One doesn’t usually think of the washing machine business as a dog-eat-dog market but the global household laundry equipment market is a ~$101b market, along with the inevitable ‘smart’ washing/drying’ category driving much of the growth.  The global market is fragmented with China’s Haier (600690.CH) the leader with a ~20% share, followed by Whirlpool (WHR) (14%), but the ‘other’ category is a big one at over 30%, pointing to a number of smaller brands that are popular regionally.  LG Electronics (066570.KS) has a lesser share (~6%) and South Korean rival Samsung Electronics is part of the ‘other category’, although in the US both LG and Samsung are popular contenders given their focus on ‘smart’ versions of washer/dryer products.
A few years back the former administration instituted trade restrictions on South Korean washers and dryers to help US brands (Whirlpool) to maintain a leading market share, and while those restrictions have been lifted, both the US and global washer/dryer markets remain quite competitive.  Samsung is adept at high volume manufacturing and cost containment in the consumer electronics space and has tried to bring that expertise to their consumer appliance business, however Samsung’s typical approach of pitting suppliers and technologies against each other to lower costs seems to have backfired when it came to their washer business.
Last year Samsung received more JD Power awards for its washers and dryers than any other brand, however they were forced to recall 86.2% of their Bespoke Grande AI front loading washers that they produced between September 2021 and May of this year due to defects in the washer door.  Those defects were allegedly caused by Samsung’s attempt to cut costs by changing suppliers and the formulation for the adhesive that seals the glass to the washer door.  Not only did the company have to issue an expensive recall, but one of its division presidents was called before the Korean Nation Assembly to explain why the defect occurred.
It seems that Samsung’s traditional adhesive supplier, Henkel (HEN.XE) found itself in competition with HB Fuller (FUL), who was providing Samsung with an adhesive that was 25% less expensive than Henkel’s silicon enhanced adhesive.  Fuller’s adhesive was less expensive because it was based on a polyurethane base, different from most washer adhesives, but to Samsung it represented a bit more margin or at least an offset to some of the raw material costs seen over the last two years.  Unfortunately the polyurethane adhesive turned out to be less resistant to heat and other chemicals in detergents and was said to be the cause of the leaky doors, the recall, and the embarrassing testimony. 

On 6/7/22 we noted that Austin Li Jiaqi, aka the ‘Lipstick King’, an unbelievably popular influencer in China saw his broadcast on Taobao Live go blank after he brought out a plate of Viennetta ice cream with Oreos and chocolate, which some say looked like a military tank.  While Austin noted in a post that there had been some technical difficulties, he did not appear two days later for his next show and has disappeared for over three months.  There has been considerable speculation on Chinese social media concerning the reason for the cancelled broadcast and his disappearance, with the most popular explanation being that the ice cream looked too much like a reference to Tiananmen Square, a topic that is forbidden in any form.
It seems that with little fanfare the Lipstick King returned to his live streaming platform on Tuesday at 7PM, showing a variety of cosmetics, skincare products, and underwear.  According to media statistics, the show began with ~100,000 viewers but an hour later was topping 25m, with over 50m by 9PM.  Given his high profile and his disappearance, most of the products he was hawking sold out immediately causing him to ask for both buyers and sellers to remain calm, asking viewers not to buy just to support his program but only if the product is needed, an unusual statement for an influencer.
All of that said, he was not the only high-level influencer to disappear since last year when the Chinese government began investigations into ‘non-state’ influencers, citing potential irregularities in content and taxation.  China’s most bankable influencer Viya, disappeared after she was handed a $201m fine for tax evasion last December and has yet to reappear.  Who says being an influencer is an easy job!

