Too malas to get links so please Google if you want details

Bad news

MyEG signs deal with China Customers on cross-border. How will this impact DNex single window?

Good news
Taiwan faces drought possibly impacting semiconductor and textile. Also brain drain to China

Huawei to launch 12nm 14nm chips later this year. China staring mass producing 14nm chips

Cup is half full?

I got this flash .. and used a translator. In summary it indicates that TSMC ASP for 3nm have gone up past 20,000

Similarly SilTerra ASP will rise proportionally.

https://www.digitimes.com.tw/tech/dt/n/shwnws.asp?CnlID=1&Cat=40&id=0000650599_5BE4DX9XL0Q1EC8JF19TV

Can those who can read Chinese please verify. TQ

There is an ongoing court case between ARM and Qualcomm.

https://www.bloomberg.com/news/articles/2022-10-27/qualcomm-fights-back-at-arm-in-dispute-over-chip-technology

This could benefit Silterra (DNEX) either way - increased licensing revenue along the lines of ARM or benefit from customers who leave ARM.

As usual this is NOT a buy recommendation.

KUALA LUMPUR (Oct 26): The Malaysia Semiconductor Industry Association (MSIA) projected the semiconductor industry to grow by 8% to 10% in 2022, followed by a weaker 2023, due to weaknesses in the personal computer and smartphone manufacturing industries, Hong Leong Investment Bank Bhd (HLIB) said.

The investment bank said in the past 12 months, a total of RM52 billion in semiconductor investment was announced, targeting to create 11,000 jobs.

https://www.theedgemarkets.com/article/malaysias-semiconductor-industry-projected-grow-810-2022

Again it's dependent on type of products the companies are producing.

Some say semiconductor is the new oil .. well ...

TSMC 3rd Quarter Results
Revenue US$20.23 billion vs guidance $19.8–20.6 billion
Revenue NT$613.14 billion (guidance NT$588.06-NT$611.82 billion)
Q3 gross margin 60.4% vs guidance 57.5%-59.5%
Q3 Operating Margin 50.6% vs guidance 47%-49%

Sanctions?

SK Hynix secures one-year waiver from U.S. govt on chip equipment for its facilities in China

https://www.reuters.com/technology/sk-hynix-secures-one-year-waiver-us-govt-chip-equipment-its-facilities-china-2022-10-12/

Some news as picked up on semiconductor

1. TSMC have informed their customers including Apple that they will be increasing wafer price from 2023

2. Couple of chip makers esp memories in Taiwan will not be cutting down their production unlike some companies in the same segment.

Thus read any which way you prefer ...

sorry for the double posting ...

The spike in USD is not just bad for importers but also for our government as we need to service our loans and bonds which are mostly in USD ...

Here is a piece of news. (I had to use a translator so I may be mistaken)

The chip giant Intel has given an order to TSMC to manufacture Intel's GFX graphics chiplet that is used with its Arrow Lake CPU.

It will be for TSMC 3nm capacity which will only start production in Q4 2023.

https://tw.stock.yahoo.com/news/%E8%8B%B1%E7%89%B9%E7%88%BE%E5%B0%8D%E5%8F%B0%E7%A9%8D%E9%9B%BB%E9%87%8B%E5%87%BA3%E5%A5%88%E7%B1%B3%E5%96%AE-%E5%A7%8B%E8%87%AA15%E4%BB%A3arrow-lake-gfx-%E6%98%8E%E5%B9%B4q4%E9%87%8F%E7%94%A2-042941493.html

There are many experts here with a wide range of opinion on DNeX. Some optimists while some pessimists.

So read whatever way you choose into this piece of news and reflect it vis-a-vis DNeX situation and Silterra a pure-play foundry.

@Toyota78

With regards to your #7

I would put it a MAYBE not YES as I believe there are execution and cost issues related

To recap DNeX through EDI-M then TimeDotCom has been operating the NSW (previously known as EDI Malaysia) for more than 30 years as a sole service provider. In addition DNeX has also other contracts from the government not related to NSW.


If you review the various announcements since Tan Sri took over there are many initiatives such as SEALnet related; JV with EXIM, China Construction Bank, CapBay on trade financing and RSW; are viable plans. But results (or data not available) do not show inline with the potential. So is it an execution problem or cost related (too many outsourced resources?)

There were also a legal dispute.

