​Apple’s (AAPL) ‘secret’ mini/micro-LED R&D center and prototype fab has been around for a number of years, and its affiliation with AU Optronics (AUOTY) and Ennostar (3714.TT) (formerly Epistar) is also a poorly guarded secret, but we seem to be getting closer to Apple actually releasing at least a mini-LED product as we have mentioned recently.  That said, project partner AU Optronics will be showing a number of mini/micro-LED products/prototypes at an upcoming event in Taiwan (Touch Taiwan 2021) this month that we hope will shed some light on how far along Apple’s (and AUO/Ennostar) micro-LED development has come.
AUO has been showing micro-LED prototypes since 2018 at similar shows, particularly a 21.3” micro-LED automotive display and more recently a 9.4” flexible micro-LED prototype developed with PlayNitride (pvt), but has yet to commercialize such a product.  AUO’s chairman indicated in 2019 that he expected the development of a Public Display or VR product to take 1-2 years, which would be this year, and an automotive product in 5 years.  We expect the company to show a modular public display system using small pitch LEDs in a module, similar to the systems sold by Samsung (005930.KS), and LG (066590.KS), but we have yet to get details on cost or availability, both of which could be problematical for real commercial production.

​Shortages in the semiconductor space have hit car manufacturers quite hard, as foundries must allocate processes above 40nm to their highest volume customers.  While TSM (TSM) and UMC (UMC) get much coverage in the trade press, the automotive space has a number of companies with their own foundries that are focused on producing automotive ICs and controllers.  TSM and UMC do get much of the 40nm and smaller node business, but ~98% of the global automotive micro-controller business, which makes up ~30% of the total automotive semiconductor market, is controlled by just 7 companies, add another 5 companies and that covers almost 80% of the total automotive semiconductor market.
Most interesting however is the difference between semiconductor BOM in fossil fuel vs. ‘new energy’ vehicles.  Fuel-based vehicles average ~$350 in semiconductor components, with $118 of that in power related devices, or 33.7%, while non-fuel based vehicles average ~$705 in semiconductor components, with $387 of that related to power management, or 55.0%. So not only do ‘new energy’ vehicles have a higher overall semiconductor component count, but also a much higher semiconductor BOM, roughly double traditional fuel vehicles.
Early last year automotive brands and OEMs reduced their forecasts based on the early stage of the COVID-19 pandemic.  Foundries shifted production to higher margin, high volume customers, locking in much foundry capacity for the first half of last year.  When automotive buyers began to see a bit of a recovery in demand, foundries were unable (more like unwilling) to shift production back to automotive customers, having a multitude of orders from high volume customers that were willing to pay a premium to avoid extended lead times. 
This left automotive foundry customers, particularly MCU (Micro-controllers) customers, to lead times that pushed mid-year orders into late 3Q or early 4Q for large customers and smaller customers vying against telecom (5G) and IT product production (drivers and TCONs) on 8” lines.  As competition for limited foundry capacity continued, automotive customers remained behind the manufacturing curve, leading to what is expected to be a 1m car shortfall in 1Q, a result of automotive semiconductor component shortages.  Given that demand across the semiconductor space continues to outstrip capacity, such shortfalls are expected to continue in 2Q, and a power outage at TSM’s plant in Nanke, Taiwan yesterday, puts another 30,000 wafers at risk, and while TSM’s loss will be covered by insurance, a portion of those wafers are expected to be scrapped tightening the automotive semiconductor space just a bit more..

“The US is losing its edge in semiconductors”, “China will take over the semiconductor market in a few years”, and “Electronics are not manufactured in the US anymore” are a few of the headlines that have made the press over the last few years, essentially calling the demise of the US semiconductor industry.  Whether to satisfy a political agenda or to help project funding at the state or local level, the end of the US semiconductor industry has been predicted for years by those who have the hubris to believe that they can predict the future[1].
Recent data, a part of a report marketed by IC Insights gave some clarity as to the ‘state’ of the global semiconductor industry, breaking down IC market share by region in 2020, with the data suggesting that the US still has more than a commanding lead in the IC market.  The data breaks down share further into IC manufacturers, essentially those operating wafer fabs and fabless producers, along with total IC share, and shows sales data since 1990 (Fig. 2).  With global IC sales share at 55% last year in the US, South Korea’s 21% is the closest challenger, and even when adding Korea, Taiwan, and China (APAC in Fig. 2), last year’s APAC share was still only 33%. 
The obvious focus of concern, rather than the US, would be Japan, whose declining share over the long-term leaves it a small contender currently (2020) ~6%, but China, the country that seems to be the focus of attention for politicians has the lowest overall IC sales share (5%) of any major IC manufacturing region, and regardless of the development programs, subsidies, and massive publicity campaigns, the Chinese IC industry will take many years to grow to a size that challenges the US.  Of course, many foundries and fabless companies have facilities in China, but when it comes to where the money winds up, it’s the US by far.


