There has been some focus on component price increases in South Korea recently, with graphics cards being singled out as one component seeing rapid price increases over the last few months.  Back in August we noted that the prices of graphics cards has fallen as China began restricting the mining of bitcoin, one of the primary demand drivers for graphics cards, and such cards had suddenly become available after being in short supply.  We went back to the same two cards that we used as examples in the 8/11/21 note to see how prices changed over the last 2 ½ months.  While China has virtually eliminated the big bitcoin farms across most of the country in their push to conserve power, as we have noted previously, bitcoin farmers have moved to other countries where power is both inexpensive and the regulatory climate is a bit more amenable to such non-essential industries.
In South Korea graphic card prices, based on 20 major graphics cards checked between August and the 2nd week in September, increased between 6% and 100% according to local tech press, while the four examples given in the graphic (Fig. 5) saw increases between 23.2% and 65.3%, with an average of 47.8%.  This compares to the two cards we used in our last check, which saw increases of 13.9% and 16.1% based on current prices, although when each of the two reached peak levels in mid-September, they were up 27.1% and 78.1% from our August check points, much closer to the increases seen in Korea.  It would seem that using the most current data here, which comes from 3rd party sellers on Amazon (AMZN) (Amazon does not have many cards directly available), we should expect some moderation to graphic card prices overseas.  That said, with the price of bitcoin nearing its all-time high, such a theory, based on what we have seen in the US, seems unlikely as the economics of bitcoin mining get better as the price/hash rises.  If you need a graphics card, despite the rise in prices, it would seem to be the time to buy, as odd as that appears.
 

Item ‘finders’ or personal trackers are a handy item to have if you are one of those people that tends to lose things of value, and by value we mean things that are essential to life as we know it, like car keys, smartphones, and wallets, but Apple’s Air Tags and Samsung’s Smart Tags, while both will (eventually) work through UWB (Ultra-Wide Band) networks, they are proprietary to the operating system, meaning Air Tags will only work on Apple products and Smart Tags will only work on Galaxy (Samsung) devices.  But there is hope, again eventually, as the third supplier Tile (pvt) finders will work on both Android and iOS systems and has just announced that it has developed its own UWB enabled tracker, the Tile Ultra.  While the Ultra is not yet available (‘early next year’) it has the advantage over both competitors in not being dedicated, but will lack the vast network of users that the Apple and Samsung products have, limiting the device’s ability to be located by passing users unless they happen to running the Tile app.
This would make the Tile Ultra best for items that are used locally, meaning remote controls or similar items, while both the Apple and Samsung tags would be more effective for items that are carried into public places, although Tile says it has sold over 35m tags since inception.  (we expect only a percentage of the early devices are still in operation given the short-term batteries)  Tile does have an agreement with Amazon (AMZN) to give access to the company’s Amazon Sidewalk Bluetooth/900 MHz network, which would include other devices on the Sidewalk network acting as additional ‘points’ of discovery, so if you left your phone at a friend’s house who had a Ring () doorbell system, there is a chance that your phone would be identified by the network.
While each tracker has its plusses and minuses, potential users should evaluate the circumstances under which they are using the devices.  If they are constantly traveling in the US, the Apple network, with its millions of active users might make the most sense, while international travelers might rely more on the Samsung product, while Tile seems to be working towards making itself as agnostic as possible, which, in the final analysis, should be the determining factor, although the big marketing machines at Apple and Samsung do exert a lot of influence over consumer attitudes.  The one thing that is common to all trackers however is the need for security, as surreptitiously placing a tracker on a person or object would allow a stalker to track that person’s or object’s location down to an accuracy of ~16’.  All three tracker companies have their own version of security, but so far none are foolproof, so it is imperative to understand the balance between the merits of trackers and the pitfalls.  They will continue to get better but caveat emptor…
For more information on UWB tags see:
8/5/21                  Air Tag Tussle
4/2/21                  The Tags are Getting Close
1/18/21                 ​Lost in the Leaves