​$5.15m to move the battery technology it has developed from the lab to commercial production over the next few years.  While there are literally hundreds of companies developing automotive batteries to feed the growing hybrid and electric automotive industries, there is a bit more promise in the technology developed here than in standard lithium ion batteries that are under development currently. 
While still in lad scale, the technology, which uses lithium itself, rather than lithium ion material to create an electrical charge, is able to last through 5,000 to 10,000 charge cycles as opposed to typical lithium based batteries which last between 2,000 and 3,000 charging cycles.  Given that 37% of the US population does not have a garage in which to charge an electric vehicle, the industry must rely on external charging stations, which take hours for a full charge. 
The prototype of Adden Energy’s solid state lithium battery takes only a few minutes to charge, essentially the time it would take for a fill up at a typical fossil-fuel gas station.  Lithium metal battery technology has been and continues to be a subject of considerable research given it superior energy density and capacity, but there is a problem with the metal as it is repeatedly charged.  When metallic lithium is charged over time it forms dendrites, small structures that ‘grow’ from the anode of the battery to the cathode, essentially shorting out the device or even causing it to overheat. 
While not stopping the growth of dendrites, Adden researchers have found a way to keep them from growing by inserting less stable formulations of lithium between stable versions, sort of adding condiments to a sndwich, and while this seems counterintuitive, it makes the entire package more stable.  The dendrites from the stable lithium (aka LPSCl) still grow, even though the next layer made of graphite, but are unable to grow through the less stable lithium electrolyte and do not reach the cathode and short the battery.  Of course, there is considerable work to be done to scale the process to the larger sizes needed for commercialization although thus far no scaling issues have been found and the recent funding will give the company the financing to begin the process, which will hopefully lead to the refinement of the process and a step forward for automotive battery technology.

The White House issued an executive order today that references the Committee on Foreign Investment in the US (CFIUS) review of transactions between US companies and foreign, specifically those transactions that involve the purchase of company assets and IP that might be considered sensitive to US security or its place in the world economy.  As ‘foreign adversaries’ are the focus, which translates to China, Russia, Iran, and a few others, what the order is considering is the potential for transactions that could affect the US supply chain, essentially those that could allow a foreign power to place restrictions on or limit completely access to assets that the US deems necessary.
Noted are technology, which certainly has been a focus, manufacturing, services, and mineral resources, and how they relate to national security, with the subset of microelectronics, AI, biotechnology and bio-manufacturing, quantum computing, advanced clean energy (including battery, storage and hydrogen), critical materials such as Lithium and rare earths, and food security.  While the EO does not change the committee’s powers, the order reinforces the mandate of the committee to protect the US when it comes to transactions by said ‘foreign adversaries’ or 3rd parties that might give access to those adversaries.  It further warned of the potential for ‘incremental investments’ that, over time, could lead to control over a sector or technology, essentially a series of transactions that might look disparate individually but could, in the aggregate, cause the US to lose control over resources or technological leadership.
Outside of the reminders on the risk of transactions made by foreign entities, the EO also focuses on cybersecurity risks, that is, allowing foreign entities access to systems and data that could be used maliciously.  US elections were referenced, as was critical infrastructure and defense, including the personal data of US citizens, especially identification and health data, citing the potential for the use of that data to target various US population sub-sectors, tacitly referring to data that would help a foreign entity to influence or incite a particular group that it believes would be more receptive to such disruptive influences.
The EO doesn’t actually change anything but it paints a more aggressive picture of the current administrations face on China as the mid-term elections approach, while making almost any transaction that falls to CFIUS one in which they could cite a threat to national security.  Should a Chinese entity want to buy a US farming firm, CFIUS could cite a potential threat to the US food supply years down the road as a reason to disallow, and any transaction that involves personal data could be nixed under the increased focus on data protection.  Unfortunate as these ‘reminders’ tighten the potential reign of CFIUS, it also increases the number of transactions that can be rejected, which will likely cause foreign entities to take less interest in purchasing US companies and could result in additional trade sanctions by those entities. China recently banned the purchase of certain medical equipment made outside of the Mainland and has at times restricted the export of even more basic medical supplies in response to increasing US trade sanctions or similar issues.  While the EO will not actually force a change, it is another public slap that will likely elicit a response, making the tightrope that we walk with China that much thinner.

