25-07-2012, 09:35 AM (This post was last modified: 25-07-2012, 09:43 AM by Zelphon.)
(25-07-2012, 12:41 AM)mrEngineer Wrote:
(25-07-2012, 12:04 AM)l0nEr Wrote: Wow. Thanks to all the industry experts for their knowledge on the industry.
I'm a little confused about how the industry work.. as in the supply chain.. anybody know a place where i can find some understanding on how the industry works? I have problems trying to figure out what UMS makes actually, and how are they different from other 'precision engineering' firms such as Hi-P, Venture Corp., etc... (i guess i shouldnt be investing in things i dont understand).
My miserable knowledge only tells me that Intel hasnt slashed capex estimates, and Qualcomm and Nvidia are still short of chips.
Hi-P and Venture are contract manufacturers which assembles the components and create a microcontroller and put the microcontroller into the plastic cover that was created using the injection moulds.
You can find more information from Kopikat's previous posting somewhere in the forum.
One more point on why replacement or maintenance capex is large in Semicon is because on top of monthly or annual prevent maintenance, the equipments parts have to be replaced to ensure that there are less particles generated after wear and tear. 1 dust particle can be the typical size of 1-2um while the nmos or pmos gate size is minimum now 22nm. So you can imagine the damage a particle can do to 1 chip. That is also why yields are so difficult to achieve as technology gets smaller.
Well, there is 1 GOOD NEWS for semicond industry..
And that is 450mm fab is finally going to get started in the next 6 mths...
But this is probably going to be located in USA...
Reason being why 450mm fab take off is so slow is because ASML, world leader in lithography tool refused to release the tools required for 450mm... All the other equipment makers are already ready to support 450mm but because ASML does not want to release the tool, the rest suck thumb...
This pissed off Intel to an extend that they bought into substantial ASML's shares and get them to release the tools...
Intel need the next generation tools to produce the next generation processors and only 450mm can justify the yield.. As mentioned by mrEngineer, the smaller the gate size, 1 dust can kill.. so to make the overall yield higher and cut down production cost, the wafer has to get bigger... that is why the Intel needs 450mm...
(25-07-2012, 08:34 AM)KopiKat Wrote: On the contrary, it was 'mrEngineer' who posted a good summary of the Electronics Industry in another thread.
Although I ought to place more focus on analysing Electronics related stocks as I do have a bit of core competence here, I find it more worthwhile to spend my time with other less complex industries (where I started with zero competence) like REITs (most times, inreasing rentals thereby raising DPU) or even SPH or Popular Hldgs. Why? Electronics biz is very cyclical (more downs than ups) and most times, highly competitive (especially on price). It's either you try to catch the cycles (buy at low cycle and sell at peak) or you have to figure out those in a successful niche or look for the leaders (the best, the strongest). I find it difficult as it keeps changing and I need to put in lots of time and efforts just to keep abreast of what's happening...
As for UMS, when I was looking at it briefly, their core competence is in Precision Engineering, with focus on Metal parts ie. machining, welding, anodising, plating,... and predominantly on Semicon Eqpt Makers, mainly on one key customer cum shareholder. Altho' their expertise can be easily ported over to other Industries like Aerospace, which uses metal parts, I don't see them as being very successful here...yet(?). I don't really know why, perhaps 'Nick' will know better. So, yes, in theory, UMS need not be too dependent on the highly cyclical semicon biz but in practice, they seem to be not succeeding in diversifying out. Perhaps the CEO is too busy developing his own biz...
I didn't follow Venture or Hi-P. But, from their websites and what I vaguely remembered, yes, 'mrEngineer' is right, they are more in the CEM biz.
Venture started out focussing on CEM biz. But, today, I see from their website, they have become more vertically integrated ie. they can do ODM and even ASIC design (highly margins here, I believe). Not sure if they had also integrated the Precision Engineering part, but not unlikely.
As for Hi-P, IIRC, the 'P' stood for 'Precision' ie. they started out as a Precision Engineering biz focussing on 'moulds' which are used to produce plastic parts eg. chassis. But, today, when I looked at their website, they have also become a more integrated co., which includes CEM. Still, I believe their core competence remains in Precision Engineering.
IMO, that's how both cos. had reacted to this declining profit margin kind of biz of the Electronics Industry ie. Vertical / Horizontal Integration.
So, you see, how much work it takes just to try to understand these biz? In comparison, it's less time consuming and more rewarding for me to pick up new knowledge in sunset industries (less exciting but not much changes) like book stores or even publishing... I hope..
Well, this is from my experience as an IA student in Shanghai Venture...
Venture is mainly involves in OEM biz.. Something like what Foxconn is doing for Apple...
Venture is involved in many types of electronics products and the bulk of the assembly line is in China... HP printer cartridges are assembled by Venture...
Venture is definitely facing cost pressure as the labour cost in China starts to rise.. especially in Shanghai..
(25-07-2012, 09:35 AM)Zelphon Wrote: Well, there is 1 GOOD NEWS for semicond industry..
And that is 450mm fab is finally going to get started in the next 6 mths...
But this is probably going to be located in USA...