ToF technology is a way of using sensors to measure distance, a logical function for setting camera characteristics when taking photos or video, but the technology has seen its prospects change a number of times since it began to be used in smartphones in 2014 as the ‘phase detection/laser autofocus’ feature in the LG (066570.KS) G3 and appeared in the Blackberry (BB) PRIV smartphone in 2015, but interest waned until 2018 when a number of Chinese brands included the technology, with Samsung (005930.KS) taking up the TOF cause in 2019 with the Galaxy S and Note series, with it showing up even in Samsung’s lower priced A80 series.  Apple joined the TOF cadre in 2019, including it in the iPhone 12 Pro and Pro Max, and it seemed the technology was off to the races.
That optimism waned a bit when Samsung decided to drop the technology from its 2021 smartphone line, which was particularly unusual in that Samsung began producing its own TOF sensors in late 2020, and while others continued to build in TOF hardware and functionality, without the backing of smartphone giant Samsung, interest in TOF seems to have cooled a bit this year.  But why would Samsung drop the technology, especially given its ability to self-source and its 2 years of promotion with little explanation other than saying it did not spur developers to come up with applications that used the TOF data and that consumers did not seem to care if it was removed.
That said, we believe it was not that simple an explanation, as there is some nuance that separated Samsung’s TOF ‘flavor’ from the TOF used by Apple, and that difference resulted in relatively poor TOF performance on Samsung’s phones and better performance on others.  That nuance is in the two possible ways in what TOF can collect distance data.  In the ‘direct’ method, such as is used in Apple’s TOF, a laser is pulsed at intervals less than 5nsec.  That light, which is outside of the human visual range, bounces off objects and returns to the device to be captured by a SPAD (Single Photon Avalanche Diode) array.  Given the speed of light is a constant, the system can calculate the distance based on how long it took for the reflection to appear; simple math, right? ( L = c 2 φ ω = c 2 φT 2π ), but such systems need lots of pixels to ‘see’ the reflections and must process a considerable amount of data, although they are fast and are able to understand ‘echos’ that occur when the laser bounces off other objects on its way back to the device.
The ‘indirect’ method, the one championed by Samsung, is a bit more esoteric in that the laser sends out a continuous signal and the system measures the ‘phase’ difference between the laser signal and the return, with ‘phase’ meaning a point in each light wave’s cycle.

​This method has less computing demand, which should make the module smaller, a big factor in smartphones, but has more trouble with echoes, which might explain the complaints that we were able to dig up concerning the TOF sensors in the Samsung phones.  We believe that Samsung was unable to compete with ‘direct’ method TOF sensors and recognized that the extra cost of the TOF sensor and the poor results were just not worth the inclusion in 2021 models.
Direct TOF modules for Apple are supplied by LG Innotek (011070.KS) and are used in the iPhone and iPad lines, but recent information has indicated that LG Innotek is also supplying TOF sensors that are expected to be used in VR headsets from Apple and Microsoft (MSFT) next year, similar to the ones it supplies to the Oculus VR headset produced by Facebook (FB), with all based on sensor arrays produced by Sony (SNE), with whom Apple has an exclusive contract.  While Samsung is said to be developing its own direct TOF sensor, we expect it would have been difficult for Samsung to have justified buying direct TOF sensors from rival Sony when it was producing and promoting its own indirect TOF sensors, thus the abandonment of TOF in the Samsung 2021 smartphone line.  If Samsung is able to develop a competitive direct TOF sensor for the 2022 smartphone line, we could see a measured return to the technology for Samsung in 2022, while Apple and others continue to develop additional applications for the technology, pushing out into new products over the next two years.

