We have noted that Samsung Electronics continues to limit the use of TOF (Time of Flight) sensors in its smartphones after championing their use a number of years ago.  The reasons for this change are a bit complex, as discussed in previous notes[1] but Samsung’s decision does not affect our expectations for VCSELs, the basic building block for TOF and other applications, as we believe there is still considerable long-term growth for VCSEL technology as demand from consumer products and automotive applications continues to grow.  Both TOF and Structured Light applications in consumer devices are based on VCSEL sensors, with imaging and security Identification being the most common applications in consumer products, particularly smartphones and other mobile devices, but LIDAR applications, essential to autonomous or ADAS applications are also based on VCSEL technology and as the cost decreases and power increases, the demand for VCSEL application in automotive and logistics will also increase.
Estimates from Yole Development indicate that broad-based growth for VCSELs will have a CAGR of 13.6% during the period between 2021 and 2026, with sub-segments growing between 5.6% and 121.9% as indicated i9n the table below.  While growth is highest in smaller segments mobile/consumer remains the largest segment, growing 3x faster than telecom, the 2nd largest, although the automotive/mobility segment sees considerable growth over the period due to the development of higher power VCSEL arrays for LIDAR and applications for in-factory electric vehicles and similar pallet moving equipment that would be outfitted with collision avoidance systems.
While most VCSEL sensors emit at near-IR frequencies of 850 or 940nm, Sony (SNE) has developed a sensor that covers the range between 400nm and 1700nm by adding a layer of Indium Gallium Arsenide (InGaIn) to the Indium Phosphide (InP) that covers the near-IR spectrum.  The additional layer expands the range of the sensor into the visible light range, but more importantly it reaches the frequencies at which OLED display materials become transparent, which implies that a sensor can be placed within the OLED display area, rather than in a separate ‘notch’ as is done with the typical 940nm VCSELs used currently, and at the same time reduces the chance of ocular injury from VCSEL arrays as power continues to increase while reducing background ‘noise’, leading to improvements in resolution and range.
While VCSEL material and process advancements also come with challenges such as lower die count, manufacturing retooling and requalification from customers, VCSEL volumes will continue to increase and overall manufacturing costs will continue to decline.  There are a number of VCSEL die level producers, such as Samsung, Sony, ST Micro (STM), and AMS (AMS.SW) but much of the press goes to companies like Lumentum (LITE), II-VI (IIVI), Vertilite (pvt), and Trumpf (pvt) who are more visible in the consumer space.
While there is much data on the characteristics of VCSELs, there is little data on pricing, as widely differing applications generate a vast number of VCSEL modalities, sizes, source counts, power levels, spectrum,  and packaging, but more to the point is that price and characteristics are closely aligned with improving characteristics commanding premium pricing while more generic modules decline.  There are step function improvements, such as when the shift from 4” to 6” wafers was made years ago that increased the die count by over 2x and reduced cost/die significantly, but we expect overall VCSEL pricing to be more of a function of the number of emitters and power capabilities, so we look more toward the growth in unit volume for indications as to how the industry is progressing.  While estimates 5 years out tend to be adjusted yearly even a 10% CAGR for the industry out 5 years would be considered formidable, even without Samsung’s help.  