Reason being why 450mm fab take off is so slow is because ASML, world leader in lithography tool refused to release the tools required for 450mm... All the other equipment makers are already ready to support 450mm but because ASML does not want to release the tool, the rest suck thumb...
This pissed off Intel to an extend that they bought into substantial ASML's shares and get them to release the tools...
Intel need the next generation tools to produce the next generation processors and only 450mm can justify the yield.. As mentioned by mrEngineer, the smaller the gate size, 1 dust can kill.. so to make the overall yield higher and cut down production cost, the wafer has to get bigger... that is why the Intel needs 450mm...
One minor point, Yield doesn't automatically improve with a bigger Wafer Size. In fact, during the initial stages, Yield is likely going to be terrible and only gets improved as they tweak the process. The main reason for bigger Wafer Size is the potential Cost Reduction as each Wafer can now house more dies (the raw form of an IC chip) and for the same factory size (after all the necessary re-tooling, etc.), production output is now higher (assuming Yield is stabilised).
This semiconductor business is very capital intensive.
Looking at the link, I already feel that I am so out of touch with the industry already... haha. I know there is immersion litho but didnt know there was extreme UV. I thought that they would have ventured into X-ray litho..
This semiconductor business is very capital intensive.
Looking at the link, I already feel that I am so out of touch with the industry already... haha. I know there is immersion litho but didnt know there was extreme UV. I thought that they would have ventured into X-ray litho..
Coming soon.. coming soon. EUV's wavelength is 13.5 nm.
Technically, it is only 3.5nm away from X-ray.
25-07-2012, 07:18 PM (This post was last modified: 25-07-2012, 11:21 PM by Boon.)
(24-07-2012, 05:17 PM)mrEngineer Wrote:
(24-07-2012, 03:16 PM)Nick Wrote: In the long run, with new Microsoft 8 OS, more tablets, more smartphones, more electronic items to cater to a growing wired world, it is likely that semiconductor plants will have to invest in new equipment to meet this demand. On the flipside, a sustained recession brought about by the West fiscal crisis and slower growth in China will reduce demand accordingly and the huge over-capacity will kill spending for a few years.
One should not assume that there is a direct correlation with the increase in new electronic products equivalent to more capex spending.
Are you saying there is no correlation at all between increase in new electronic products and increase in capex spending? If it is a finished semiconductor product, capex had already been spent to produce it. If it is a semiconductor yet to be produced, maintenance capex (at least) as you said needs to be spent. Therefore, I think there is a correlation.
Research, research and research - Please do your own due diligence (DYODD) before you invest - Any reliance on my analysis is SOLELY at your own risk.
Since I was spending so much time on this thread, I decided to take a small position today, inspite of the misgivings I'd posted. Will decide on the fate of this stake after their Q2 results (which ought to recognise IMT-S & IMT-US contributions?). Still, I may not be able to control my itchy fingers...
“Demand for Semiconductor Manufacturing Equipment (SME) is directly linked to demand for semiconductors, which, in turn, is linked to the strength of the global economy”
“The health of the SME industry is tied to the semiconductor industry and follows its global fortunes”
“Several key characteristics of the SME industry are responsible for this recurring volatility, the most important being the direct link between demand for SME and demand for semiconductors An adage notes, “When the economy gets a chill, electronics catches a cold, semiconductors come down with pneumonia, and the equipment industry dies".
All of the above are quoted from the attached publication by the US International Trade Commission, titled “Industry and Trade Summary- Semiconductor Manufacturing Equipment)”.
Though it is a relatively old publication (June 2006), it does give a good overview of the SME industry then. I believe most factors and key drivers affecting the industry mentioned remain relevant today.
Research, research and research - Please do your own due diligence (DYODD) before you invest - Any reliance on my analysis is SOLELY at your own risk.
27-07-2012, 12:51 AM (This post was last modified: 27-07-2012, 12:53 AM by mrEngineer.)
(25-07-2012, 07:18 PM)Boon Wrote:
(24-07-2012, 05:17 PM)mrEngineer Wrote:
(24-07-2012, 03:16 PM)Nick Wrote: In the long run, with new Microsoft 8 OS, more tablets, more smartphones, more electronic items to cater to a growing wired world, it is likely that semiconductor plants will have to invest in new equipment to meet this demand. On the flipside, a sustained recession brought about by the West fiscal crisis and slower growth in China will reduce demand accordingly and the huge over-capacity will kill spending for a few years.
One should not assume that there is a direct correlation with the increase in new electronic products equivalent to more capex spending.
Are you saying there is no correlation at all between increase in new electronic products and increase in capex spending? If it is a finished semiconductor product, capex had already been spent to produce it. If it is a semiconductor yet to be produced, maintenance capex (at least) as you said needs to be spent. Therefore, I think there is a correlation.
Hi Boon,
You have to take into context on my earlier statement on the correlation that I am referring to the supply has outstripped demand into context at this point of time. In the early 2000s, there were low tech fabs in China and almost none in Middle East, but now many fabs have established in these lower cost countries and have pushed the ASP of wafers down that the industry is suffering.