​The Chinese government, particularly provincial and city governments in China, are maligned for providing what seems like unlimited funding for 5G projects, particularly infrastructure related funding that results in statistics that can show that China is competing in or leading global markets, and there are few countries that have a more positive attitude toward growth.  However, this also leads to some bitterness given that smaller or less financially viable countries malign China as the funding removes some of the growth barriers that limit publicly or privately funded companies that have to prove themselves a bit before early funding or must live or die in the public markets without outside support..  We admit that our attitude toward China’s growth in the display space is colored a bit by the fact that the industry was, and still is, funded by state and local governments, so we poked around to see what the US has to offer in the way of subsidies for broadband and we were surprised at the number of broadband funding projects and their size.
Here are a few[1]:
Infrastructure Investment & Jobs Act – 11/2021 - $65b for Broadband – “Access to affordable, reliable, high-speed broadband is essential to full participation in modern life in the United States.  The persistent ``digital divide'' in the United States is a barrier to the economic competitiveness of the United States and equitable distribution of essential public services, including health care and education.  The digital divide disproportionately affects communities of color, lower-income areas, and rural areas, and the benefits of broadband should be broadly enjoyed by all.  In many communities across the country, increased competition among broadband providers has the potential to offer consumers more affordable, high-quality options for broadband service”.
 
The plan funds projects that provide broadband service to locations that have no service or speeds less than 25MBs.  $100m is allocated to each state, including the US Virgin Islands, Guam, Samoa, and the Mariana Islands, with the remainder allocated on an application basis.  No more than 5% of the funds can go to planning and no more than 2% for grant administration.  All project requests must show a 5 year plan for deployment to be considered.
 
Broadband Equity, Access & Deployment Program (BEAD) - $42.45b for new Broadband equity – Each of the 50 states receive an initial allocation of $100m with the remaining funding being distributed based on coverage maps to be provided by the FCC.  Each state must submits a 5 year plan including prioritized locations.  State grants require that the applicants offer at least 100MBs speeds and is open to cooperatives, non-profits, public/private partnerships, private entities, utilities and local governments.  The program began in January of this year and by May 32 states had requested program participation.
Affordable Connectivity Program aka Emergency Broadband Benefit Program – Currently $14b – Allocates a $30 credit to qualifying families toward their internet service and $75 for those on tribal lands.  Eligible households can also receive a one-time discount of $100 toward the purchase of a laptop, desktop computer, or tablet if they contribute between $10 and $50 toward the purchase.  Eligible households must have an income at or below 200% of Federal Poverty Guidelines or participate in USDA breakfast/lunch program, SNAP, Medicaid, WIC, public housing, or are on a veteran pension. Under this program an additional $7.75b has been set aside for additional smaller projects for distance learning, telemedicine, and ‘middle-mile’ infrastructure.   In June of this year there were over 12.2m participants in the program.
American Rescue Plan Act of 2021 - $350b for all projects – States decide how much goes to broadband.  Virginia allocated $700m toward providing universal broadband access across the state, while California allocated $3 toward a $6b commitment to expand middle and last mile connections.  Late last year the US Department of the Treasury issued guidance as to how states should use money from the $10b Capital Projects Fund that was created by the ARPA, emphasizing its use for investments in fiber technology, broadband networks owned or operated by local governments or non-profits.  The awards for Louisiana, New Hampshire, Virginia and West Virginia total $582.8 million and will cover projects designed to deliver connectivity to 200,373 locations.  Louisiana has qualifying broadband coverage across only 75% of the state and the new funding will serve 88,500 locations while the funding for West Virgina will cover 20,000 locations and New Hampshire’s allocation will cover 15,000 unserved locations.  Virginia was approved to receive the entirety of its $219.8 million Capital Projects Fund allotment, which will allow it to deliver broadband to an estimated 76,873 locations.
Emergency Connectivity Fund - $7.2B – Administered by the FCC, this fund is to help schools and libraries purchase broadband devices such as routers, tablets, and hotspots or add connectivity if it is not currently available.  By May of this year, about 9 months since the initial program funding, there was only $1.5b remaining as 7,369 schools or school districts, 628 libraries and 133 related groups had applied for grants.
Of course even the most abbreviated summary of broadband funding in the US would not be complete without the December 2020 Consolidated Appropriations Act (aka Huawei (pvt) ‘Rip & Replace’), a $5b program that was the original funding for the Emergency Broadband Benefits Program and allocated $1.9b to carry out the ‘Secure & Trusted Communication Networks Act of 2019 which required carriers to remove network equipment from Huawei and ZTE (000063.CH).  By February of this year the FCC had received over $5.6b in reimbursement requests for the replacement program and the FCC indicated it will ‘prioritize’ the funding starting with service providers with 2m customers or less (Priority 1) and then move to public or private educational institutions (Priority 2), although the FCC indicated that all priority 1 requests must be covered before funds are allocated to priority 2 and lower priority requests, which means that the FCC will head back to Congress for more funding.
These are only a few of the many broadband funding programs available in the US, others like the Rural Digital Opportunity Fund ($20.4b), the USDA’s Broadband Reconnect Program ($1.5b), and a myriad of smaller ones are or have been allocating capital toward expanding the US broadband network to potential users that are outside of typical coverage zones.  In most cases these are not 5G related and based on fiber, although speed is the requirement regardless of the connection mode, so that infrastructure would be up to the carrier or government agency running the project.  All 50 states have applied for funding from at least one program although we note that some of the states have allocated funding toward creating broadband ‘offices’ as part of the state government, so not all the dollars will wind up in the ground.  All in, after even this quick look, we look back at China’s 5G funding with a slightly less biased view given the extent of funding for just these few programs.  As to whether the dollars will be spent wisely, that is fodder for another note but China is now facing a number of internal investigations as to where some of its funding programs actually went, so we would expect a similar situation here down the road if history is any teacher[2]