[1] 7/20/22, 1/13/22, 10/14/21, 4/19/21, 3/5/21
​

​Last year Samsung Electronics (005930.KS) shipped 270.5m mobile units, up 5% y/y and had high hopes for smartphone shipments this year, at least at the end of last year, but the most recent prognostications from the share leader now call for mobile shipments to be down slightly for the year, after a 14.4% decline in shipments in 2Q and the specter of excessive inventory both at Samsung and across much of the smartphone industry.  Samsung indicted that it is focusing on the premium smartphone segment and expects to see sales and ASPs increase in that category as it garners increasing shipments of its foldable smartphone line, but it is still competing with Chinese smartphone brands in the mid-range price tier, where the highest unit volumes take place.  
In that regard, Samsung reduced orders from component suppliers in 2Q and has hinted about other cost saving measures some of which relate to cameras and camera module reductions in mid-tier models.  While the full impact of those changes has not been felt, the overall component order reductions are just coming to the surface as South Korean camera and related component suppliers begin to report 2Q results.  Most recently Partron (091700.KS), a long-time Samsung camera module supplier, reported a 21% drop in sales q/q and a 5% drop y/y/, with operating profit declining 61% q/q and 45% y/y, and Samsung Electro-Mechanics (009150.KS), 23.7% of which is owned by Samsung Electronics, saw its optical solutions division’s 2Q sales decline 10% q/q and 4% y/y.  SEMES has focused on supplying its parent with camera modules for its high end lines, such as the Galaxy S and Note series, but has more recently been supplying modules for the A series, again a mid-tier line where the competition is quite intense.
There are a number of other similar suppliers in Korea that have likely faced the same challenges in 2Q as suppliers of optical components and cameras to Samsung but have not yet reported results, such as MCNEX (097520.KS), Power Logics (047310.KS), Cammsys (050110.KS), Namuga (190510.KS), and Coasia Optics (196450.KS), but more to the point will Samsung’s longer cost cutting measures magnify the slower sales of smartphones for the remainder of the year and into next year?  We expect that the ‘camera wars’ which have led to the proliferation of multiple cameras, with some models having 5 or 6 front and rear cameras, has come to an end, although we expect the focus to now shift to ever higher resolution cameras and better image stabilization systems rather than an excessive number of cameras.  That said, it would seem that at least the next few quarters will be bumpy for Samsung’s local camera component supply chain as Samsung and the industry readjust order levels to lower demand.

MLCCs (Multi-layer Ceramic Capacitors) are tiny things, ranging from as small as 1/4mm to as large as ¾” and while the price varies across the hundreds of MLCC types and sizes, they typically sell for pennies.  With ~70% of MLCC demand coming from consumer electronics products, tracking inventory, shipment, and pricing levels can give anecdotal insight into the CE space, and given their use in a wide variety of other industries, particularly automotive, MLCC metrics can point to medium and long-term trends in those industries, and with a relatively small number of manufacturers and a typically conservative viewpoint toward capacity expansion, MLCCs are reliable indicators of the health of the global electronics market.
MLCC unit count in various CE & Automotive products:

• Laptops – 400 – 800 units
• Smartphone – 200 – 1000
• LCD TV – 500 – 800
• Gas Engine Vehicles - ~5,500
• EVs - ~12,500
• ADAS - ~15,000
• Autonomous Vehicles - ~17,500

 
That said, electrode and termination materials in MLCCs are Nickel, Copper, Tin, and Palladium, and while very small amounts are used in all but the high voltage MLCCs used in EVs, the fact that over 1 trillion MLCCs are produced yearly, makes raw material price swings quite significant, with the cost to produce MLCCs representing ~75% of industry sales.  However, given that MLCCs have the lowest cost structure for all capacitors, they represent the highest profitability levels for capacitor producers.
Early this year many CE companies had an optimistic view of 2022 product prospects, much of which was based on the demand levels for a number of CE products due to COVID-19 in 2021, but there were a number of red flags underlying that optimism, particularly the deterioration in LCD TV panel prices, which began in July of last year.  While this was specific to a particular CE segment, leaving IT (monitor, notebook, and tablet) prices to continue rising, if foreshadowed the more general decline in panel prices seen in recent months, despite the insistence of most panel producers that the weakness was limited to only TVs.  Concerns about semiconductor supply shortages coupled with this optimistic view, caused CE companies to overestimate demand and build component and product inventory based on 2021 results and 1Q ’22 results, which were still being driven by the optimistic 2022 targets.
Now that reality has set in and targets are being lowered, we see component demand weakening and prices following which heads us back to MLCCs.  Taiwan-based Trendforce is now estimating that MLCCs will see 2Q prices decline by 3% to 5% and by 2% to 4% in 3Q, citing the continuing decline in TV panel prices as an indicator, along with the continuing weakness in smartphones, which include both LCD and OLED.  While those quarterly price reductions will likely do little to affect the long-term growth trendline for MLCCs seen in Figure 2, it will serve as a cautionary note toward plans for MLCC capacity expansion, with most of the large suppliers having been through at least one over-expansion cycle. 
Around July of 2021 there were thoughts that MLCC’s were heading into a shortage situation and some manufacturers were beginning to plan capacity increases, especially for automotive oriented MLCCs, but the industry as a whole does not seem to have fallen into the pattern of over expanding into the demand inflection point, which gives us reason to maintain a positive long-term view of MLCC growth, despite the near-term weakness.  For component suppliers MLCCs are still one of the best places to be and will continue to see application growth over the long-term.