​As we have noted previously, there has been a shortage of driver and timing ICs for display panels.  This has limited panel shipments slightly but has built on itself with panel producers accumulating inventory levels that are above normal as a safety net for further potential driver/TCON shortages.  While this has proved to be a good solution for much of the year, the current slowdown in demand for notebooks and TVs would be a point where panel producer demand would slow, especially as foundry costs for such products continue to rise in 4Q, however monitor demand is still robust that panel producers have continued to place driver orders, which is expected to push driver prices up a bit again in 4Q.  With relatively high driver inventory levels at LCD panel producers going into 2022, we expect demand to slow a bit further and LCD driver prices to stabilize or weaken. 
At the smartphone level the situation is a bit different, with LCD smartphone demand waning and small panel producers working out of inventory.  Small panel OLED producers face similar demand challenges, however OLED drivers are limited to a smaller production base, which is expected to keep OLED driver demand strong and prices high for the remainder of the year.  While OLED smartphone expansion will continue next year, unless there is a resurgence in smartphone demand, we expect flat OLED driver pricing in 1Q ’22 and weaker LCD driver pricing during the same period.

Samsung Electro-Mechanics (009150.CH) has decided to make a big investment in semiconductor substrates and is currently waiting for board approval to start the ~$845.3m expansion project.  Given that Samsung EM spent $609m last year for facility maintenance and expansion last year, across the entire company, which also includes production facilities for passive components and camera modules, the amount being spent is significant.  In fact, the company asked the board to approve a $1.44b investment but the board lowered that to between $845m and $930m.
As almost all semiconductor substrate products have been in short supply for almost a year, made worse by a series of fires at Unimicron (3037.TT) in October of last year and again in February of this year, and the potential closing of SEMCO’s RF PCB business later this year, delays in deliveries of PCB based components have been high enough to cause product roll out pushbacks and rising prices.  There are a limited number of substrate suppliers, and while the global market is large (Estimates range from $53b to $64b last year), many of the suppliers have a particular product niche.
In that regard, Samsung EM is expected to spend much of the capital on Flip-Chip BGA substrates where there are few suppliers, and is expected to be dedicating some or all of these lines to a “US Chip Company”, likely Intel (INTC), who will also be making an investment in the lines.  FC-BGA is most commonly used for CPUs, GPUs for servers and mobile devices (smartphone application processors).  Samsung EM currently generates ~$423m from FC-BGA, which is expected to grow to over $1b when the expansion is complete.  No timetable for the expansion has been set, although we expect more detail once the board gives final approval.

Vietnam has become a major production base for many companies that have been looking for alternatives to production in China, with an ample workforce and positive government support for global businesses, but recently (early July, the country, which has a vaccination rate of ~26.6% has seen a big spike in new cases, limiting travel and logistics, and closing plants during the spike.  There have been a number of commentaries and announcements concern where the outbreaks have been most prominent and what companies were affected, but little definitive data on the extent of the issue has been circulated.  The American Chamber of Commerce in Vietnam[1] has done a number of ‘flash surveys’ to gather some data concerning the status of US and foreign companies operating in Vietnam and at what level of production they are operating.  The latest survey was taken between August 23 and 25, which is close to the peak infection point for this cycle, so it should give some insight into how difficult conditions are for manufacturers in the country.
Survey Results:
Vaccinations have seen a sharp uptick, especially around Ho Chi Minh City, the country’s largest city, where almost 50% of companies polled indicated that almost 50% of staff had received at least one dose of the vaccine, up from 25% in a survey taken 3 weeks ago.

• 13% of those surveyed indicated that their companies had ceased operations or were operating with a skeletal staff, with nearly 50% operating at under 50% of normal capacity.
• 20% of respondents have already shifted some production, with another 16% discussing options.
• While AmCham members and families were worried about the poor and disadvantaged, who were being impacted most severely, 71% expressed ‘concern about quality of care if you or a family member develops serious symptoms.’

 


[1] AmCham Vietnam represents 650 corporations and 2,500 individual business representative. The US is Vietnam’s largest export market

The survey data indicates that over the 20 day period between surveys, those businesses that had seen the least impact from the COVID-19 outbreak (operating at 50% of normal or better) saw things recover, however those that had been most severely affected saw things get worse, with workers hesitant or unable to return.  The government of Vietnam has implemented a “3 on Site” model in July, which involves eating, sleeping, and working without leaving the factory site, however as the outbreak became worse, particularly in the more populous southern provinces, the wisdom of such a policy was questioned and suspended in Ho Chi Minh City where over 3,700 businesses had spent considerable sums to implement the practice.