[1] Thanks to Diana Goovaerts @ Fierce Wireless for building the dataset.

[2] As far back as the Andrew Jackson administration (1829) political monetary corruption was a hallmark of US politics.  The 1st we could find was Samuel Swartwout, a presidential appointee to the New York City Collector’s Office, who embezzled $1.25m in customs receipts to purchase land.  He fled to Europe before prosecution.  There were many earlier scandals but none involving the theft of funds.

There are few winners in the trade wars between the US and China, with the Chinese semiconductor industry now severely limited as to its access to what is basic semiconductor technology, along with existing restrictions on advanced tools.  The recent round of restrictions, which restrict IC design software were another blow to China's chi’ advancement, however the even more recent ban by the US government on the sale of NVidia’s (NVDA) A100 and H100 to firms in China, will put a dent on China’s burgeoning AI project portfolio, and foster a growing fear in China’s AI community that restrictions will be placed on all NVidia products, which are the heart of most AI learning systems.
 
Nvidia owns almost the entire GPGPU (General Purpose Graphics Processing Unit) market, and while these devices are the heart of computer gaming systems, they are also contain a GPU that is able to perform ultra-high speed mathematical transformations that are used to mine bitcoin.  That said, more importantly is their use in systems that ‘teach’ AI devices by processing large datasets that allow for various types of matching or recognition by the AI.  As the larger the dataset, the more ‘examples’ the AI can learn from, the necessity for high speed is absolutely essential, with the Nvidia A100 delivering 312 TFLOPS (TeraFloating Point Operations per Second or 312 million decimal based simple math functions/second).
While the ban does not go into effect for US companies with data centers in China until March of next year, license requirements for direct sales to Chinese firms begins immediately.
 
