Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/samsung%E2%80%99s-u-s-taylor-facility-reportedly-becomes-top-choice-for-customers-looking-beyond-tsmc-as-intel-struggles-with-%E2%80%9Cexecution-challenges%E2%80%9D.24395/
        )

    [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] => 2030770
            [XFI] => 1060170
        )

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

Samsung’s U.S. Taylor Facility Reportedly Becomes Top Choice for Customers Looking Beyond TSMC, as Intel Struggles with “Execution Challenges”

Fred Chen

Moderator
Muhammad Zuhair Jan 19, 2026 at 11:59am EST

Samsung's chip operations in the US are now gaining the spotlight, as a report reveals the Korean giant is attracting massive interest from fabless customers.

Samsung's Prior Commitments With the Likes of NVIDIA & Apple Gives Them a Massive Edge Over Intel Foundry​

Ever since fabs have started investing heavily in the US, there has always been a debate about the supply chain hierarchy, or who customers would go to as their secondary foundry option. Traditionally, the discussion hasn't taken much traction, given that the primary supplier, TSMC, has managed to meet chip demand for every fabless client, but with the AI frenzy, there's a severe supply chain bottleneck. This has prompted firms such as AMD, NVIDIA, Qualcomm, Apple, and many others to seek a secondary supplier, and according to Deutsche Bank (via Jukan), Samsung Foundry is leading the race.


It is claimed that Samsung's Taylor fab is the next place for foundry customers looking for semiconductor services in the US, arguing that Intel's execution is facing challenges. The report notes that customers such as Qualcomm and AMD are looking to place orders with Samsung, likely around the company's 2nm process (SF2 and derivatives). The attraction of the Korean giant stems from its having already secured clients like NVIDIA and Apple for its foundry division, so customer confidence is higher relative to the likes of Intel.

A timeline chart compares 'FinFET' and 'GAA (MBCFET)' technologies from 2019 to 2027, highlighting nodes like 'SF5 (5LP)' for FinFET and 'SF3 (3GAP)' for GAA.
Image Credits: Samsung

Recently, reports disclosed that Samsung has decided to prioritize its Taylor facility by shifting production lines to 2nm and that the company is also investing in advanced packaging lines. Fabless manufacturers are looking for nodes that serve as viable replacements for TSMC's N3 process for now, and the only options available are Intel's 18A and Samsung's SF2, which are superior in terms of node class. The inclination of US manufacturers towards both Intel and Samsung clearly shows that there is room for more players in the chip industry.

It's important to note that foundry orders have yet to materialize in production lines, and this is the next major step for both Samsung and Intel as fabs. We do know that both chip manufacturers are racing to gain access to customers looking to invest in a 'dual-sourcing' chip strategy, so it will be interesting to see how the competition unfolds.

 
Last edited by a moderator:
I could not find the reference to "execution challenges", but it seems a couple of reasons Samsung would be currently preferred over Intel are Samsung is already well-known (though maybe not entirely favorably) in foundry world as a TSMC alternative, and people are waiting to see how Panther Lake turns out.
 
I know Samsung is cutting deals to get customers. I am wondering if Intel will as well. and then whether customers will look to have multiple sources.

Once metric to look at: Over the next couple years does the increase in dollar external wafer foundry revenue grow faster at Samsung or Intel. Obviously Intel has to have larger percentage growth (they have no revenue today). But just in dollars.

When TSMC says no, who is getting more of the orders?
 
I could not find the reference to "execution challenges", but it seems a couple of reasons Samsung would be currently preferred over Intel are Samsung is already well-known (though maybe not entirely favorably) in foundry world as a TSMC alternative, and people are waiting to see how Panther Lake turns out.

Samsung Foundry is well known for horrible yield and broken promises.

Since Intel 18A is already in HVM and Samsung 2nm HVM is targeted for 2027 (hopefully) I give Intel the advantage. The question is price. Samsung has in the past dumped wafers to get production rolling. Will customers take the Samsung risk for a 20-30% price cut on 2nm wafers? Will Intel be able to be price competitive to Samsung?

Given the size of TSMC 2026 CAPEX you can bet that N2 will again dominate the top 10 semiconductor companies for high volume products. Intel, Samsung and Rapidus will be scrambling for 2nm table scraps.

Intel Foundry 18A is focused on 6 big customers but not necessarily for high volume products. If successful at 18A, Intel 14A could get some high volume products. It should be an interesting call today. Let's see what Lip-Bu has planned.

I am really hoping that Apple gives Intel 18AP a try for the M series products. I seriously doubt Apple, AMD or Nvidia will work with Samsung.
 
Nvidia is the most likely to work with Samsung out of the three

Maybe but Samsung does not have competitive packaging. They would have to use a TSMC CoWos packaging partner. Packaging really has turned out well for TSMC, great execution.

I think QCOM will use Samsung 2nm for Samsung products. Tesla also says they will. I'm sure the Taylor fab will be full since they can only do 50-100k wafers per month with maybe 50% yield?

I heard recently that Broadcom is playing with Rapidus and looking at Intel 18A again. The NOT TSMC demand is real. Exciting times in the semiconductor ecosystem, absolutely.
 
Maybe but Samsung does not have competitive packaging. They would have to use a TSMC CoWos packaging partner. Packaging really has turned out well for TSMC, great execution.
I was thinking of client dGPU Nvidia did client dGPU on Samsung 8nm they can always use Samsung for client where advanced packing is not necessary.
 
Back
Top