The State Internet Information Office in China has released a provision relating to the use of “Deep Synthesis of Internet Information”, which more loosely translated would be the use of AI in media services of any type on the internet and in other broadcast venues.  What this refers to is the use of artificial characters that might look like actual persons or are life-like.  This has become a relatively common practice in China, where artificial avatars, who mimic existing media personalities, are used to report the news, or make public commentary.
The new rules force those using such forms of information dissemination to visibly identify them as such, and to monitor the information that is synthesized by AI-based news processors.  In many cases AI is used to sift through large volumes of news from a wide variety of sources and synthesize a short summary of a story that seems to be of consequence.  This is done through the AI system, which looks for recognizable patterns in news flow, and puts together a composite of those stories that it or others have selected by combining source information.
The Chinese government is attempting to move the responsibility for accurate reporting to the media source, making sure that the media owner knows he or she is responsible for making sure the information disseminated is accurate and legal, in that it does not violate existing laws that “endanger national security and interests, damage national image, infringe on social public interests, and disrupt the economy”.  While this seems to be an attempt to ‘keep things real’ by making sure AI-based information is disclosed as such and is legitimate, it also makes clear that the responsibility for meeting all existing laws, some of which are obviously censorship related, are followed, and that blaming misinformation or illegal (meaning different from state rhetoric) information on the AI is not a valid excuse.  Owners of the distribution source must keep concise records of any issues relating to information sources and log and report any discoveries of incorrect or illegal material generated by the AI system, but the rules state that the responsibility to meet applicable laws are the responsibility of the owner, no matter what the AI might broadcast, and that the public must know when Ai generated information is being used.
“For the public good” shows up a number of times in the new regulations, but the rules also seem to have a repressive tone, seemingly extending the government’s ability to suspend, terminate, or prosecute the accounts of those who might find themselves at odds with the state.  Given the recent protests in China, it seems that this is another way to plug up any holes in the media censorship that exists in China, and while it has technology related overtones, it is really a way of reminding the media that it lives under strict rules and technological mistakes or omissions will result in the same repressive results that would occur in any other form of media that exists on the Mainland.
Examples of AI Generated ‘Digital Twins’ are shown below courtesy of Deep Brain AI (pvt), a South Korean firm that is considered the preeminent provider of such technology.
https://renew.deepbrainai.io/renewal_resources/videos/Use+cases+%26+Industries/Media_4_CCTV.mp4
https://renew.deepbrainai.io/renewal_resources/videos/Use+cases+%26+Industries/Media_2_LG.mp4
​

Semi.org is a large trade organization whose mission is to “…advance the growth and prosperity of our member companies’ ecosystems…”, sponsoring a wide variety of committees, expositions, market research, and influencing public policy.  As part of their research, they forecast a number of semiconductor categories, some of which we have highlighted in notes earlier this year (5/4/22, 11/07/22).  This week Semi introduced its forecasts for semiconductor equipment spending, which indicated a strong year in 2022, ($108.5b) up 5.9% y/y[1], but a decline of 15.9% in 2023 ($91.2b), albeit with a positive 2024, up 17.5% from the 2023 forecast, although even with the substantial spending gain in 2024, the absolute spend ($107.16) in 2024 is expected to be lower than this year’s record number. 
That said, we note also that the previous forecasts for 2022 and 2023, which were made in September of this year were $117.5b and $120.8b, which puts the new estimates down considerably from the previous forecasts, which were up 14.5% y/y for 2022 and up 2.8% y/y for 2023.  Things certainly have changed over the last few months for the semiconductor business, which has gone from shortages across numerous, if not most categories, to shortening lead times and lower utilization  in many, so we do not fault SEMI.org for the revisions, more the industry itself for assuming that the rules of the business had changed and capacity expansion will solve all supply chain issues.  With a number of regional incentive packages being offered to chip producers, construction was started on 23 new fabs last year, 33 this year, with 28 expected next year, and while China will lead the way with 20, the America’s represent 18 and Europe 17, with Taiwan in 4th place at 14. 
While each fab has its focus, it is easy to see how Semi derives the upswing in equipment spending in 2024, but little has been said about how closely these fabs will map with demand, a far more relevant question, and if we use the display industry as a guide, an accurate match-up would likely not be the case.  It is hard to separate potential business prospects for each semiconductor producer from the overall effect each capacity addition has on the industry, but again, using the display space as an example, semiconductor producers are less interested in the overall industry than they are in what they report to shareholders, so if we had to speculate on a broad basis, we would expect one of two scenarios in 2024/2025, one being a over-capacity in many semiconductor segments, and the other, more rational one, being less spending than forecast in 2024, as some of that capacity is pushed out.
Perhaps the semiconductor industry is a bit more rational than we are used to, and maybe we are a bit more negative after hearing nothing but massive optimism toward semiconductor growth as a solution to the shortages the semiconductor industry faced earlier this year.  However, human nature being what it is, and business being what it is, such unbridled capacity growth carries substantial risk to the semiconductor apace and therefore consumer electronics in general.  In the long-term, such near-term growth will likely find the demand it needs, but we expect it will be a bit of a bumpy ride