My guess is this "house" is not in order yet for various reasons. Hence I put a MAYBE

Contrast TSMC's revenue

Intel's revenue will decline 17% in 2H22 y/y and 7% sequentially, based on CEO comments at Evercore. 2H y/y decline happened only twice since 2009 (2012 & 2020). Revenue will come in around $65 b for 2022 (an $11b lower than Apr 22 guidance). $NVDA is also set to see a 2H decline

Slowdown?

TSMC's revenue in August leaped 58.7% year-on-year to a fresh all-time high NT$218.13 billion (US$7.06 billion) and up 16.8% over July. Jan-Aug revenue hit NT$1.43 trillion, up 43.5%.

https://pr.tsmc.com/english/news/2959

As I have often repeated it depends on sector or business within of semiconductor which is very wide.

TSMC is big enough to absorb the slowdown in memories unlike Samsung.

We should take a mid-long term view of DNeX/SilTerra

One door closes another one opens ...

Nvidia is caught up in the new sanctions the U.S. is imposing on China. The company said it will not be able to export two of its AI chips to China, its second-largest market. That will likely cost Nvidia some $400 million in lost sales for this third quarter and interrupt some production that happens in China.

This will impact TSMC as posted earlier as Nvidia accounts for 10% of TSMC business.

August 23, 2022

Semi Capex On Pace For 21% Growth to $185.5B This Year
2020-2022 expected to be the first 3-year period of double-digit capex growth since 1993-1995.

IC Insights has adjusted its 2022 worldwide semiconductor capital-spending forecast that now shows a 21% increase this year, to $185.5 billion (Figure 1). The revised outlook, which was included in IC Insights’ recently released August 3Q Update to The McClean Report, represents a decrease from $190.4 billion and 24% growth that was forecast at the beginning of this year. Though lowered, the revised capex forecast still represents a new record high level of spending. In fact, if industry capital spending rises as forecast by a double-digit amount this year, it will mark the first three-year period of double-digit capital expenditure gains in the semiconductor industry since 1993-1995.

Wafer fab utilization rates at many integrated device manufacturers (IDMs) remained well above 90% through the first half of this year and many semiconductor foundries operated at 100% utilization rates, as orders remained robust during the economic recovery from the Covid-19 pandemic.

Combined two-year semiconductor capital spending in 2021 and 2022 is now expected to reach $338.6 billion. IDMs and foundries are spending heavily on new manufacturing capacity for logic and memory devices built with leading-edge process technology. However, strong demand and ongoing shortages of many other essential chips such as power semiconductors, analog ICs, and various MCUs, have caused suppliers to boost manufacturing capacity for those products as well.

While all of that is positive news, a menacing cloud of uncertainty looms on the horizon. Soaring inflation and a rapidly decelerating worldwide economy caused semiconductor manufacturers to re-evaluate their aggressive expansion plans at the mid-point of the year. Several (but not all) suppliers—particularly many leading DRAM and flash memory manufacturers—have already announced reductions in their capex budgets for this year. Many more suppliers have noted that capital spending cuts are expected in 2023 as the industry digests three years of robust spending and evaluates capacity needs in the face of slowing economic growth.

https://www.icinsights.com/news/bulletins/Semi-Capex-On-Pace-For-21-Growth-To-1855B-This-Year/

As I often mentioned the chip industry is very wide and heterogenous in terms of applications and functions.

have a good and safe weekend

From Bloomberg

Don’t celebrate yet
Bad news for chipmakers is generally good news for customers: In their earnings reports this season, Intel Corp., Nvidia Corp. and Micron Technology Inc. gave forecasts that indicate chip supply is outrunning demand. The implication—that a supply glut could be looming—is potentially dire for the companies.

But finally, it seemed, customers who had been struggling to get enough silicon, and not pay dearly for it, would get some relief.

Unfortunately for both customers and chipmakers, the reality is more complicated. The chip companies that are struggling with the buildup of unused stockpiles and falling orders are primarily concentrated in one area: the PC industry. In the pandemic, the rise in remote work sent demand for personal computers skyward. Now, people are returning to offices and schools, and deciding they don’t need that fourth home laptop.

Meanwhile, the customers that aren’t looking to buy PC chips are still struggling. Cisco Systems Inc., the maker of machinery that basically runs the internet, said it still can’t get enough components to meet demand. Like many others, Cisco is suffering shortages of a lot of types of chips, including ones that may perform relatively simple functions but are nonetheless essential to complex devices such as network switches or routers.