[1] The financial markets generally are unpredictable. So that one has to have different scenarios... The idea that you can actually predict what's going to happen contradicts my way of looking at the market. George Soros

​While taking cues from voice assistants can be a bit risky (ask those whose GPS told them to “Make a right turn now”, only to find themselves in a lake or steam), Siri seems to have given an indication that Apple (AAPL) will be scheduling an ‘event’ for April 20th.  When Siri was asked if (he.she.it) has any information as to when the next Apple event would be held, some users received a response from Siri of April 20, next Tuesday.  Not all ‘Siri’s’ seem to have this information, as iPhone and iPad ‘Siri’ does not, while Homepod ‘Siri’ does, although that seems to be somewhat different for each user.
In any case, such an event has been rumored for weeks, with speculation that the long awaited and long delayed AirTags are to be announced, along with two new iPad Pro, although rumors about AirTag releases have been around for years.  The only indication that something tag oriented might be in the works was Apple’s recent announcement that it is opening its Find My network to 3rd party producers (note 4/2/21), which could be a prelude to the actual release of Apple’s UWB AirTags, but we have been down that road before.
More important would be an announcement that Apple would be supporting mini-LED displays, regardless of the product, as the LED industry has been waiting for such an event to unleash another round of marketing about how this ‘new’ technology will improve existing LCD and QD/LCD displays and rival OLED display metrics.  Apple will have to tread a fine marketing line that does not denigrate its OLED displays while extoling the benefits of mini-LEDs.
Unfortunately, at the same time Siri has been leaking possible event dates, there have been stories that shortages in the mini-LED space have caused delays in production and even production postponements due to component shortages.  While this is easy to understand given the circumstances surrounding semiconductor foundry capacity, Apple does not necessarily have to give a release date for such products and can merely introduce the concept of mini-LEDs, or whatever ‘retinated’ name the marketing department decides to choose.  While Apple is certainly not the only potential supporter of mini-LED technology, it never hurts to have another high profile CE company jump on the bandwagon and add a few dollars to the mini-LED marketing pot.

Samsung has announced plans to build a semiconductor plant in the US and is evaluating sites for what will be a ~$17b project that will encompass ~1,900 jobs.  Under consideration are three sites, one in Austin, Texas where Samsung already has a fab, another in Phoenix, Arizona, and a 3rd in Genesee County in Upstate New York.  While the Austin location was considered the most likely location given its proximity to Samsung’s S2 line, a 90,000 WPM 300mm fab, the power issues that caused the plant to shut down during a February snowstorm might bring into question Texas’s unusual power grid issues and add some additional risk to that potential expansion site.  Also being considered are two locations to the west and east of Phoenix, Arizona.
The site in New York State is one that has been developed by the state and Genesee County (near Buffalo) and features low-cost hydroelectric power (30mi from Niagara Falls), although from the data we have seen, power is less expensive in both Austin and Phoenix, a 24” natural gas pipeline, and two redundant high-capacity water lines feeding the 1,250 acre site.  With Cornell, RIT and 55 other colleges and universities nearby, the site promos offer over 1m potential employees within a 30min commute and is accessed by two major highways and is 4mi from a major interstate. 
That said, Samsung has been said to be requesting tax reductions between $800m and $1b to settle in any of the three locations, with the alternative of setting up another fab in South Korea if demands are not met.  Right now it is certainly a buyers’ market for such facilities, especially after the President’s recent EO concerning the semiconductor shortages, which sanctioned another study that should be completed around June, roughly when Samsung is expected to make a decision on the final location of its new fab.    “Where money talks, nobody walks”[1]


[1] – Charlie Greer – WABC for Dennison Clothes. 1967
 

​Back in January (1-21-21) we noted that Chinese CE brand TCL (000100.CH) and affiliate Chinastar (pvt) were considering a project that would entail the development of a new Gen 8.5 LCD fab in Guangzhou, China, with the potential for another fab dedicated to producing flexible and rollable OLED displays using printing technology.  TCL has now signed an agreement with the Guangzhou Municipal People’s Government and the Guangzhou Development Zone, both state-owned organizations, to build a new Gen 8.5 LCD fab, owning a 55% stake, with the government’s stake at 45%. 
The project, which is expected to cost 35b yuan ($5.35b US) is expected to begin construction in September of this year, although it is contingent on TCL raising 12b yuan ($1.83b US) in a private share sale to up to 35 investors, of which 9b yuan ($1.375b US) will be used for the new fab.  According to the company, the new fab will take two years (from construction start) to begin production and will have a stated capacity of 180,000 sheets/month, which is in line with our original expectations, although the fab’s start date is a bit earlier than our expectations if they meet their stated goal.  The backplane technology is expected to be IGZO and the fab is expected to be focused on producing IT oriented displays, such as monitors, notebooks, automotive displays and similar mid-sized panels.
While still based on the capital raising contingency, we expect TCL has already found close to the required investment needed for the project and it seems that the government has helped to move the project along more rapidly than expected.  No word on the status of the OLED fab has been made as part of this project agreement.
 

