasses should have come off last July.

​Japan was one of the strongest proponents of ‘at-home’ learning during the early days of the COVID-19 pandemic, and the country's GIGA program, which was actually developed in 2018 (pre-pandemic) to move Japan’s 13m primary and secondary school students more quickly into the digital age, as tight educational budgets had limited the number of computers available in the country’s 35,000 schools, but as COVID became an issue, the plans were pulled forward.
The program’s mantra was “1 device for 1 student”, with 2/3 of the program’s budget going toward PCs, with the rest going toward programs to help teachers develop digital learning environments that will help them learn about the best ways to search for information, assemble presentations, and become familiar with digital testing, with the ‘added benefit’ of helping shy students express their ideas without having to raise their hands, although we find that more of a crutch than a panacea.
While the GIGA program, at least the hardware side, has ended, the Ministry of Education has decided to start testing the use of digital textbooks at elementary and high schools, providing free digital textbooks, primarily to learn English, to 5th and 6th grader and junior high school students.  Both digital and paper textbooks will be given so that the Ministry can study the effect of the digital implementation, although the program is expected to go fully digital by 2024.
90% of Japan’s municipalities have already completed the distribution of tablets to school children, but many schools are said to still be struggling to integrate them in daily teaching routines, but the digital textbooks will allow students who are learning English to hear how to pronounce words to see if that helps to improve language skills, and those schools that have an interest can pick one additional subject (math, science, music, arts & crafts, technical arts, homemaking, or health & physical education) to be supplied digitally.
While it is hard to gauge the impact of digital textbooks on such a narrow subject field, and while the Ministry is still trying to come up with a plan as to how to introduce digital textbooks effectively, it seems they have made up their minds about a full-scale introduction in 2024, which will mark the end to paper content in Japanese schools.  As to the program’s effect on tablet or notebook sales in Japan, with much of the student population already equip with a laptop or tablet, the effect should be minimal unless students find that those devices given out in 2020 and 2021 are a generation or so behind what is then necessary for digital textbook content in all subjects.  At that point however, we expect the burden and cost of upgrading to a more modern laptop or tablet will fall to families rather than the government, which we expect will limit the upgrade cycle a bit.

Taiwan-based Yageo (2327.TT) is arguably the largest producer of passive components (capacitors, resistors, and inductors), and while passives are essential for the operation of all CE devices, they get little of the recognition and fanfare that silicon-based products receive, and remain in the shadows in the CE supply chain.  While CE products usually contain a few silicon-based components, they can have hundreds or even thousands of passives, and while they are small and quite inexpensive singly, CE devices can run without them.  With growth rates in the 2% to 3% range in good years, components tend to be relatively low-risk investments, but ones that are highly cyclical, so component company investors tend to look for sub-categories that are seeing a technology or demand change, such as MLCCs (Multi-Layer Ceramic Capacitors) that are used to regulate current flow and remove noise from circuitry and have increasing applications in mobile devices and electric vehicles.
But we digress as this note is not about the virtues of MLCCs or components in general but about inventory, much of which in the component space, is carried by distributors such as Arrow Electronics (ARW), Avnet (AVT), TTI (BRK.A), Future Electronics (pvt), and Digi-Key (pvt), who stock millions of components for their customers who are typically not large enough to buy directly from the producer.  Some are supplier authorized or franchised, while independent distributors have no direct affiliation with component producers and buy in the open market. 
In the distribution business, there are a few factors which are most important, first, are the distributor’s components legit?, and while this seems a bit odd, if a CE company finds that product failure is coming from a specific component, they will want to trace that component back to its source and that is why working with an authorized distributor has a bit premium attached.  Second is availability, which can  translate into lead time if larger quantities are needed, and again the added leverage with producers that authorized distributors have can shorten those wait periods and keep CE product roll-outs on schedule.  Last is price, and this is where independent distributors have the advantage as they can negotiate deals in the open market, which are not constrained by producer list pricing.
Back to Yageo, who just came of the best quarter in the company’s history, up 3.9% q/q and up 13.0% y/y, but warned that it is expecting sales to drop between 2% and 3% q/q as both direct customers and distributors hold back orders as they try to reduce inventory levels, which are unusually high against weak smartphone and notebook demand.  The company expects to reduce factory utilization rates in 3Q by ~10%, from 70% top 60% for standard passive components, in order to bring down inventory levels from 130 days to between 100 and 110 days, which is considered normal, by the end of the year or during 1Q ’23.  Premium passive products are still seeing strong demand from automotive, networking, and industrial customers, and will still see utilization rates between 90% and 100% at the company, but commodity passives will be the burden for the remainder of the year.
Optimistically, and perhaps rightly so, Yageo expects only minor impact from the expected pricing pressure for the remainder of the year, as the premium segment accounts for ~75% of the company’s sales (1Q), but much will depend on how quickly inventory can be drawn down to more normal levels, with the company’s view being about six month at 60% utilization to bring inventory levels down by ~20%, while little was said about how and why those inventory levels continued to build in light of the many indicators that have been pointing to a slowdown across a number of Yageo’s end market segments for some time.  We guess that if distributors are asking for product, regardless of the macro picture, Yageo had little choice but to supply whatever they were asking for, although knowing how close to the edge of a precipice they were heading and we expect other passive component manufacturers felt the same way, leading up to what will probably be a weak 2H for all..