Older types of chips, like power convertors, are made in factories that have long since ceased to be state of the art. Because of the industry’s fascination with producing chips using hyper-advanced technology in facilities that cost as much as $20 billion, there’s been very little investment in expanding capacity of older, less complex manufacturing.

Cisco forecast that the issues will persist through its fiscal 2023, which ends in July next year. Scott Herren, Cisco’s chief financial officer, said his hope is that factory space freed up by less production of PC parts will be devoted to making more of what he needs, but that will take time to play out.

Chip stocks have taken a dive recently on supply glut fears. But it’s worth nothing that not everyone is hurting. Analog Devices Inc., a maker of chips used in vehicles and industrial machinery, said its backlog of orders is still growing. Prices for its products are stable, it said. This is from a company whose worst quarter this year featured 72% revenue growth.

Other companies, like NXP Semiconductors NV and STMicroelectronics NV, are seeing strong demand as well. Their results suggest the chip industry is no longer as reliant on PC and smartphone growth to stay healthy. Semiconductors are used in so many products these days—cars, factory equipment, appliances—that chipmakers have an extra cushion.

Still, that’s not placating investors who react to the slightest mention of a slowdown. Even after Analog Devices gave a better-than-expected forecast—saying it was conservative, at that—the stock fell 5% on Wednesday.

A chip glut may well be coming, but for much of the industry it's not here yet. Applied Materials, the biggest maker of machines that manufacture chips, said Thursday that it’s still receiving more orders than it can fill—because it can’t get enough chips itself. That situation is going to persist for several more quarters, Chief Executive Officer Gary Dickerson told us.

This raises the question, if the maker of chipmaking equipment can’t get enough chips, what chance do others have? —Ian King

Sabah to cancel work passes for Hibiscus Petroleum units by Oct 1 unless arrears paid — report

https://www.theedgemarkets.com/article/sabah-cancel-work-passes-hibiscus-petroleum-units-oct-1-unless-arrears-paid-%E2%80%94-report

Here's some news on the semiconductor front

Power ICs - (SilTerra's expertise) is a $21B market fueled by applications such as electrification of transport, autonomous driving, control of battery of power and medical tools, and building and factory automation.

Fabless suppliers hold record 34.8% share of global IC sales
2011-2021 CAGR for fabless company IC sales (10%) was double that of the IDM IC suppliers.
Fabless - semicond companies like AMD etc do not own any fab.
IDM (Integrated Design Manufacturers) design and manufacture ICs eg Nvidia
SilTerra is a pure play foundry (fab) - produce for other companies. However SilTerra has some toolkits (building blocks) certified by automotive industry that can be used by customers to build their own products.

Foxconn Industry Internet (FII), the Shanghai-listed subsidiary of Hon Hai Technology Group, has acquired a minority stake in China’s state-backed chip maker Tsinghua Unigroup.
Hon Hai - also known as Foxconn, the world’s largest iPhone assembler – invested 5.38 billion yuan (US$788 million) in Tsinghua Unigroup via Guangzhou-based Xingwei Industrial Investment Partnership Fund, Taiwan-listed Hon Hai Technology said in a stock filing on Thursday.
The Taiwan government is reportedly considering fining Foxconn up to T$25 million (US$835,600) for investing in Tsinghua Unigroup without first seeking regulatory approval, Reuters reported on Friday, citing a source.

TSMC's profit rose a whopping 76 per cent in the second quarter. The company did warn of inventory build-up, though, and predicted a “semiconductor down cycle” in the first half of next year.

Meanwhile Kenanga gave a TP of 4.10 for P.I.E. Industrial (7095)

https://klse.i3investor.com/web/pricetarget/research/63892

PIE is 51.42%-owned by Taiwan’s Pan-International Industrial Corp, which in turn is 27.33%-held by Foxconn, more formally known as Hon Hai Precision Industry Co Ltd.

TSMC sales soar 44% in another sign of resilient tech demand

https://www.theedgemarkets.com/article/tsmc-sales-soar-44-another-sign-resilient-tech-demand

3nm 5nm 8nm ... each have their own benefits and uses

The world marches on ...

