Race to the bottom?

[Barnsley]

As TSMC keeps making strides in advanced nodes, its Taiwanese foundry peers may be facing tough competition on mature nodes from Chinese rivals, including SMIC, Hua Hong Group, and Nexchip, as per Economic Daily News. The report indicates that recently, prices for 12-inch wafers from Chinese foundries have been slashed by as much as 40% compared to the offer from Taiwanese companies.

In addition to the rock-bottom price for 12-inch wafers mentioned above, major foundries in China are also offering an additional 20-30% discount on 8-inch wafers, as indicated by the report.

12-inch Wafer Prices Show the Most Significant Decline

According to the report, Chinese foundries are targeting deals for driver ICs, power management ICs, and MCUs, and they are offering varying discounts depending on the nodes or product types. The steepest price reductions are said to be seen in the 40/45nm nodes, with 12-inch wafer prices—which had previously held relatively firm—showing the most significant declines.

The move has reportedly triggered a wave of order transfers to China from Taiwan’s IC design clients, impacting Taiwanese foundries such as UMC and TSMC affiliate VIS, Economic Daily News notes.

Is it a win if you are competing on price and not quality?

What would be the reason that Chinese companies wouldn't be supporting the Foundries is China?

https://www.trendforce.com/news/new...es-with-an-up-to-40-discount-on-mature-nodes/

 

[OneEng]

I can't speak for all companies of course, but in mine, we have been intentionally "de-risking" production of any Chinese supply issues we are able to and have a strategic backup (albeit more expensive) plan for any parts/services that we do end up sourcing from China.

It is no longer just a matter of quality of Chinese parts, it is a matter of business continuity.... which is even more serious IMO.

With talk of 100% and more tariffs and war in the South China sea, I can't see this situation changing much in the next 5 years at least.

 

[count]

Eventually, as Moore's law slows down, the silicon based chip foundry business will become commoditized. That might be 5 years,10 years, or even 20 years from now. This is the nature of any business, they all go through a cycle of growth, maturity, stagnation, and decline.

I think foundries are a mature business where the leaders have emerged, progress is steady and incremental, and weaker players are failing.

 

[OneEng]

Eventually, as Moore's law slows down, the silicon based chip foundry business will become commoditized. That might be 5 years,10 years, or even 20 years from now. This is the nature of any business, they all go through a cycle of growth, maturity, stagnation, and decline.

I think foundries are a mature business where the leaders have emerged, progress is steady and incremental, and weaker players are failing.

True. I agree with your analysis of the trajectory. It doesn't do any good to make a process that no one can afford to use. Sure, you can say that high value military applications where "cost is not of any concern" will still exist and chips can be produced at any cost for this; however, the military has already been down that road and has fallen back on off-the-shelf chip tech (for the most part) as the development time for a military proprietary chip that rivals anything off the shelf is prohibitively long ..... so long that by the time it is released for use, off the shelf solutions exist that greatly outperform it.

So I believe that this leaves us with fabrication companies that invest carefully in technology nodes, and operate processes that are very efficient .... and provide customers with tools that allow their services to be on-boarded quickly and efficiently .... and produce products at a competitive, market friendly price.

The beading cutting edge of fab tech has become both too risky and too expensive to live on IMO.

 

[Arthur Hanson]

Eventually, as Moore's law slows down, the silicon based chip foundry business will become commoditized. That might be 5 years,10 years, or even 20 years from now. This is the nature of any business, they all go through a cycle of growth, maturity, stagnation, and decline.

I think foundries are a mature business where the leaders have emerged, progress is steady and incremental, and weaker players are failing.

Where do you think chip technology will go from here? Stacking different functions in a single chip may be the future with many surprises to come, never has the semi segment been static for long and there may be many surprises in the future, from technologies, cost, power consumption, memory integration, the possibilities are endless. Any thoughts or comments appreciated.

 

[count]

Where do you think chip technology will go from here? Stacking different functions in a single chip may be the future with many surprises to come, never has the semi segment been static for long and there may be many surprises in the future, from technologies, cost, power consumption, memory integration, the possibilities are endless. Any thoughts or comments appreciated.

Semiconductor technology will continue to improve with diminishing returns, following the classic S curve. Right now we are on the upper side, of the S curve where things are slowing down. Eventually, some new computing paradigm will come along (quantum?) where initially it'll be worse than silicon technology, but it will follow a steeper S curve and blow past silicon making silicon transitions obsolete just like vacuum tubes were made obsolete by silicon.

It should be noted that quantum is currently on a much faster improvement curve than silicon, with the number of qbits in state of the art quantum computers doubling every year or so (and poised to accelerate). This is much faster than silicon where we are down to improvements of only maybe 15% a year these days.

 

[Daniel Nenni]

TSMC is in a strong position to undercut competitors but they may face monopolistic scrutiny. I highly doubt TSMC will risk that. TSMC really is wishing for Intel Foundry to be successful, no matter what Morris Chang says. For Morris it is personal, for CC Wei it is professional.