[1] Given the previous three quarters worth of data, the 2022 forecast assumes semiconductor equipment spending of $28.64b, down 0.4% q/q but up 4.6% y/y

​Senator Marco Rubio (R-FL) has introduced bi-partisan legislation to ban TikTok (pvt) from operating in the US.  The bill titled “Averting the National Threat of Internet Surveillance, Oppressive Censorship and Influence, and Algorithmic Learning by the Chinese Communist Party Act”, has a stated purpose to “protect Americans from the threat posed by certain foreign adversaries using current or potential future social media companies that those foreign adversaries control to surveil Americans, learn sensitive data about Americans, or spread influence campaigns, propaganda, and censorship”, using the presidential powers of the International Emergency Economic Powers Act.   The bill proposes that all transactions of Bytedance (pvt) and TikTok be blocked, preventing commercial operation in the US or with a US person (intelligence activities are excluded) and covers potentially similar threats from Hong Kong, Macau, Russia, Iran, North Korea, Cuba and Venezuela that would qualify as social media companies.
There has been talk of such proposals a number of times over the last few years, with little support, and certainly continued resistance from the ~80m active TikTok users in the US, and with the controversial takeover of Twitter (pvt) by E. Musk, such a ban would garner even more focus from users, who are obviously not concerned about their personal data being harvested by nefarious governments.  If you need any reinforcement of idea that TikTok users are certainly not afraid to reveal their innermost secrets or reveal their knowledge level (or lack thereof), the video below from TikTok user SSSniperWolf, a 30-year-old actress (aka Alia Marie Shelesh) who has posted over 3,000 videos and has over 33 million subscribers and over 20 billion (billion!) views.  A few minutes will reveal the ‘sensitive data’ that is of concern to the esteemed senator from Florida.
https://youtu.be/FJXD5zMRUhM
 

Back a bit over a year ago, Samsung Electronics (005930.KS) announced that they had chosen a site in Taylor, Texas for their new 6 million square foot semiconductor fab that is expected to have a $17b price tag.  As shown in Figure 3, the site was essentially open farmland, although Figure 4 shows the current status of the site work.  Samsung has just begun to order basic cleanroom equipment from a number of South Korean suppliers, with deliveries scheduled for January of 2023.  The equipment is thought to be specialized air conditioning and filtering units that are external to the cleanroom, which can be placed before construction is completed.
The town of Taylor has also approved the addition of 9 new applications from Samsung for additional facilities likely to be built in the future, expanding the tax rebate zone to 1,268 acres, which would allow for additional fabs to be built on the site without additional tax abatement approval.  The town has also approved a $271m application from Linde (LIN), which will provide co-located industrial gases used at the plant.  Samsung has been offering an internship program to the town’s independent school system, donated $1m to 4 local non-profits, and is sponsoring job fairs to residents.  The plant is expected to need ~2,000 workers once it begins full-scale operation, and while we expect senior management will come from Samsung in Korea and from Samsung’s other fabs in Texas, the job picture for the region seems positive, especially as suppliers commit to local facilities.
The ground-breaking ceremony, which was expected earlier this year, has still not taken place, site work has continued and foundations are being laid, with the initial production scheduled to begin in the 2nd half of 2024, although we expect things are a bit behind schedule.  That said, these early equipment orders bode well for clean room construction to begin early next year, and signal that Samsung is fully behind the project and the timeline is progressing although there is still considerable equipment that has yet to be ordered, which leads us to expect a few months might be added to the production date.