Samsung Electronics starts 3-nanometer chip production ahead of TSMC
news / 4:59 PM GMT+8•June 30, 2022


Samsung Electronics said Thursday it has kicked off mass production of 3-nanometer chips, becoming the first company to do so globally, as it aims to beat Taiwan Semiconductor Manufacturing Co, or TSMC, the world’s most advanced foundry chipmaker.

Samsung said it’s using gate-all-around (GAA) transistor architecture, which allows these first-generation 3-nm chips to have 16% smaller surface area, 45% reduction in power usage and 23% performance improvement compared with current 5-nm chips. The South Korean company also said in a statement that the second generation of the 3-nm process would allow 50% lower power consumption.

The company is currently producing the first generation of 3-nm chips and plans to start the second generation of the 3-nm process production in 2023, a spokesperson at Samsung Electronics told TechCrunch.

Samsung has been competing with Apple chipmaking partner TSMC, which also said in June that it would begin mass production of a 3-nm chip process to volume production in the second half of 2022. The Taiwanese company plans production of 2-nm chips by 2025. (The smaller number of nanometers, which are hard to develop, the more advanced chips, according to industry sources.) The spokesperson explained that smaller nodes allow more transistors to be placed on a given area, which enables the chip to be more advanced and more power-efficient.

The announcement comes amid a global chip shortage, fueled by the coronavirus pandemic, that has threatened manufacturing companies that need the most advanced chips for the next generation of products.

The South Korean tech giant will produce the advanced 3-nm chips at its Hwaseong semiconductor production lines and its third chip plant in Pyeongtaek, the world’s largest semiconductor facility.

President Joe Biden visited Samsung’s Pyeongtaek chip plant in May on his first trip to Asia to strengthen semiconductor alliances.


Samsung announced last year it would make a 171 trillion KRW ($132 billion) investment in the logic chip and foundry business by 2030. Samsung Electronics is setting up a $17 billion semiconductor plant in Texas. Meanwhile, TSMC said in April it would invest $100 billion to expand its chip fabrication capacity for the next three years.

“Samsung has grown rapidly as we continue to demonstrate leadership in applying next-generation technologies to manufacturing,” the head of foundry business at Samsung Electronics Siyoung Choi said in its prepared statement. “We will continue active innovation in competitive technology development and build processes that help expedite achieving maturity of technology.”

@nicholas99

agree but the reverberation or tremors (even slight) will be felt in the east.

FWIW .. Eastern Taiwan got hit with a 5.9 scale earthquake.

There were several quakes before. Hence construction of wafer fabs amongst many types is costlier and most probably extra QC checks needed. Chips manufacturing especially fabs are very sensitive to movements. Much so with lower nm these days.

888 you wrote

"Dx would not be involved with EV. And Silterra would not be short of Customers, in fact ChipOne LTA is more than deliverable."

An EV has up to at least 3,000 chips. SilTerra is already in EV

https://www.thesundaily.my/business/silterra-unveils-new-technology-for-chips-in-vehicles-HC8260779

In fact process has already being qualified for EV related.

"SilTerra’s 180nm Bipolar-CMOS-DMOS (BCD) automotive grade process technology is qualified to meet the AEC-Q100 Grade Zero standard, which is the automotive industry standard that specifies the recommended new product and major change qualification requirements and procedures."


SilTerra has not one but two LTA - second one with Taiwan Ilitek

https://www.businesstoday.com.my/2021/11/29/silterra-inks-agreement-with-taiwans-0ilitek/

The LTAs should eat up about 65-70% capacity while i believe the target is 80% even with the new expanded capacity of 10 million masks in 2023. (Up from 8.7 millions now.) RM 645M sourced internally. So already found and utilised and WIP.

So watch out for more LTAs.

This exclude the capacity planned under the MOU.

An extract from a trade report on semiconductor

Semiconductor growth still seen at 11% despite 2022 headwinds
New quarterly update shows higher gains in microprocessors and power discretes this year, but lowers sales increases in optoelectronics while the global economy faces greater risks.

This is following a 25% increase in 2021 and an 11% increase in 2020


Note SilTerra makes power discrete amongst other segments.

#StaySafe

For those who are interested in EV and EV related counters, BloombergNEG just reported on the Zero Emission Vehicle Transition Council latest report.

You can find the summary here on sales, forecasts, etc

https://about.bnef.com/blog/zero-emission-vehicle-transition-council-progress-dashboard/

Full report

https://assets.bbhub.io/professional/sites/24/BloombergNEF-Zero-Emission-Vehicles-Progress-Dashboard-May-2022.pdf

If you follow i-Saham categorisation I believe more than 10 counters are involved in EV.