While the collateral damage to Nvidia is substantial, AMD (AMD) who also faces some restrictions on its high-end GPUs says the impact will be small, but while the bans will put a dent in US company sales, it will limit China’s ability to grow its AI development, a key piece of Chinese technology strategy.  However there is a silver lining for one company, albeit not a US one, and that is Shanghai Tainshu Zhixin Semiconductor (pvt) aka Iluvatar CoreX, the only company in China able to produce a GPU that compares to the Nvidia or AMD CPUs.  The company was formed in 2013 as a joint venture between the Shanghai government and Taiwan based Via Technologies (2388.TT).  The company has raised $334m in two rounds, the most recent being $149m raised in July of this year (previous was 3.2021).  The company has developed a high speed GPU that is specifically designed as an AI ‘training’ processor and does not contain the additional hardware for gaming that is found in the Nvidia GPU system.
 
The Tiangai 100 chip, Iluvar’s first release took about four years  to design and produce  and is said to have received ~$33m in orders since its announcement at the World AI conference, has a performance rating of 147 TFLOPS (company), but things have not been all that pretty for the company up until the recent announcement.  It seems the company’s chairman, Diao Shijing, previously the co-president of Tsinghua Unigroup’s (state) semiconductor business and executive director of YMTC (pvt) (see our note of 9/12/22), have been under investigation by the Chinese government and as of the middle of last month were still ‘missing’.  The investigation centers around the bankruptcy of Tsinghua Unigroup, a company in which Taiwan based Foxconn (2354.TT) is hoping to make an $800m investment, although that investment is being investigated by the Taiwanese government due to Tsinghua’s ties to the Chinese military.
 
All in, while there continues to be considerable collateral damage on both sides as the US/China trade war continues, China continues to fund almost anything that is semiconductor related, and while still struggling under some of the more rigorous US semiconductor restrictions, seems to be making at least a bit of progress toward trying to reduce the semiconductor technology gap with the US.  The Chinese government’s strategy of developing a technological industry by funding everything and hoping that some of the spaghetti sticks to the wall can be a messy one, as much of the capital is loosely supervised, but that strategy worked with LEDs and LCD displays, so we expect more funding and a few more ‘disappearing’ semiconductor executives over the next few years.  While South Korean panel producers saw the handwriting on the wall for LCDs years ago and made the transition to OLED, ceding LCD to China, we doubt the US has the mindset to even imagine that China could one day become an equal in the semiconductor space and will fight such a battle to the last man or woman, cutting the tentacles from any Chinese semiconductor successes with new sanctions, like a Sushi chef preparing Tako.
 