Microprocessors get all the press, with their billions of transistors squeezed into a tiny piece of silicon, but who does all the real work these days?  Its us, the passive components that serve and protect the microprocessor and its fancy TSVs, 3D stacking, and metal highways.  We do all the work and they get all the glory, but it hasn’t felt so good when you microprocessors have to sit on the shelf because manufacturers have to wait months to get us lowly passive components.  Yeah, we only cost a few pennies but there is usually only one microprocessor and hundreds or thousands of us passives in a system, so without us there is no system.  Okay, we are anthropomorphizing passive components, but these simple components are the traffic cops, bus drivers, and garbage collectors for many CE devices and while they seem immaterial when compared to transistors and the complex systems they can be molded into, they are as necessary to those systems as water is to humans.
The ECIA (Electrical Component Industry Association) surveys its members via an independent 3rd party (Technology Partners Consulting) on a regular basis to get a sense of both sentiment across the components industry and a read on inventory and pricing levels.  As this data is for North America, it is not conclusive for the global market but does give some major data points as to where the industry has been and is headed.  As can be seen in Figure 2, while m/m sales increases peaked in March of last year, they have remained consistently above the 100 mark, which indicates stable sales gains, however the inset shows that expectations for June, while still above 100 for individual component categories, indicates that the end markets will fall below 100 for the first time since the COVID-19 pandemic began. 
More specifically, as shown in Figure 3 and Figure 4, members were pessimistic about the prospects for sales in 1Q, relative to 2020 and got more pessimistic as the quarter developed, and while that pessimism about 2nd quarter was at a low in January, it continued to rise as 1Q developed.  With expectations for the industry expected to fall below 100 in June, it would seem that a positive view of the industry from its members has been replaced with a negative one, although individual component segments, while down in June, have remained above the 100 mark, sort of a “…well the industry is crappy, but we’re still doing o.k….”, with ~90% of respondents indicating they see a book-to-bill of 1 or greater at their company.
We see the worry coming from the fact that while general component inventories are all relatively close to normal for most categories, concern that inventory levels will continue to grow in 2Q is the case for ~45% of respondents, while only 8% expect those levels to decline.  These metrics are actually down from those heading into 1Q, but component inventory levels were artificially low early this year as supply chain issues and high raw material prices limited the supply of a number of components, both active and passive.  Since then supply pressure has eased a bit and component suppliers have tried to normalized shipments, but are concerned that seasonal factors will make it necessary to build inventory further without underlying demand.  
All in, component suppliers remain pessimistic about the state of the industry but seem to be more optimistic about their individual segments.  In most cases raw material and energy prices continue to rise and with component inventories at normal levels, it becomes more difficult to pass on price increases, so the carrying cost of potentially higher inventories becomes more of a burden in 2Q and potentially 3Q, while demand returns to more modest pre-COVID 19 levels.  We believe this is the basis for the industry pessimism seen above but with sales remaining above the 100 line for most component segments each seems to have a more positive view of their own segment than the industry overall.

​ToF (Time of Flight) sensors were a hot topic two years ago, and we admit we were certainly intrigued by the ability of these devices to create an image map that allows for accurate depth resolution that can be used by imaging applications to make adjustments to focus based on object location.  LG Electronics was the first to incorporate such sensors in 2014 as part of the LG G3 smartphone, but the first well-known application for ToF in CE products was the Microsoft (MSFT) Kinect motion sensor used in conjunction with Xbox consoles back in 2010, which was a major hit for the company from a publicity perspective.  The real kicker came when Apple adopted the technology in the iPhone, and dozens of smartphone brands have adopted the technology to assist camera applications.
The problem came when Samsung Electronics revealed that the Galaxy Note 20 would not contain a ToF sensor and the Galaxy S21 would also abandon the technology, with Samsung saying it lacked a ‘clear advantage’ for the consumer, which took much of the ToF momentum out of the market.  There was a good reason why Samsung did not focus its attention on ToF, and that has to do with a decision the company made as to the type of ToF sensor it would produce years before, but leave us say that it was not based on a lack of consumer interest but on a profit motive for Samsung.  While not carrying the cache of Samsung’s marketing, ToF sensors are still found in a wide variety of smartphones by Apple, Sony (SNE), Honor (pvt), Motorola (MSI), Huawei (pvt) and others, but are more the focus of those in the automotive space, tying them in to obstacle recognition systems, but it seems they have also found a new home.
Infineon (IFX.GR) and PMD Technologies (pvt) have developed a 3D depth sensing ToF system that will be used in the Magic Leap 2 (pvt), an AR headset designed for enterprise customers that is due out later this year, which will ‘capture the physical environment around the user’ to give the device an understanding of how to interact with the environment.  The system will give the device accurate image coordinates and mapping of hands, faces, and objects in real-time, and is able to work in both bright light and complete darkness.  The sensor chip itself has been optimized to reduce heat and power consumption to maintain battery life.
One feature of the Magic Leap 2 AR headset is its ability to reduce the ambient light surrounding a projected image without changing the image itself.  This allows for better focus on the AR image but also makes it necessary to maintain some perspective on distances without the light that would normally illuminate other objects.  Given that ToF is UV laser based, the lack of light does not affect its ranging capabilities and can provide the necessary depth information to keep users from losing track of where they are relative to the projected image, almost converting the device from an AR to a VR headset.
As a specialty device, especially one geared toward business use, the Magic Leap 2 will not push ToF sensors back into the limelight, but can bring ToF back into the CE nomenclature.  While there will be other silicon based features in the Magic Leap 2 and other AR/VR headsets, the fact that an AR headset can be ‘dimmed’ without losing the depth perception necessary for safe use, is a step toward expanding the use of ToF sensors in the CE space, and we expect over time to see ToF sensors become an integral part of the XR world.