Here's one to DNEX executives. Please update.

Kuala Lumpur 14 July 2021 – Today, China Construction Bank (Malaysia) Berhad (“CCBM”) and China Construction Bank Corporation Labuan Branch (“CCBL”) (hereinafter collectively referred to as“CCB Malaysia”) have signed a Memorandum of Understanding (“MoU”) with SealNet Sdn Bhd
(“SealNet”), Bay Supply Chain Technology Sdn Bhd (“CapBay”) and Labuan Digital Authentication Centre Inc (“DAC”) to collaborate together to build a comprehensive Regional Single Window (“RSW”).

http://www.ccb.com/malaysia/uploadfile/ggxx/202107181626581862/20210718121821173564.pdf

Just let other small companies do the integration part - PLUS style ie build the roads and you just tax cars and lorries that uses it. Build once use many.

There was report of SEALnet hitting 300% growth albeit from a low base.

https://themalaysianreserve.com/2021/07/28/dnex-expects-sealnet-revenue-to-grow-300/

Intel CEO now expects chip shortage to last into 2024

PUBLISHED FRI, APR 29 202212:53 PM EDT UPDATED FRI, APR 29 20226:45 PM EDT

- Intel’s Pat Gelsinger now expects the semiconductor industry to suffer supply shortages until 2024.

- The CEO told CNBC the extended timeline for the chip crunch is now due to a lack of manufacturing equipment.

- Intel on Thursday reported weaker-than-expected guidance for its fiscal second quarter.

https://www.cnbc.com/2022/04/29/semiconductor-shortage-intel-ceo-says-chip-crunch-to-last-into-2024.html

What does it mean?

The market is there so it depends on how SilTerra - a pure-play global semiconductor foundry - executes with existing equipment and products. Let's hope the equipment for the current expansion (going live early 2023) are already delivered and under commissioning and QC runs which typical will take anything from 3-6 months.

Perhaps a bit of cobbler's children mindset could come into play ie producing customer's orders / devices rather than SilTerra's own maybe for a higher ASP.

Or maybe SilTerra's own MEMS and photonics will fetch a higher price.

Don't forget there is also an arrangement with Finland's Okmetic to supply C-SOI wafers for MEMS and photonics. No update on this.

Endless possibilities.

So stay cool people - the name of the game is us to make money.

Firstly this is not a speculative counter relying on business performances from QR to boost prices especially contribution from the Technology / semiconductor segment.

DNEX took over a very sick company in all aspects including management. This has been corrected including using funds to improve reliability, productivity and quality of EXISTING facilities. This has been done and it shows in the numbers including rising ASP.

You may have contracts including long term utilising 60-80% of your capacity but you don't get paid if you don't ship quality parts or have same capacity for EACH of those products from front to back.

A simple analogy will be a car manufacturer. you can make 10K engine blocks but your transmission can only make 3K auto and 2K manual. So you can only sell if you are lucky a max of 5K

I do not think quality has been spectacular in the past.

The second phase which is to expand capacity by 10% is expected to complete in early 2023.

So you can expect revenue to go up after that period. Not overnight.

Then there is the recent MOU which have not much details at the moment so I will refrain from making any comments except to point on raising of capital by DNEX (endless possibilities) and time taken for the first product to roll-out - expect 2.5 to 3 years.

Therefore please do not expect massive contribution from the Technology segment so soon and translate that to share price or contribution from the segment.

The other segment that is showing profit is IT despite of a monopoly of trade facilitation while consulting / system integration is a very challenging environment. Trade facilitation would remain the main growth in this segment especially if trades take off and regional trade RSW implemented efficiently.

E&OE

One more .. hope you don't spill your coffee ...

Silicon wafer is the key components to start the chip manufacturing processes.

From Bloomberg (I am not a subscriber so just give headlines from 2 days ago)

Key Supplier of Wafers for Chips Has Sold Out Through 2026
- Sumco expects supply-demand imbalance to last five years
- Silicon wafer maker sees little room for factory expansion

Further back in February 2022


Wafer woes could drag out chip shortage

Major vendors not meeting demand.
Two silicon wafer vendors, which together represent more than 50 percent of global supply, have warned that they won’t be able to meet demand during 2022.

https://www.itnews.com.au/news/wafer-woes-could-drag-out-chip-shortage-575853

You could argue about using alternative to silicon wafer such as Gallium Nitride (GaN) and Silicon Carbide (SiC) which DNeX has some products but capacity is not that high or the quality that good (same with other foundries).

