Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/foundries-the-best-business-model.5751/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Foundries, the Best Business Model

Arthur Hanson

Well-known member
The Foundry/Ecosystem business model is the best business model for driving forward technology, the economy, technological diversity, business diversity, lower costs and creativity. This applies to almost any sector substantial Foundry/Ecosystems are set up, such as semiconductor, bio, 3D/Advanced Fab, and Autos (Mclaren/Lotus). Foundries/Ecosystems allow for greater technological diversity and speed of development by allowing companies to pursue projects where literally massive costs are widely spread by the foundry and their ecosystem partners. TSM is one is probably the world's largest, most successful foundry with the largest ecosystem. Like Morris Chang has stated many times, even the largest competitor is small when compared the entire resources a foundry and its ecosystem bring to bare. The foundry/ecosystem models across many areas are starting to feed on themselves and have changed all the business rules in a way that is bring unprecedented development and adoption speed unimaginable just twenty years ago. It has brought more opportunities to people and businesses than any other development in the history of man. It transcends borders, cultures and entrenched, protected business monopolies, letting more people and businesses realize their full potential. It is the most efficient economic model yet developed and even gives the outsider opportunities to participate in this boom by investing in it. Companies from the world's largest, Apple, down to one man operations have used the foundry model to extend their reach and opportunity by either using it or participating in the ecosystem.

The Foundry/Ecosystem model will by its very nature change and evolve. Since so many of the SemiWiki community participate in this model I hope to hear about many points and observations that I have overlooked. Modern communications and transport have allowed this model to be all encompassing world wide, making it an even more powerful tool through vastly increased scale. I'm sure there are many adaptions and opportunities in the foundry model and many more foundry opportunities in areas where they aren't currently. This is the area I'm looking for opportunities in myself. I hope many, many others can do the same.
 
Last edited:
Good points.

When I look at Samsung I see both a foundry and an IDM combined. This combination will soon make Samsung the #1 semiconductor company in the world, as measured by revenue, displacing Intel.
 
Daniel, It will be interesting to see what company ends up with the largest market cap just for it's semiconductor division. I feel the inherent conflict of interest between IDM and foundry like Samsung has, leads to too many conflicts of interest like those between Apple and Samsung. I wish I had just 10% of what the two of them spent on legal action. If they are willing to abuse their relationship with Apple, imagine what they would do to everybody else. I think we are going to see a rise of pure play foundries with surrounding ecosystems around the world. This can bring a staggering amount of resources to bare on any challenge. Look at what this model has done to the oil industry, we went from peak oil to at least a five hundred year supply using numerous technologies that are still advancing at a staggering rate. The next step will be to bring several foundry/ecosystems together to tackle the challenges the world faces. These foundry/ecosystems are the result of communications and transport not available economically just thirty years age and are still a work in progress and I see combining several of them together will have profound effects on the world as a whole. The other half of the equation will be companies that aggregate different foundry/ecosystems together to meet ever greater challenges. All this is going to happen at an ever accelerating rate that are going to have profound unintended consequences both good and bad. This is why it is critical to think about this now. Power corrupts and absolute power corrupts absolutely. Knowledge is power and multiple foundry/system models bring unprecedented power from around the world into the hands of a few. Apple as the world's largest corporation by far, with more free cash than the majority of countries bring an entirely new wrinkle to the society of the entire world. This forum is the ideal place to open this discussion get actual participants in it thinking.
 
Arthur, you may be missing a point here in that IDMs like Intel and Samsung can in theory fund their fab development costs out of product sales (phones for Samsung, CPUs for Intel) which make more money than wafer sales by foundries like TSMC, where the product margin goes to their customers. Then they could sell wafers at unbeatable prices because the fab is already "paid for".

Not saying that they *do* this, but they could. They could then undercut TSMC on pricing, it's effectively a cross-subsidy from product sales but there's nothing illegal about that.
 
Big companies are definitely not as innovative as small ones, for a variety of reasons. So to the extent the foundry model has unlocked the possibility for smaller companies to operate in semiconductor chip making, it has enabled innovation.
 
Big companies are definitely not as innovative as small ones, for a variety of reasons. So to the extent the foundry model has unlocked the possibility for smaller companies to operate in semiconductor chip making, it has enabled innovation.

I certainly don't want to imply that small companies are not innovative, but in this area we have:
  • Apple, the biggest company of all by market cap, fabless
  • Qualcomm, the inventor of the original CDMA technology and the leader in chips for mobile
  • Intel, TSMC and more developing FinFETs to high-volume manufacturing
  • Cadence and Synopsys completely revamping their physical design and analysis tools (Innovus and ICC II)
  • Mentor driving Calibre to deliver all the physical verification and OPC needed to make any of this stuff work
  • Not enabled by fabless since memory companies don't do fabless, but 3D NAND flash from Samsung, Intel, Micron
 
Arthur, you may be missing a point here in that IDMs like Intel and Samsung can in theory fund their fab development costs out of product sales (phones for Samsung, CPUs for Intel) which make more money than wafer sales by foundries like TSMC, where the product margin goes to their customers. Then they could sell wafers at unbeatable prices because the fab is already "paid for".

Not saying that they *do* this, but they could. They could then undercut TSMC on pricing, it's effectively a cross-subsidy from product sales but there's nothing illegal about that.
t
IanD, One thing you have to consider on cross subsidies is that Samsung is in many low margin businesses like appliances and TVs. Only Apple makes big money from cell phones and much of that is their ecosystem. Their competition in many areas would like them to drain money into other areas.
 
t
IanD, One thing you have to consider on cross subsidies is that Samsung is in many low margin businesses like appliances and TVs. Only Apple makes big money from cell phones and much of that is their ecosystem. Their competition in many areas would like them to drain money into other areas.

The fact remains that the profit margin on a Samsung phone is an order of magnitude bigger than on a Samsung phone CPU, this is the benefit of being an IDM. Apple is different to everyone else in the phone industry due to their application tie-in and profit from this, and high product margins. Intel is different to everyone else in the chip industry because they ship large volumes of expensive high-margin products on the back of dominating the x86 CPU market.

I have heard rumours of unfeasibly low Intel foundry wafer pricing which could not be sustained by a foundry like TSMC who make all their money from wafers, maybe another example of Intel "contra-revenue"; I don't know if Samsung do the same, but they could if they wanted to.
 
Back
Top