While we don’t often mention coal in our CE notes, it is the globe’s primary source of material for energy production and recent issues facing China and a number of other countries that have begun to limit the exportation of coal to prevent shortfalls at local power-generating facilities, have brought the material to the forefront in recent months.  In particular, Indonesia, the largest producer of thermal coal, issued a ban on coal exports, which began on January 1, when the country’s mines failed to meet a government requirement that 25% of their annual production must go to the state-owned electric company PT PLN at a maximum price of $70/ton.  While the Indonesian government indicated that it is also reviewing export restrictions on other mineral resources, with indications that copper and bauxite (85% of bauxite is converted to aluminum) are among the one being considered, no determination on these materials has been noted thus far, with Indonesia being the world’s 6th largest copper exporter in 2020 with a 4.2% share and the 5th largest bauxite exporter last year.
But the good news is that the country has started easing the coal restrictions for those local companies that have met their domestic market obligations, allowing 75 ships to load coal from complying companies while 12 more ships have been allowed to load under written assurances of compliance and penalties from others being closely monitored.  The coal stockpiles at Indonesian power generation plants has now reached 20 days, which the government finds is enough to maintain consistent operation.  The original ban pulled the mining licenses of over 2,000 producers, many of whom had been lax about meeting government imposed requirements, but the release of the ban, at least in part, is a welcome relief for a number of countries that are large coal importers and rely on Indonesia to supply same, such as the ones shown below.

According to Trendforce, while NAND Flash prices are expected to drop in 1Q, they have revised their original forecasts upward.  Citing the lockdown of Samsung’s (005930.KS) and Micron’s (MU) lines in Xian, China due to a CVID-19 outbreak and strength from PCIe 3.0 products (PCI or Peripheral Component Interconnect Express, a motherboard interface for PCs, graphics cards, Wi-Fi boards, and disk drives), causing delivery times to extend.  In order to shorten those delivery times, customers are more willing to accept less of a price decline.  The same lockdown issues have pushed SSD customers to step up orders, while lower shipment levels of Intel’s (INTC) Alder Lake CPU, which supports PCIe 4.0, has shifted demand back to PCI3 3.0, lessening the decline in SSD prices.  The change across total NAND Flash pricing, while still down 5.5% (single point) for the quarter, represents a 26.6% increase in expected pricing.

In our 10-28-21 note, we indicated that, among other countries, France took umbrage at Apple’s removal of chargers and earphones from the box when purchasing an iPhone.  The government of Brazil fined Apple for the change considering such devices as ‘defective’, and France cited the potential harmful effects of cell phone radiation on young children, forcing Apple to include headphones within the iPhone packaging.  Apple’s theory that most users already have chargers and headphones, leading the inclusion of same to product waste harmful to the environment, is a bit dubious, especially during a period when inflation is causing many components to become more expensive, and goes to the same but more subtle issue of how many consumer staple companies have reduced product weight/package to keep price consistent. (Example:  Sweet & low packets used to be 0.04 grams but are now 0.035g, a 12.5% reduction in weight).
That said, France has changed its mind toward Apple and has decided to overturn the law requiring headphone inclusion, now citing the same ‘waste’ issue that Apple’s lawyers had presented, and that the health argument is not enough to justify the additional waste, although it still requires that ‘compatible headphones are available for sale during the life of the product.’  According to Apple related sites in France the company has already sent a memo to carriers in France indicating that it will stop shipping iPhones with headphones after 1/24/21.  There is still the potential legislation in the EU that all smartphones must have a C-type port to connect to a charger, which would eliminate the necessity for a separate charger for Apple smartphones, but if it were that easy to reverse the headphone issue in France, the EU mandate might never see the light of day.