Infineon across the road from DNeX is building some facilities there expanding towards that area.

So you could translate all of the above into some sort of revenue for the next few QRs ...

DNeX - BIH aka Hon Hai aka Foxconn MOU

Is the cup half full or half empty?

Extract from PR

4. SALIENT TERMS OF THE MOU

4.1. Scope of MOU
The Parties are discussing on the collaboration to set up a joint venture company for the Proposed Project.
The MOU is an understanding of the Parties’ intentions and is not intended to create any legal relations nor to be legally binding on the Parties.

4.2. Definitive Agreement
During the duration of the MOU, the Parties will endeavour to negotiate to enter into definitive agreement(s) based on the structure of the Proposed Project and its financial details.

4.3. Duration
The MOU is valid for a period of one (1) year commencing from 17 May 2022 and may be further extended by mutual agreement by the Parties (“the Term”). Notwithstanding the foregoing, the MOU will terminate prior to the expiry of the Term if:

(a) A definitive agreement has been executed between the Parties; or
(b) Terminated by either one of the Parties by giving one (1) month notice of termination in writing.



Just smoke no fire yet .. so hold on your coffee mugs. Also no many billions being mentioned. Who or how the bill is going to be split and paid? Nada ...

Foreplay only? Or marriage and consummation? (Note - not just marriage!). Here are the attempts based on info published

1 Hon Hai 3+3 initiative
2 Foxconn was amongst the bidders for Silterra from Khazanah
3 DNeX won the bid
4 Hon Hai / Foxconn bought 5% DNeX to become substantial shareholders
5 Dr Bob Chen of Hon Hai is on the board of DNeX and SilTerra

Have a great weekend peps ...

@ironshirt

I agree with your point. MOU is nothing and will not impact DNeX bottom line. It also take 2-3 years before the first product roll out eg construction, installation and qualification / test.

Those MOUs with Chinese firms are slightly different. Why?

Because Hon Hai has already owned 5.03% of DNeX. They already have a foot past the door.

Here are a few more pointers

Hon Hai aims for 10% of EV market

Under Hon Hai 3+3 initiative - 3 emerging industries of EVs, robotics and digital health care applied AI, semiconductor, and comms technologies - DNeX / SilTerra is a good fit as they have products in many of those areas including health care.

From PR the new plant will use "matured" processes which reduces the need for more RnD and risks.

However the journey has just started.

hi @ogre2021

I don't know.

Tan Sri / BOD will be in better position to answer about the planned future expansion to 12M mask layer per year in 2024 in light of this JV.

Perhaps the JV could replace that plan thereby saving DNeX a chunk off RM700M. Also offers many restructuring options.

Would be the JV be be at the holding (DNeX) or Silterra's level?
How would DNeX funds the JV?
Spin off and list Silterra?
CGP Fund is a PE so would they exit once the JV is in place?

I am not a mind reader or would like to predict things except to say that there is a ready "in-house" customer for the products from the JV.

As the JV is with a "western friendly country" the products and the production technology would be free of sanctions.

Thus we have access to market and technology.

Not a slam dunk but its up to the company to screw things up.

Peace

Posted by ogre2021 > May 18, 2022 12:31 AM | Report Abuse

hi @jwblacklabel
today dnex announced mou for jv with foxconn subsidiary to build & operate 12-inch wafer with capacity of 40K/month.

what are your thoughts on this wrt silterra planned future expansion to 12M mask layer per year in 2024 for an additional cost of rm700M?
pls refer to link below, based on the NST report, the current silterra expansion to 10M mask layer will be ready by early 2023.

https://www.nst.com.my/business/2022/03/782006/dnex-eyes-group-wide-ex...

Chips are not just for or about EV ... repeat 1000X

@king36

Re:future of chips manufacturing

These are notes from a recent roundtable of industry players

Summary - Growth outlook is strong for ICs, both individually and in packages, but getting costs under control is a huge challenge.