​AU Optronics (2409.TT) has said it is considering building a module/assembly plant in the US.  The company, among the top 5 panel producers globally, produces panels in Taiwan and China, is expected to make a decision in 2023 at the earliest, with mass production scheduled for 2025 if approved.  Other than a sales office in Farmington, MI, a bit outside of Detroit, that is a point for AUO’s automotive display business, they have little representation in the US, and while their panels are in a number of TV sets under major brand names, they are not well known in the US.
While a module/assembly plant is a far cry from a display panel fab, it would be unusual for a panel producer to build such a facility in the US given its high cost of labor.  Typically North America is served by assembly plants in Mexico or Brazil, so such plans would be out of the ordinary, but given the incentives the US has put on bringing foreign electronics manufacturing back to the US, the net result could make such a venture a bit more competitive, especially given the high cost of transportation CE producers have faced recently.  While an unusual decision, and one that could also be used to leverage negotiations with potential assembly sites in other countries, it would be nice to see tangible results from the current administration’s efforts to bring CE and semiconductor manufacturing to the US.

​LG Display (LPL) has decided to end TV panel production in South Korea by winding down production at its P7 LCD fab , producing only to fulfill long-term existing contracts.  LGD is also expected to reduce LCD TV panel production at it fabs in Guangzhou, China from 210,000 sheets/month to 120,000 sheets/monthly, leaving that fab as the only one that will continue to produce LCD panels for the TV space. LGD’s P8 fab, which is a Gen 8 fab will produce LCD panels for IT displays (notebook, monitor, and tablets).
LG Display has expressed its intentions to reduce or eliminate LCD TV panel production in South Korea in the past, however the rise of TV panel prices during the COVID-19 pandemic in 2020 and 2021 , gave them reason to postpone those changes, likely with encouragement from parent LG Electronics (066570.KS), who, along with most other TV set manufacturers faced short supply from other sources.  Since last July the price of LCD TV panels has declined precipitously, making the decision a bit of an assumption.
Both Samsung Display and LG Display have been facing aggressive competition from Chinese panel producers, who are able to produce generic LCD TV panels less expensively due to lower labor costs and government subsidies.  In recent quarters, LG Display, along with other non-Chinese panel producers, have shifted away from direct competition with the Chinese and tried to change their mix toward higher margin more specialized products.  While this has offered a bit of protection from competition, weakness in the overall display market has lessened the benefits of those changes and pushed LGD and potentially others to restructure fabs in order to maintain a profitable mix.
Some have moved toward OLED, which we expect will be the case for LG Display in the future, converting P7 to either an OLED IT panel fab or an OLED TV panel fab, but given the current macro circumstances, such decisions can wait for a quarter or so to see how the holidays pan out and what the longer-term picture might be.  While it is expensive to have an idle fab of the size of P7, it is certainly less expensive than running it at low utilization rates or producing at cash costs or below, so the decision was imminent, although we will have to wait until January to get more information on how quickly the implementation will take place, the cost, and if any plans for the future of the fab have been made.

​While we have mentioned the issues that Apple is facing in the EU in relation to the fact that Apple uses a proprietary charging and data transfer connector/system, the reality is that those issues will likely change the way Apple connects to the outside world globally.  With a hard deadline of 12/27/24, Apple will then be required to support the USB-C standard in the EU and considering the additional cost of producing some Apple devices with USB-C and others with Apple’s Lightning connector, we expect the change will be made for all Apple devices. Apple already uses the USB-C connector on its MacBooks and iPads, so the concept is not new to the company, but the iPhone remains the target for the EU, which will have to change its connector by the deadline.
 
There have been rumors on a regular basis since 2020 that Apple will shift some (premium likely) models to USB-C before the deadline, however there seems to be no rush on Apple’s part to make the change.  With Europe overall averaging  ~24.1% of Apple’s total revenue, second only to the Americas (46.3%), over the last two years, Apple can ill afford to continue to fight the change, but there is another way Apple can avoid the issue, even after the deadline date.  Apple does have the option of going wireless, meaning that both charging and data transfer would be done without a wired connector, and would thus avoid the controversy.  This would be a far more radical step and could have implications for those with older iPhones who want to preserve data when transferring to a new iPhone, but it is certainly not out of the realm of possibility for Apple, and would likely be a positive longer-term solution given that the concept of wireless charging has been developing since 2012, and over 1B smartphones(installed base) have wireless capabilities. 
As of the end of last year 22% of smartphones in Europe had wireless charging capabilities (21% in North America), and while the elimination of any other means of charging would be a big step, it is not as radical as it might seem.  We have no knowledge that Apple is pursuing that path, so we cannot make such an assumption, but we would expect that the general direction would be to eliminate any connector over the long run, if for no other reason than to reduce costs and create space for other hardware features.