The promise of global broadband is an attractive one and a worthy one from a ‘civilization’ standpoint, with the entire global community able to communicate with each other, sharing ideas, and bringing to light the richness of cultural diversity.  Of course there are some drawbacks, such as a lack of individuality as that diversity melds into a global uniformity that lacks the cultural identity that made them unique, but that is for a different discussion.  Global networks have been around for many years, starting out as telegraph communication networks based on the Morse system developed in the late 1830’s and championed by railway companies that used them to both communicate and to activate signaling systems used to prevent train collisions.  During the mid-1850’s submarine cables were made available, connecting the continents.  
Twenty-five years later analog telephones became available on a point-to-point basis but eventually morphed into an exchange based system that were connected via trunk lines to become networks and eventually was converted from analog to the digital systems used today.  IP networks opened the internet to billions of users and the GSM mobile network became even larger, expanding on the PSTN (Public Service Telephone Network).  However each network type has its drawbacks, some requiring copper or optical fiber connections between nodes, while others need many towers due to topographical issues that block signals, making it difficult or expensive for such networks to reach isolated customers.
Satellite networks offer a chance for potential connection availability for anyone with an open view of the sky, but in order for a satellite network to be fully available there has to be a vast number of satellites in what is called low-earth orbit, so there is a satellite overhead at all times.  LEO satellite networks have not proven to be economically viable for the average user and satellite networks such as Iridium (IRDM) and Globalstar (GSAT) have not quite reached the cost levels that would make them available to a large swath of the global population, which is why Elon Musk came up with the idea of Starlink (pvt), a mesh network of LEO satellites that will blanket the earth by the end of 2023.
Starlink’s almost 3,000 working LEO satellites create a high-speed, low-latency broadband network that costs $110/month (after a one-time hardware fee of $600) for regular internet service or $500/month for business users ($2,500 hardware fee).  As the satellites are in orbit at 341.75 miles above the earth, as opposed to geostationary satellites, which are ~22,236 miles above the surface, during the time it takes for a signal to reach a geostationary satellite and return (~240ms) the Starlink satellites can receive and send 70 signals, resulting in a latency of ~20ms vs. 600+ for geostationary satellites.  The antenna, which is ~13” or 24” depending on your applications is mounted on the roof, pole or wall, and is self-aiming, with the only requirement being an open view of the sky.  The antenna connects directly to a router (supplied).
The Starlink satellites travel at ~16,776 mph, orbiting the globe in between 90 and 110 minutes and are organized in three layers or shells that are at three different heights above the ground, allowing not only communication between satellites of a particular level, but also between levels.  The Starlink satellites are designed to be small, allowing a typical launch to carry between 34 and 60 satellites, leaving room for potential failures, and after 62 launches with 3,293 total satellites in the payloads, 3,025 are in orbit with 2,989 in operation.  The rest are either not working properly, have lost propulsive capabilities, have reentered the atmosphere or are in a decaying orbit for a 96.77% operational success rate and one that has improved markedly from early launches (see Figure 2).
Of course this all sounds wonderful but it is a bit more complex than the Starlink press might lead one to believe.  Mr. Musk’s has plans for as many as 42,000 satellites in order to get full global coverage (12,000 in the initial 3 shell project), which means the full roll-out will take a considerable amount of time and money.  Earlier this year one launch of 49 satellites ran into a geomagnetic storm which disabled 40 of the satellites, which are said to cost ~$500,000 each, and while the cost/unit will likely continue to decline, the cost of ~39,000 more satellites could result in less coverage or higher prices for consumers.  As to the service itself, a report earlier this year indicated that the Starlink consumer service median download speed was 104.97MB/s with upload at 12.04MB/s as compared to the median for broadband providers of 131.3MB/s download and 19.5MB/s upload, although it is faster than other satellite networks.  That said, it varies considerably by location, with Starlink ahead in a number of countries.
Given that Mr. Musk has estimated the cost to fully build out the Starlink network to require an additional $20b to $30b, aside from Mr. Musk’s personal fortune, the company is always looking for government funding to offset the cost of the system and the Rural Digital Opportunity Fund, an FCC program established in January 2020 to bring broadband to over 5m homes and businesses that have been unable to gain voice and broadband service with at least 25MB/s.  With $20.4b to be awarded over 10 years the FCC is auctioning off blocks of locations in auctions, with the winners required to meet a number of performance requirements. 
SpaceX was one of two winners (LTD Broadband (pvt) was the other) in the $9.2b auction that took place about a year and a half ago, but the FCC has issued a decision that concludes that neither winner would be able to provide the service given the requirements.  Starlink’s winning bid would have provided the company with $885.5m to provide 100MB/s/20MB/s service to 642,925 locations across 15 states and while the FFC stated “Starlink’s technology has real promise. But the question before us was whether to publicly subsidize its still developing technology for consumer broadband—which requires that users purchase a $600 dish—with nearly $900 million in universal service funds until 2032.  After careful legal, technical, and policy review, we are rejecting these applications.  Consumers deserve reliable and affordable high-speed broadband.  We must put scarce universal service dollars to their best possible use as we move into a digital future that demands ever more powerful and faster networks.  We cannot afford to subsidize ventures that are not delivering the promised speeds or are not likely to meet program requirements.”
Starlink is not going down easy and has filed an Application for Review with the FCC, essentially asking for a review of the decision on the grounds that it was flawed in both policy and law and it contained data that was off the record and ‘cherry-picked’, showing the FCC’s bias toward fiber, and that under the funding rules it is not required to meet the RDOF speed requirements until 2025.  Strangely the FCC Commissioner Brendan Carr agreed with Starlink, citing the faith that other government agencies have in the Starlink network including a recent $1.9b deal with the Air Force to provide internet service at military bases and that it would cost ~$3b to run fiber to the same areas that would be served by Starlink.  No date has been set for the review, but we expect there will be much in the way of legal fireworks before it is all sorted out, although we expect Starlink to continue its launch plans regardless of the outcome.
 