In the next five years or so the semiconductor market is expected to double to $1 trillion. Anirudh Devgan, president and CEO of Cadence

There are new applications and more industry verticals. .. Fabs are fully booked for the next few years, but they are adding capacity.,, Niels Faché, vice president and general manager at Keysight

(note comments about new apps and verticals - NOT just EV but areas like complex IoT, heavy computing power needed in deep learning (AI) etc. Most probably towards system in package )

DNeX / Silterra's current capacity is almost fully booked with couple of long term contracts. Matter of execution, managing cost and once USP is increased revenue will follow likewise.

Oil? I know nuts about that .. sorry ...

PS ..

It does not matter at present whether 200 or 300 or 400mm mostly. What you get is more DPW plus some thickness variability.

At this stage should focus more on the backend of line (BEOL) capacity as each cluster of equipment is dedicated for certain product / technology example aluminum or copper etc ..

My guess - and I could be wrong - they are adding equipment in that area so there could be more linearity in the production flow rather than start-stop for each type.

Eg for those who are familiar with the prawn mee in the MPPJ area - that guy (don't know if he is still there) at peak hour will only serve bee hoon-mee. Nothing else. That's linearity .. same product over and over.

PPS 200/300/400 seems to be boys toilet comparing length of errr the birdie ..

Have a pleasant evening ..

It's timmmmmeeeeeee ... not UFC but another semiconductor 101 (and a some corrections)

I am going to use very simplistic explanations so you can get a flavour.



We know SilTerra's capacity is 46,000 wafer starts per month. (I am not going to factor losses or yields)

SilTerra processed 200mm wafer. A wafer will contain many dies (circuit) aka chip. Assuming the dimension of the chip as 10X10mm. This according to the semicond thesaurus should work out to be around 250 die per wafer (DPW)

So working backwards on Foxconn's EV plans of producing 55K cars and assuming 10 chips (easier on my mental maths ya) per car.

Thus for the year Foxconn would need 55,000 X 10 = 550,000 chips

This will translate to 550,000 / 250 = 2,200 wafers

That is equal to less than 5% of SilTerra's capacity for a month.

So how are they going to fill in the other 95%?

That's why I say the best thing SilTerra can do manufacture anything and everything for anyone (since they are a pure play foundry) to their maximum capacity. Not just EV but also MEMS, etc.

Of course it's an oversimplifications but i hope you can appreciate the scope and shout less about EV and what's not. (I appreciate the enthusiasm of folks here but let's not go overboard with Foxconn EV in particular).

BTW no EV chip is the same for all producers. Even your laptop processors are different (esp Intel and AMD).

BUT customers can either use SilTerra BCD tools to develop their chips or bring their photomask for production at SilTerra knowing that SilTerra's technology and processes are already certified Grade 0 for automotive technology standards.

Guys @nicholas99 @lalasai9

In semiconductor proximity does NOT matter.

Chips for Windows PCs as Intel and AMD (then) built and tested in facilities that were situated across in the road in Bayan Lepas.

Intel's major assembly and test facilities are now in Kulim. Like wise many many other MNCs including WD in HDD are in Malaysia. It's because of the people.

Sometimes due to criticality of delivery time or security the plants actually had staff hand-carrying the chips to other site.

Foxconn did bid for SilTerra. Their plans were not clear. Even if they intended for EV where are they going to assemble the chips after the wafer are produced in Kedah. Perhaps in Singapore.

Foxconn have been building their own semiconductor capabilities with several plants in China. They are more than EVs. If you look at their production numbers planned for Thailand that is only about a month's of SilTerra capacity.

#valuechain

@lalasai9

Sorry .. Foxconn EV has gone to Thailand.

https://techcrunch.com/2021/09/21/foxconn-says-thailand-ev-factory-will-begin-producing-50000-units-by-2023/

Do you think our current policies (and politics) are investor's friendly unless to build on something that is already in place or expanding existing operations? Or something which we are a leader on?

Thus ended the demand for QR until next time.

Let's end the EV / semiconductor chapter as well.

here is a good article

How the Electric Car Industry has Impacted Semiconductor Demand

https://www.azom.com/article.aspx?ArticleID=20351

Please check the resource and extra reading section.

#notallchipsarecreatedthesame ...

@JJChan

192M? no problemo. according to some of the projections here that works out to 1 week of wafer starts for SilTerra excluding yield losses.

Yes this is the place of the ridiculous to the sublime!

LOL @Solomon7

yes indeed SilTerra is the throes of genesis ...