There are some points that you should consider:
1) You made an assumption that the study you have shared in the forum to be relevant with current times. I disagree as no doubt barrier of entries of semicon industries are high, countries like China, Korea and Middle East have actively joined in the game recently because of government incentives on increasing the tech exposure for the country etc. This means that the oversupply situation is still ongoing and I am unsure how the industry is consolidating.
2) The study does not correspond with what I see on the actual ground (which was 2 years ago). Like I mentioned in my earlier post, most new tech product that was recently tech qualified are only expected to be in the market in the next 2-3 years time. This means that when Windows or Google announces a new gadget, there are already Fabs working with the fabless IC designers to use the old chips and using qualified processes and equipments to produce the product. The fab will work the best it can to improve the yield so as it can fulfil as much order as possible.
3) You have check whether UMS supports Applied Materials more on their new equipments or replacements. Looking at their precision engineering expertise, I am leaning towards replacement capex as I can imagine alot of chambers, robotic arms, belts, rubber protector needs precision engineering work. This may be the highlight of UMS.
My fault was that I quickly assumed UMS only supports Applied Material serving local fabs as their segmentation in Singapore is very large. However, after knowing that Applied Materials Singapore manufactures and supplies their equipment worldwide, then it is a very different story. In that case, UMS is in a unique position to enjoy the growth of global semicon equipment industry if overall business condition picks up and more 200mm fabs shut down to make way for the bigger 300mm or 450mm wafers as Zelphon has shared.
(24-07-2012, 03:16 PM)Nick Wrote: In the long run, with new Microsoft 8 OS, more tablets, more smartphones, more electronic items to cater to a growing wired world, it is likely that semiconductor plants will have to invest in new equipment to meet this demand. On the flipside, a sustained recession brought about by the West fiscal crisis and slower growth in China will reduce demand accordingly and the huge over-capacity will kill spending for a few years.
One should not assume that there is a direct correlation with the increase in new electronic products equivalent to more capex spending.
Are you saying there is no correlation at all between increase in new electronic products and increase in capex spending? If it is a finished semiconductor product, capex had already been spent to produce it. If it is a semiconductor yet to be produced, maintenance capex (at least) as you said needs to be spent. Therefore, I think there is a correlation.
Hi Boon,
You have to take into context on my earlier statement on the correlation that I am referring to the supply has outstripped demand into context at this point of time. In the early 2000s, there were low tech fabs in China and almost none in Middle East, but now many fabs have established in these lower cost countries and have pushed the ASP of wafers down that the industry is suffering.
There are some points that you should consider:
1) You made an assumption that the study you have shared in the forum to be relevant with current times. I disagree as no doubt barrier of entries of semicon industries are high, countries like China, Korea and Middle East have actively joined in the game recently because of government incentives on increasing the tech exposure for the country etc. This means that the oversupply situation is still ongoing and I am unsure how the industry is consolidating.
2) The study does not correspond with what I see on the actual ground (which was 2 years ago). Like I mentioned in my earlier post, most new tech product that was recently tech qualified are only expected to be in the market in the next 2-3 years time. This means that when Windows or Google announces a new gadget, there are already Fabs working with the fabless IC designers to use the old chips and using qualified processes and equipments to produce the product. The fab will work the best it can to improve the yield so as it can fulfil as much order as possible.
3) You have check whether UMS supports Applied Materials more on their new equipments or replacements. Looking at their precision engineering expertise, I am leaning towards replacement capex as I can imagine alot of chambers, robotic arms, belts, rubber protector needs precision engineering work. This may be the highlight of UMS.
My fault was that I quickly assumed UMS only supports Applied Material serving local fabs as their segmentation in Singapore is very large. However, after knowing that Applied Materials Singapore manufactures and supplies their equipment worldwide, then it is a very different story. In that case, UMS is in a unique position to enjoy the growth of global semicon equipment industry if overall business condition picks up and more 200mm fabs shut down to make way for the bigger 300mm or 450mm wafers as Zelphon has shared.
Yup..
450mm will happen but not so soon..
200mm fabs are still going strong because the equipment depreciation is zero and any wafers being sold or manufactured = PURE PROFITS..
Now the whole industry is in consolidation mode...
Big equipment makers become bigger but M&A..
Likewise big foundries become bigger due to closure of weaker rivals..
There are many cases of bankruptcy in Japan and Taiwan...
Chartered was sold off to to become Global Foundries...
This trend continues to persist because of over capacity issues driving DRAM prices etc to record low...Depressing margins...
Look at the TV, smartphones, etc..
All getting cheaper and cheaper..
The korean fabs are bascially fighting a price war with the taiwanese on cost... and Korean are winning... Samsung squashed HTC big time in smartphones...
So let's narrow down the big picture to UMS..
UMS is basically supporting Applied Materials Manufacturing..
As Applied gets bigger, they can still command the price margin as they monopolised certain segments of the market...
So when there are new fabs (i.e. new capex), Applied tool sales will go up.. Meaning UMS will have higher revenue..
The key is whether will there be new fabs considering the shitty state of the world economy...
450mm is just starting.. That is a confirmation..
But whether the pace will accelerate, that will depend on global economy sentiments...
Homepage
PR Newswire
