Samsung Delays $44 Billion Texas Chip Fab Because 'There Are No Customers' (tomshardware.com) 38
An anonymous reader quotes a report from Tom's Hardware: Samsung is reportedly delaying the launch of its Taylor, Texas, fab, citing difficulties in securing customers for its output. Sources told Nikkei Asia that even if the South Korean chipmaker brings in the necessary equipment to produce chips at the new plant, the company cannot do anything with them due to the lack of demand. Aside from that, the original planned process node for the Taylor plant is no longer aligned with current demand, highlighting the rapid pace of semiconductor technology.
The chip maker started construction on the Taylor fab in 2022, with an initial investment of $17 billion. By 2024, the company decided to double this to $44 billion, with the addition of another advanced fab and expanded R&D operations. This move is supported by a $6.6-billion CHIPS Act subsidy, which was finalized in December last year, despite multiple delays and setbacks. Samsung C&T, the primary contractor for the Taylor fab, states that construction of the site is progressing. Documents from the company show that the site is almost 92% complete as of March 2024. Work on the site was originally scheduled to finish the following month, but regulatory filings indicate that this was moved to October.
No reason was given for the delay, but multiple sources indicate that it occurred due to a lack of demand. It was initially planned for the Taylor Fab to produce chips for the 4nm process node, but this has since been upgraded to 2nm, to compete with TSMC and Intel. A supply chain executive told the publication that there is little demand for the originally planned 4nm process node at the site. "Local demand for chips isn't particularly strong, and the process nodes Samsung planned several years ago no longer meet with current customer needs," the executive said to Nikkei Asia. "However, overhauling the plant would be a major and costly undertaking, so the company is adopting a wait-and-see approach for now." Although it has already declared its intention to upgrade the site to manufacture the 2nm process node, that is a resource-intensive task in terms of time, effort, and money. Despite the lack of customers, Samsung says it will proceed with opening the Taylor Fab by 2026 -- a necessary move to qualify for CHIPS Act funding and avoid falling behind competitors like TSMC. Delaying further could jeopardize billions already invested in the project.
The chip maker started construction on the Taylor fab in 2022, with an initial investment of $17 billion. By 2024, the company decided to double this to $44 billion, with the addition of another advanced fab and expanded R&D operations. This move is supported by a $6.6-billion CHIPS Act subsidy, which was finalized in December last year, despite multiple delays and setbacks. Samsung C&T, the primary contractor for the Taylor fab, states that construction of the site is progressing. Documents from the company show that the site is almost 92% complete as of March 2024. Work on the site was originally scheduled to finish the following month, but regulatory filings indicate that this was moved to October.
No reason was given for the delay, but multiple sources indicate that it occurred due to a lack of demand. It was initially planned for the Taylor Fab to produce chips for the 4nm process node, but this has since been upgraded to 2nm, to compete with TSMC and Intel. A supply chain executive told the publication that there is little demand for the originally planned 4nm process node at the site. "Local demand for chips isn't particularly strong, and the process nodes Samsung planned several years ago no longer meet with current customer needs," the executive said to Nikkei Asia. "However, overhauling the plant would be a major and costly undertaking, so the company is adopting a wait-and-see approach for now." Although it has already declared its intention to upgrade the site to manufacture the 2nm process node, that is a resource-intensive task in terms of time, effort, and money. Despite the lack of customers, Samsung says it will proceed with opening the Taylor Fab by 2026 -- a necessary move to qualify for CHIPS Act funding and avoid falling behind competitors like TSMC. Delaying further could jeopardize billions already invested in the project.
Re:MAGAs will kill it (Score:4, Insightful)
Re: (Score:1)
Not a cell phone in sight. Just people living in the moment. Democrats want to take this away from you.
Re: (Score:1)
The real problem is we can't return to being a key chip manufacturing nation overnight. You have to fall down to learn how to walk.
Re:MAGAs will kill it (Score:4, Interesting)
Re: (Score:2)
Re: MAGAs will kill it (Score:3)
Uh, did you not finish reading even the headline?
If no one wants to buy their output, why build the Fab?
The concept you want to research is called 'market saturation', and if you have the time, consider looking into the concept of 'overproduction'...
Re: (Score:2)
He has a point.