Me bad as was thinking of Stevie Ray V and Texas Flood ..

Have a great evening!

Bro @huat_aah

I am comfortably numb with your Clapton analogy.

SilTerra should be able to make money right now so long as their equipment are running according to their capacity as orders are full. CEO TS said that their order books are filled until somewhere - can't remember now but definitely for many quarters.

As I have posted before it depends on what the "philosophy" is. Whether to run as per order or run non revenue generating engineering lots for their RnD or qualifications. Basically scheduling and disciplined starts.

The recently hired Group CFO is a fab man - I think ex Moto and On. Don't think he wants to foresake his and his colleagues ESOS.

@SinGor raised some valid concerns but he visit the SilTerra website and check out the management team he will find that there is a new team - all in acting position but I believe are existing staff which will ensure continuity. The acting COO has been around for several years and specialisies in yield.

SilTerra's challenges are IMHO leadership and talent (aquistion, rention and performance). Perhaps with the heads of Hydra being removed the ahem .. cream will rises to the top.

I am always a believer that people wants to do a decent job for a decent rewards and management has to create an environment for them to excel.



PS would be interesting if Clapton picks up a Gibson ...

Sir @huat_ahh

I am just sharing knowledge in hope that we will have more knowledgeable investors who are then able to provide some form of check and balance to the Board and its executives.

I believe financially many here are able to do that and challenge the boad if needed.

Worst case - touch wood - if they (and I) lose money on the counter at least we know why ..

Have a great weekend sir!

EV EV EV everywhere ..

Frankly if you are talking EV about SilTerra you are doing a major injustice to the company.

SilTerra IS MORE THAN EV

SilTerra has a broad range of its own products like in MEMS and bio-science realm.

SilTerra has general equipment that makes kind of products based on the designs of other - this is what is meant by PURE PLAY foundry.

Another simple explanation - SilTerra is like those shops where you go develop your holiday shots. You bring the negatives or they develop it for you and then print out the photos. That's what SilTerra does.

I want to bring a sense of reality and expectation to people here.

Heck we don't even know how much EV chips can SilTerra produce. The 46K is where they starts. Then you have to account for breakages and defects. And you also need to account for the back end equipment of which NOT ALL can finish the process for EV chips.

So net of net of the 46K?

So please don't just EV here, there and everywhere ...

It's more than EV here and the demand is across the board of useage ..

Hari in dalam sejarah

Carsem or Carter Semiconductor has been in the OSAT Outsourced Semiconductor Assembly and Test since the late 1970's or early 80's. HL through MPI bought them over in the 1990s' (thereabouts).

The other grand-daddy of local OSAT is KESM also in late 1970's early 80's. They however does only the "T" portion - testing and burn-in. Chips need to be baked and tested before shipping to customers.

Let me explain in a very simple manner.

Imagine if you are baking some cupcakes. you would get OSAT equivalent - A to make the mixture for you then send to T who will bake them for us using moulds (burn-in boards)

Except often that not those moulds are customised. Hence you have listed companies and others making those burn-in boards.

Some like bunnies, teddy bears, sharks etc in shape la ..

If i can stay awake for next 30 mins i might get a list of burn-in boards manufacturers that are listed. Worth taking a look.

#playsafe

Take a look at iSaham analysis of DNeX


https://www.isaham.my/stock/dnex

Scroll down towards the end Summary .. and you will find this

Manipulation Indicator: Very hard to manipulate.

Have a good day! #buysafe

Sir @trevor777

nothing to do with techie or not .. more in your garden - evaluation and company's worth ...

Assuming those IPs are real have they been evaluated?

Sir @trevor777

I know you cover DNeX / SilTerra in-depth.
Did you come anything about IP evaluation of SilTerra BCD libraries? After all its certified to Grade 0 automotive electronics standard. Must be worth something.

BTW you cannot evaluate their processes as mostly dictated by the equipment they use. Just like a Nespresso machine.

hmmm .. too many food references

Sir @nicholas99

Is SilTerra worth USD 4B?

Glass half full. We can also say DNeX / BCGPF bought a fab dirt cheap.

End of the day if SilTerra cannot produces wafer on time and with quality it will be worth nothing.

The potential is there. And that is what we should be investing as well as mitigating the risks. Not base on QRs.

There are some who are traders. Good luck to them.

One man's meat is another man's bah kut teh ...