​China’s Sichuan province was hit with a magnitude 6.8 earthquake in a relatively remote area near the Tibetan border, causing 66 deaths, 250 injuries, and the evacuation of over 5,000 people from the area.  This follows the lockdown of the Sichuan Provincial capital Chengdu, a city of ~21m, last Thursday, in order to try to control a COVID-19 outbreak and the closure of many factories in August due to power blackouts caused by the record heat and lack of rain, causing hydroelectric power systems to fail to meet typical power requirements.  Despite the quake, Chengdu remains on lockdown.
There are a number of LCD and OLED fabs in Chengdu, particularly those of BOE (200725.CH) and Tianma (000050.CH).  We don’t have information on the status of these fabs due to the lockdown, however given that Chengdu is ~125 miles from the quake’s epicenter, we expect the damage to the city was relatively light, although power was lost in Chengdu and Chongqing nearby for some residents.  A quick look at display facilities in Chengdu show that ~8.6% of BOE’s LCD capacity is produced in Chengdu fabs while 18.4% of the company’s OLED capacity is produced in the local area, so it would seem that local shutdowns, if they were to or have fully closed, would be relatively minor unless sustained for an extended period.  BOE has made no public comment about either the effects of the lockdown or the earthquake to date.

“Did I do that?” was the catch-phrase of Steve Urkel, a character played by Jaleel White for the sitcom ‘Family Matters’ that ran between late 1989 and mid-1997.  As the annoying protagonist, the Urkel character typically made a mess of whatever situation in which he appeared, either by saying or doing something foolish, absurd, or stupid.  When confronted he feigned surprise and stated “Did I do that?”    Then there is the time when, after a night of bar hopping, you came home and began a rant on Twitter (TWTR) about how “…that stupid Barbara T. was a skank for talking to the guy YOU were talking to at the bar, and how she was always in EVERYBODY’S FACE, and was FAT and couldn’t get a guy even if SHE WAS NAKED!”, and then you fell asleep, only to wake up the next morning to realize that instead of Brenda T., who was the skank that you were mad at, you put ‘Barbara T. in the Tweet, with Barbara T. being your boss.
The frantic texts and calls to Twitter about deleting the Tweet came to no avail and the meeting the next day with HR and Barbara T. did not go well, leaving a mangled career path and the swearing off of all drinks with umbrellas, but the Tweet remained as a reminder that restraint is a complex emotion and not an easy one to foster, with or without alcohol, but those days are soon to be over…for a price.  It seems that Elon Musk, in his attempt to exit his $44b offer for Twitter, seems to have pushed management to get off the fence about a way for users to retract their Tweets, that is, if they are awake and if they pay $5.00/ month.
Twitter is ‘testing’ its newly developed ‘edit’ button, that would allow users of Twitter Blue, essentially a paid subscription to Twitter that is ad free and has a few ‘custom’ features, but most importantly will allow the ‘edit button to undo any Tweet for 30 minutes, giving one time to reconsider the impact of their latest Tweet.  Of course, the user needs to be conscious and coherent enough to realize, “yes, it was you who wrote that…” and not to make it worse by editing the original rant, but the subscription is a requirement and the Tweet will be marked as edited, although that seems a small price to pay to keep your friends, family, or job from firing or ghosting you forever.  But there is another catch, and that is as the app stands now, the modified Tweet will have a history attached “to help protect the integrity of the conversation and create a publicly accessable record of what was said…” according to Twitter, which would then open the original Tweet up to those taking the time to check the time-stamped history.  Perhaps once this highly requested addition is added to Twitter users will request another ‘level of editing’ that could wipe out the original (for an additional fee?) or a ‘Did I do That?’ button that holds all Tweets for 24 hours if your smartphone smells alcohol.  The possibilities are endless… 