China Adds Display to Import Tax Emptions
​

​Last month we noted that the Chinese government had added semiconductor materials, equipment, and products to its import tax exemption list, giving Chinese semiconductor producers a lower cost base than other semiconductor producing countries, or at least taking import taxes out of the cost equation for local companies.  Yesterday the Ministry of Finance of the People’s Republic of China issued a notice adding display raw materials, consumables, and both foreign and domestic display production equipment to the list of products exempted from the import tax.  This includes LCD, OLED, and micro-LED materials and products, and will remain in effect until 2030, with a claw back provision for import taxes on the items specified back to January 1 of this year.

​Samsung Electronics has signed an agreement with UMC (UMC) to produce image sensors and display drivers at UMC’s P6 plant in Nanke, Taiwan, which will have a production capacity of 27,000 wpm, running primarily at 28nm.  Most of the plant’s capacity is expected to be allocated to Samsung when the plant opens in 2023.  As part of the deal, Samsung is expected to sell 400 units of unspecified fab equipment to UMC. 
Samsung’s image sensor demand has outstripped its production ability even as Samsung is planning to expand its internal production capacity by 20% to 30% this year through the conversion of one of its DRAM fabs.  A number of South Korean equipment suppliers are expected to benefit from Samsung’s CIS expansion, so if Samsung’s UMC deal changes that long-term timeline, expectations would have to be adjusted, but given that the UMC fab is not expected to be in production until 2023, we expect Samsung’s current internal CIS expansion plans will continue this year.

As we have noted, LG Electronics (066570.KS) has finally decided to exit its money-losing smartphone business after 6 years of losses totaling $5.4b.  One of the last smartphones the company released (October 2020) was the unusual LG Wing 5G, a 6.8” smartphone with a rotating 3.9” second screen, with only 3 low-priced LG smartphones released this year.  The original price of the Wing when it was released was $999, although AT&T (T) was selling it for $1,049.  LGE has promised to continue to update the OS on its existing phones for three years, and FlipKart (WMT) will throw in a one-year warranty and 5 years of service if you buy one, but the real action is with the phone’s price. 
The difference in full (no trade-in or new line discounts) price for the same phone (256 GB) among the three carriers is unusual, with T-Mobile (TMUS) the lowest at $599, Verizon (VZ) at a more normal $999, and AT&T (T) still selling it at $1,049, and at various other on-line retailers, the price can go as low as $489, but if you are interested in buying what might become a museum piece years from now at the best possible price, you will have to travel to India, where you can buy that same phone for $399.99, 60% off of the listed price.
History will tell us whether the LG Wing 5G will have any value years from today, but if we look back at a particular Samsung (005930.KS) smartphone failure, the Galaxy Note 7, released in August of 2016, stranger things have happened.  The Galaxy Note 7 was infamously known for busting into flames, causing Samsung to informally recall the units in the US and replace them with ones with a different battery.  Unfortunately those also caught fire and a worldwide recall of the Galaxy Note 7 was reported on October 10, 2016, causing an estimated $17b in costs to Samsung.  What makes that very negative incident unusual is that in July 2017, 9 months after the recall, Samsung released a limited release refurbished version of the Galaxy Note 7 called the Galaxy Note FE (Fan Edition) for $610, a 29.5% discount from the original price.  While relatively limited in its release, the Galaxy Note FE was a hit and became a collector’s item.

​According to OMDIA, Samsung Display (pvt) is expected to make a decision on the 2nd of 4 potential phases of capacity expansion for the QD/OLED project later this year, which would add another 30,000 sheets/month to the project’s potential capacity by 2023.  That same timeline would begin phase 3 build-out in 2022, with that additional capacity (30,000 sheets/month) slated for production in 2024.  That said, SDC has a number of hurdles to cross before those decisions are set, and is still researching whether the lines will use an OLED material as a light source or whether it will switch to nano-rods, a quantum dot structure.
While there are many technological challenges in bringing this new display technology to market, there are also production challenges as some of the production equipment will be specific to the process chosen and therefore developed specifically for SDC, either by an outside supplier or an affiliate.  Given that such equipment development takes time and continuous refinement, timelines for high volume production are tenuous at best, and yield improvement, regardless of the final process will be the most important challenge, even if the technology is proven successful.  If SDC can’t make money with the process, either because of equipment cost or low yield, we doubt they will finance the 2nd or 3rd phase expansions unless parent Samsung Electronics gives the ok, and that remains an open question as they have not been ready to fully endorse the project quite yet.