Samsung don't have to tell you what the real reasons are. Corporations lie all the time, and sometimes they lie for diplomatic reasons while walking quickly away from you.
Local Demand? (Score:2, Interesting)
"Local demand for chips isn't particularly strong, and the process nodes Samsung planned several years ago no longer meet with current customer needs," the executive said to Nikkei Asia.
Local demand? Where did they think those gigantic factories were, Oklahoma?
Re: (Score:1)
No, Samsung has problems...
Re:Local Demand? (Score:4, Interesting)
"Local Demand" in the context of large factories, means "country-level"
Basically the reason for the plant existing is in case South Korea is invaded or bombed by the North. Same with TSMC having their arm twisted to have a factory in the US. The last thing the US needs is for both their secondary supplies of chips to be gone.
Remember TI (Texas Instruments), what ever happened to them? They acquired 13 companies since their inception, including National Semiconductor (Remember Cyrix before it was give away to S3/Via before getting offloaded to AMD)
Like I kid you not, All Samsung has to do is make SSD's and there will be local buyer. The reason there is no local demand is because they are looking for Apple, Nvidia or even Intel or AMD to use the fab. They don't want to risk making a product that will flop or have to be firesale'd
But really they want the US Government to use them, why else would the US want to subsidize a chip plant at all.
Re: Local Demand? (Score:2)
They planned and designed a Fab to build type of chip nobody wants - it's back to the drawing board.
The planned Fab was going to make chips using a technology no one wants now, markets change, why build a fab if you can't sell the products it makes?
Re: (Score:2)
But really they want the US Government to use them, why else would the US want to subsidize a chip plant at all.
Because politicians want PR opportunities, and they think that encouraging local manufacturing attracts votes.
Re: (Score:2)
So if the chip is produced in the US then it gets shipped to a contract manufacturer that put onto a board that gets sent to a factory that then puts the boards into a product.
This is the problem right here. The middle levels of manufacturing are all overseas. So the 'Made in Texas' chips would have to go overseas, placed on boards in foreign factories and then in products made in other foreign factories. And after all these foreign factories get involved, tariffs still apply. Or if they are proportional to content origin, too damned tough to track. And if some Asian war shuts down shipping, we still don't have the middle levels of manufacturing in place over here to take over bu
Re: Local Demand? (Score:2)
Who would have guessed? (Score:3)
How could anyone have predicted in 2022 that technology would have advanced by 2025? It's almost like competent management means planning for the future.
Re: (Score:2)
The problem is that Samsung's technology isn't advancing.
Re: (Score:2)
Nah the problem is you're talking out of your arse. Samsung's transistor density in its current node is on par with Intel's best, and only slightly behind TSMC. They are targeting 1.4nm node by 2027, and all three companies have demonstrated next gen CFET production recently.
Re: (Score:2)
Re: (Score:2)
It depends, which "4nm" node are we talking about here? TSMC N4P or Samsung 4LPP? N4P is still in use and in fairly high demand. 4LPP not so much (ditto for 4LPE).
So They're Returning the Subsidies, Right?! (Score:2)
Samsung bleeding customers (Score:2)
They aren't keeping up with TSMC, period. Even their cutting-edge nodes like 3GAP (or whatever the fuck they call it now) aren't hitting acceptable yields or attracting foundry customers. They are in trouble.
Re: (Score:2)
Re: Nintendo should use it (Score:2)
Yeah, that justifies a $44BN chip fab to lower the price of Nintendo game consoles...
They're all waiting for the tidal wave from China (Score:3)
Chinese companies were happily churning out 20nm chips for washing machines, then someone in Washington thought it'd be a great idea to sanction Chinese semiconductors for, um, reasons. The Chinese government noticed and within six months it rolled out a plan: By 2030 have a world leading full stack semiconductor industry using 100% Chinese owned IP from top to bottom.
Because of our sanctions in just four years we're going to have the world's undisputed leader in mass production churning out semiconductors in quantities and at prices that'll make semiconductor investment in the west a money losing proposition. The only solution for the west is to put up a wall preventing the import of Chinese semiconductors, but the problem is that the west no longer makes anything that the rest of the world can't get from China faster, cheaper and better.
The western semiconductor industry knows its in deep trouble.
Re: (Score:1)
Re: (Score:2)
How about Raspberry Pi? (Score:2)