Wi-Fi is confusing.  In the ‘old’ days, Wi-Fi versions were called ‘802.11’ followed by a letter, which was usually shortened to just the letter, such as Wi-Fib or Wi-Fin.  However, since the specifications for Wi-Fi are developed by engineers, there were lots of ‘in-between’ Wi-Fi specifications that kept Wi-Fi designations from being sequential.  The Wi-Fi Alliance, the organization responsible for such specifications decided to convert those older designations into a simple form as shown below, hopefully simplifying things for consumers, who have had to deal with trying to understand the compatibility between their router and their smartphone when the router says “compatible with 802.11 a,b,g,n” and their phone says “Wi-Fi 802.11 a/b/g/n/ac).  Unfortunately things are still a bit opaque as the current version of Wi-Fi is 802.116e, which doesn’t have a new and simple designation, making the name changeover a bit messy, but don’t worry because Wi-Fi chip supplier Mediatek (2454.TT) has just shown a demo of its new Wi-Fi 7 chip, which it expects to release in 2023, cutting short the life of Wi-Fi 6e.
The good news is that specifications for Wi-Fi 7 are almost completed, which means that we should start to see such demos from other Wi-Fi chip suppliers later this year, paving the way for release next year and widespread use in 2024.  Not only does this simplify the Wi-Fi nomenclature, but the specifications for Wi-Fi 7 are considerably better than those for Wi-Fi 6e, and will allow for data speeds up to 46Gbps, which would mean you could download a 25Gb Blu-ray movie file in a few seconds, and AR/VR headset Wi-Fi connections will be able to take advantage of those higher connection speeds[1].
Bu wait, there is more…  Aside from the faster connection speed, Wi-Fi 7 will provide for the use of 16 antennae, twice that of Wi-Fi 6, which means a router should be able to handle more devices with less bandwidth (speed) sharing and interruptions, with the possibility that final specifications will allow for all three frequency bands to be used by such devices at the same time, further helping any bandwidth contention, all of which could put Wi-Fi speeds ahead or wired connections for the first time.
All in, Wi-Fi 6/6e will have a shorter life than originally expected, as Wi-Fi 7 is moving quickly toward commercialization.  While still two years out for the mass market, the incremental change will reduce some of the limitations that home VR systems face today, and while there are still many mountains to climb before users will no longer have to spend all of their time in the real world, its one less brick in the wall.


[1] Connection Speeds are theoretical as actual connection speeds will be lower.
 

On 12-10-21 and 01-07-22 we updated readers on the status of one of the more important components in current CE products, MLCCs or Multi-Layer Ceramic Capacitors and back in August of last year we noted that Murata (6981.JP), the world’s largest producer of MLCCs (between 31% and 45% share depending on inclusions) had shut down its plant in Takebu, Japan for a week due to COVID-19.   It seems that once again Murata has been plagued with COVID-19, but this time the plant, which accounts for ~20.7% of the company’s capacity has been only partially affected, given some measures taken after the shutdown last year.  Expectations are that the partial shutdown will affect MLCC production for some high-end smartphones and servers but current inventory is able to make up for the shortfall.
While MLCCs are typically quite small, ranging from ~0.8” x 0.6” to 0.01” x 0.005” (.254mm x .127mm) they are integral to many CE devices, particularly smartphones, and even more so 5G smartphones.  Murata’s production capacity is primarily in Japan (56%), China (36%), Singapore (3%), and the Philippines (5%), although production in the Philippines is primarily for the automotive space.  Singapore is focused on high-end specification MLCCs for smartphones, while China is focused more on consumer oriented (lower spec) MLCC production, leaving Japan as the primary for servers, 5G base stations, notebooks, and additional automotive production.  With the recent outbreak of the COVID-19 Omicron variant, MLCC production at Murata is again at risk, with Japan and the Philippines the current hotspots for the company.  With the relatively high price of MLCCs and moderate inventory levels, we expect little additional pressure on the MLCC market as the Omicron surge seems to be relatively short-lived thus far.