​While the concerns of businesses over the spread of the virus continues to grow, it seems that the biggest issue right now is a lack of coordination from the government.   Provincial governments do not seem to be operating with a coordinated approach to vaccination programs or rules and regulations concerning protection from the spreading infections.  There does seem to be some improvement from the higher vaccination rate, but low efficiency ports, either for inflowing components or outgoing product are unacceptable for companies, especially those in the CE space where the selling season is limited.  That said, a lack of government coordination seems to be an issue for many countries including the US, so while Idaho remains the state with the lowest vaccination rate in the US (46.4%), that seems to be a voluntary issue, where the 26.6% rate in Vietnam is due to a lack of vaccine and a lack of government coordination.  Welcome to the real world.

Much focus has been given to semiconductor based component shortages, and rightly so as capacity has been behind demand for many months and demand has remained relatively strong throughout, yet more basic components face similar shortages and get much less attention.  In most cases basic component shortages tend to push out lead times, which tends to be a good indicator of the supply/demand balance.  There are almost an infinite number of small components such as resistors, inductors, and capacitors and such a large variety of each makes tracking price almost an impossibility, leaving lead times as the alternative.
Fig. 1 shows the long-term lead time trends for capacitors on a general basis and Fig. 2 shows lead time changes for the last few months.  We have focused our attention on MLCCs ((Multi-Layer Ceramic Capacitor), as they have become an even more integral part of a number of consumer electronics products (5G smartphones, electric & hybrid automotive systems, etc.), and they continue to see lead time increases and increasing demand and sales, with 3Q expectations for a 15% q/q increase in sales, but lead times for other capacitors (aluminum, tantalum, etc.) have also been increasing as supply disruptions in Malaysia and rapidly increasing raw material costs take their toll. 
Tantalum concentrate alone has seen a 33.02% increase in price over the last 12 months while ruthenium, also a basic resistor material, has increased 158.6% over the same period.  While some of the factors that have caused price and lead time spikes over the last year are COVID-19 related and are considered ‘transitory’ by some, the net is that we have to live with what is available and for capacitors, which have typically been easily sourced, that could mean waiting as long as 48 weeks for certain products.  With well-planned ordering, lead times of 20 weeks to 24 weeks are barely tolerable, but for products where lead times are greater than 6 months, we would expect either end product delays or eventual product cancellations unless you have the buying power of Apple (AAPL) or Samsung (005930.KS).
​

Samsung is the major competitor to Sony’s dominance of the image sensor market.  Even after losing business from Huawei (pvt) due to the US trade ban last year, Sony still had a share just a bit less than double that of Samsung, the number 2 contender in the almost $21b image sensor market, but Samsung continues to push product limits and has been gaining ground over the last two years with an aggressive new product schedule.  In August 2019 Samsung released the first image sensor (IOSCELL™) with over 100m pixels which has been used in over 35 smartphone models by Samsung itself, Xiaomi (1810.HK), RealMe (pvt), Motorola (MSI), Honor (pvt), Infinix (688036.CH), and Vivo (pvt), and while since then Sony and others have released CIS sensors with higher pixel counts (Sony – 127MP), Samsung has just announced another step up in the image senor wars.
Samsung announced today the ISOCELL HP1, a 200MP image senor and the ISOCELL GN5, a 50MP version with high speed auto-focus.  The new sensors use a process that Samsung calls Chameleon Cell, which senses the lighting environment and can combine between 4 and 16 adjacent pixels to gather more light in low light environments.  If light is sufficient, all 200 0.64um pixels are used individually.  The sensor is 1” x 1.22” slightly smaller than the 1” x 1.33” 108MP ISOCELL HM3 and can shoot 8K video at 30 frames/second and 4K video at 120 fps. 
We would expect the next Galaxy S series, due out in January 2022 to be among those sporting the new sensor, which will undoubtedly be followed by high pixel count versions from Sony and others, but it seems that Samsung has pushed the image sensor boundaries a bit faster than others and continues to set the pace.  While pushing higher quality image sensors into the smartphone market is certainly a plus for consumers, we still try to grasp at what point such features reach a point of diminishing returns.  Hopefully there are enough consumers who would be able to tell the difference between images taken with such high pixel count sensors and would count such as a reason to pick smartphone models with same, but the size of that consumer base seems to get smaller with each iteration as we expect most look at their smartphone cameras as utilitarian devices that can capture selfies, food pics, cat videos, and dogs with Frisbees when called on and are less concerned recapturing Ansel Adam’s-like images for coffee table books.