​Sanan Optoelectronics (600703.CH) completed a private equity sale of $1.13b US to be used to increase the capital of the company’s Mini/Micro LED project in Hubei province.  The project has turned profitable according to the company as of September of this year on a net basis and the company feels the expansion is warranted.  The shares were purchased by a number of institutions, some state-owned, and are locked in for 6 months.  The project was originally announced in 2019 with a 3-year construction window and a production goal of one year after construction was completed, although there have been reports that Sanan had been producing production samples in March of this year.
Sanan has signed agreements with Samsung (005930.KS) for the supply of Mini-LEDs, likely being used for watch displays, a JV with Chinastar (pvt) for Mini-Micro-LED R&D, and has been  approved as a supplier to Apple (AAPL), so the company has a legitimate reason for raising capital to expand production, at least from a customer perspective, although Apple has also been working with Epistar (3714.TT) and automotive lighting specialist AMS OSRAM (pvt) for a number of years on various Mini-LED and Micro-LED projects.  We expect the latest capital raise is Sanan’s push to become the dominant volume supplier to Apple, and Apple is certainly one to encourage additional capacity investment from suppliers.
The Mini-LED market is a relatively new one and is still evolving and while we expect the Mini-LED market, which is an adjunct to LCD displays, to grow steadily over the next few years particularly for high-end LCD displays, we see such vast differences in estimates and forecasts for the Mini-LED space (estimates range from $39m to $174m for 2020 and from $81m to $411m for 2021), that it seems a bit problematic when it comes to evaluate such diverse forecasts without knowing what is included.  The same goes for longer-term forecasts, which vary even more greatly out in the 2027 – 2030 timeframe, and the prospects for Micro-LED, which, in theory, would compete with LCD and self-emissive displays, are even more tenuous.  Sanan, already a major LED producer, is among other Chinese LED producers who have been investing is Mini-LED capacity this year, including $717m from Jiangxi Zhaochi Semiconductor (pvt), a similar capacity increase from Xiamen Changelight (300102.CH) and a 289m investment in HC Semitek (300323.CH) by China’s leading LCD display producer BOE (200725.CH), with these firms also looking to become major Mini-LED suppliers as the market develops.

​A recently proposed amendment to the US National Defense Authorization Act, that would have prohibited Federal agencies from purchasing semiconductors from three Chinese companies, has been scaled back after protests from trade groups and the US Chamber of Commerce shed light on the difficulties and costs associated with the increased restrictions.  The bill would have implemented almost immediate prohibitions of agencies purchasing semiconductors from Semiconductor Manufacturing International, aka SMIC (688981.CH), Changxin Memory Technologies (pvt), and Yangtze Memory Technologies, aka YMTC (state), with two of the three having been named by US officials as linked to the Chinese military.
The trade groups indicated that the cost of dissecting every device purchased by US government agencies to see if they contained semiconductors from the three target companies, would be enormous and incredibly time-consuming, while being virtually impossible within the timeframes originally suggested in the amendment.  This seems to have helped to change the legislation from almost immediate implementation to a 5 year phase-in plan, in keeping with the funding plan of the CHIPS & Science Act.  Th revised bill also calls for an initiative with the private sector to identify and replace Chinese semiconductors across the federal supply chain, rather than leaving that burden solely on federal agencies, with the goal of replacing all Chinese-made semiconductor products with American made semiconductors to avoid global chip shortages.
All in, we have never been fans of trying to eliminate a competitor through trade sanctions that are politically based, but there is certainly some merit to using those sanctions to avoid funding adversaries that are producing devices or electronics for an aggressive military power.  That said, it is a very fine line to walk without hurting the overall global effort toward developing more sophisticated semiconductors, of which China is a part, so there is a balance to be found on a global scale.  On a country-wide basis, the US is certainly capable of building the infrastructure needed to become independent of China’s manufacturing prowess, but it does have a cost, as became apparent when the US government began forcing carriers to remove telecom equipment made by Huawei (pvt), which has yet to be completed and is considerably over original cost expectations. 
The bill promotes the idea that US jobs will be created by the expanded semiconductor ban, which is likely correct, but those gains must be balanced against the cost of implementation, the higher price of non-Chinese semiconductors, and the risk that the US and its allies will take on more of the cost burdens during the nadir of such a cyclical industry.  Its easy for politicians to use anti-China sentiment to stir up potential voters but such actions have deeper and longer lasting implications than can be expressed in a two page bill summary, and we expect that the political world will quickly move on to another issue and leave the global semiconductor industry to deal with this new disruption.