During the last week there have been a number of stories in the trade press indicating that inventory levels ‘in certain products’ have returned to normal after brands cut orders last months, although we have seen little or no data indicating that such has occurred.  We expect there is a considerable amount of ‘promotion’ in such statements, as suppliers try to entice brands back to placing orders by hinting that others might be doing so, with much of the focus being on notebooks, which saw a shipment decline of 15.4% q/q and a decline of 13.0% y/y in 2Q, while notebook display shipments declined by 25.8% q/q and declined 22/1% y/y during the same period.
Our assumption would be that if inventory levels for notebooks are now returning to normal (at least for some brands and product types) there should be a bit more stability in the retail market as to pricing, however pricing notebooks is a difficult task as there are an almost infinite number of variations as to type (laptop, Chromebook, 2 in 1, etc.), display size (13.3”, 15.6”, 17.3”, etc.), memory, drive size, and a host of other characteristics, which makes price comparisons difficult.  That said, we went to Best Buy (BBY) and listed the ‘most popular’ notebook models and cross referenced that list against Amazon (AMZN) pricing for the same models.  If we could not find an exact model match on Amazon, we eliminated that model, but we were able to come up with a list of 10 models that were on the Best Buy ‘Best sellers’ list that were also represented on Amazon.  We note that in some cases the reason some of the Best Buy models do not appear on Amazon is that they are either too new, too old, or have an unusual configuration, which throws a bit of a question as to how Best Buy determines the ‘Best seller’ list, but that is a question for Best Buy.
All of that said, the average discount from list price for all 10 models at Best Buy was -26.7% and in comparison the average discount from the highest price on Amazon was -18.4% and the change from the lowest price on Amazon was a slight increase of 0.2$, with the detail being all but two models were at or below their lowest price on Amazon, which, as a single datapoint would indicate that recent discounting to move inventory has not quite ended.  For reference, and this was a bit of a surprise to us, all but one notebook model saw higher current pricing on Amazon than at Best Buy, with an average difference of +14.1% higher on Amazon.  We had assumed that Amazon pricing would have been equal to or lower than Best Buy, but that was not the case, although the basis for the lest was Best Buy’s ‘Best sellers’ which implies that list is attractively priced.  That said, Amazon tends to be lower priced for such items in most cases, unless the Amazon pricing comes from 3rd party sellers, rather than Amazon itself, which was the case for a number of models.  These are anecdotal data points that are, at the least, helpful in checking direction and momentum.
All in, our conclusion is that it is a bit early to be calling for a return to what  has been ‘typical’ ordering for notebook brands and CE products generally, and while we expect seasonality will help to stimulate some additional ordering as of late August early September.  Typical 3Q notebook panel shipment seasonality (5 year average) calls for an 8.3% increase q/q, although that same 5 year average for 2Q was +13.8% while the industry saw a 25.8% decline, so we only use those seasonal averages as reference.  There have been a few signs that the rapid increases in inflation might have leveled off a bit but as an incentive to consumers who deal with ‘non-seasonally adjusted’ prices for essentials, expensive CE products are likely not high on the list of things they ‘need’ to purchase over the next few months, so we see a more gradual demand recovery that could easily extend into 2023. Especially given the aggressive demand cycle the industry went through in late 2020 and early 2021 as a result of COVID. 
Bringing inventory levels back to normal is only half of the problem, with the other half being the stimulation of a new demand cycle, and that we expect will take lower prices.  Whether that comes from a disinflationary environment next year, or a desire for CE brands to dump high cost inventory before year-end (a practice not unfamiliar to many CE companies, particularly those outside of the US where the focus is on asset quality over sales growth), will determine how the holiday season plays out.  While we expect some cost relief across the CE supply chain over the next few months, we expect that the holiday season this year will be mediocre for most CE products.