​Last Thursday we noted that Murata (6981.JP), the leading supplier of MLCCs (Multi-layer Ceramic Capacitors), a small, but essential component used in 5G smartphones, automotive electronics, and a myriad of other consumer electronics, had closed its factory in Takebu, Japan due to an outbreak of COVID-19 adding to the already tight market for such devices.  With much of additional MLCC production coming from factories in Malaysia and the Philippines, where COVID-19 remains a significant public health threat, other suppliers are facing potential issues in terms of production, with estimates that Samsung Electro-Mechanics (), the 2nd largest producer of MLCCs, has ~40% of its capacity in those countries.
We noted that lead times for MLCCs have been increasing and we have seen data indicating that the industry fell short of demand this month by between 2% and 10%, with shortages expected to continue until early 2022, and while supplier Yageo’s (2327.TT) plan to reduce MLCC prices by 10% to its Chinese customers is still an enigma, if they are looking to generate share, China represented over 45% of MLCC demand last year, making it the largest regional buyer.  China itself does have capabilities for MLCC production but has yet to establish itself as a global competitor in terms of quality, but as a market that has grown from $8.55b in 2016 to $16.25b last year, it presents an attractive opportunity for Chinese suppliers looking for growth oriented markets.
Fenghua Hi-Tech (000636.CH) has indicated that it will be spending $1.16b in three phases to increase its MLCC production capabilities by 45b units/month when fully completed or 10% of global capacity according to the company.  A second Chinese MLCC supplier, Chaozhou Sanhuan (300408.CH) has also indicated that they will be raising at least $600m to increase MLCC capacity, although we have no timeline as of yet.  With both companies increasing capacity by roughly 100% between 2016 and 2020, China is certainly making an attempt at penetrating the MLCC market, but much of the proposed capacity is still in the planning stage, so we expect little significant change on the supply side this year and with the problems surrounding COVID-19 in Malaysia, the Philippines, and Japan, we expect the market to remain tight this year despite Yageo’s plans.

​It was only 3 or 4 months ago when we were being told by suppliers that customers didn’t seem to mind the price increases that were being put into effect during the latter part of 1Q and early 2Q.  Component suppliers commented numerous times that customers “were just happy they could get product, and didn’t mind the cost”.  While there are still many CE products and components that are in short supply and are seeing price increases, particularly semiconductors, we have noticed a number of times recently that the easygoing attitude toward price increases has waned and the seller’s market seems to have lost a bit of its vigor.
Much of this is anecdotal, although the weakness in TV panel prices seems to indicate a certain amount of reluctance toward the price increases that have already been built in over the last year, and IT panel buyers (monitors and notebooks) are becoming more cautious with inventory into the holiday season.  This is not to say that price increases will still not occur, but we sense a ‘disturbance in the Force’, rather than the complacency seen over the last 12 months.  It is hard to gauge what underlies this change, but it would seem that the enthusiastic demand seen as the global economy emerged from a shelter-in-place environment has perhaps been sated, or at least has shown that it is not all consuming and has limits as to how far it will go to be satisfied.
We note that there have been smaller ups and downs throughout the last year in terms of CE demand, but, while we don’t believe that price increases or shortages in a number of components will abruptly end, the sense that those price increases need to be far more justifiable than in the past seems to be the case.    Those brands who came into the year with very high expectations are now beginning to question the sustainability of those forecasts and targets.  There is still some time before the key selling months (November & December) and final adjustments, positive or negative, will still be made, but the bloom is off the rose so to speak.