Samsung Foundry in Trouble, Might Cancel 1.4 nm Node High-Volume Manufacturing

[AleksandarK]

Samsung Foundry could abandon its 1.4 nm (SF1.4) process node initially targeted for 2027 production, according to industry leaker @Jukanlosreve. This decision comes amid ongoing yield problems with the company's 3 nm SF3 node and follows the shutdown of underutilized 5 nm and 7 nm production lines. This could significantly impact Samsung's technology roadmap, which had positioned SF1.4 alongside automotive-focused SF2A and SF2Z nodes. According to Korea Economic Daily data, the foundry division's market share has fallen to 8.2% compared to TSMC's 67.1%. Samsung continues the development of Exynos 2600 on the SF2 node and maintains orders from Japan's Preferred Networks but has failed to attract major clients beyond Chinese firms, avoiding US sanctions. Internal restructuring appears imminent, with reports suggesting the Exynos design team may move under the Mobile Experience division.

While Samsung may prioritize improving yields on existing nodes rather than pursuing SF1.4, this approach risks competitive disadvantages in high-performance computing and AI markets. The SF2Z node with Backside Power Delivery Network (BPDN) technology remains under development, though its commercial viability depends on resolving broader manufacturing issues. Samsung's retreat would further fall behind advanced node manufacturing, competing with TSMC and potentially Intel, raising questions about competition in leading-edge semiconductor fabrication. The company's decisions in the coming months will determine whether it can regain manufacturing credibility or face production of trailing-edge semiconductor nodes. Manufacturing advanced silicon remains a challenge for everyone except TSMC.



View at TechPowerUp Main Site | Source

 

[x4it3n]

Terrible news for the Industry as a whole, we need more competition! TSMC almost have a Monopoly right now

 

[wNotyarD]

When was the last time I've seen news on Samsung Foundry having success?

 

[Athena]

Ugh, this means we are going to be supply constrained for pretty much every product launch for the foreseeable future

hopefully they will fix the yield issue(s)

 

[Daven]

We are nearing the end of the road for node shrinks. It’s time to stack!

 

[Wirko]

We are nearing the end of the road for node shrinks. It’s time to stack!

The best we can do right now in that regard is to put one EUV scanner, or maybe two of them, on top of another EUV scanner. Many-layer DRAM and two-layer logic (CFET) are only coming around 2030 if we're lucky.

 

[Jermelescu]

Ugh, this means we are going to be supply constrained for pretty much every product launch for the foreseeable future

hopefully they will fix the yield issue(s)

As long as you are not that desperate for cutting edge stuff, you should be good

 

[kondamin]

Is there anyone that has gone back a couple of nodes and tried power over via's and GAA-FET on larger nodes?

 

[x4it3n]

As long as you are not that desperate for cutting edge stuff, you should be good

You mean not everybody is waiting for the RTX 6090 already?

 

[Wirko]

Is there anyone that has gone back a couple of nodes and tried power over via's and GAA-FET on larger nodes?

China might do it, at least the BPD part, to squeeze every bit of performance from older nodes.

 

[kondamin]

China might do it, at least the BPD part, to squeeze every bit of performance from older nodes.

Seems like a logical thing to do looking at it from a surface level.

If it can reduce heat generation you should get higher clocks

 

[ScaLibBDP]

>>...according to industry leaker @Jukanlosreve...

Wow! Just Wow! Who is that Magician Guy? I mean @Jukanlosreve...

Let me put it in a different way. Techpowerup is an absolute spreader of tech-rumours and, very often, tech-bullshit-like news!

The red line is crossed and this is my last post. I'll Never visit Techpowerup again.

 

[TheinsanegamerN]

As long as you are not that desperate for cutting edge stuff, you should be good

Even regular stuff is affected by supply constraints. See: the 5070's supply issues.

>>...according to industry leaker @Jukanlosreve...

Wow! Just Wow! Who is that Magician Guy? I mean @Jukanlosreve...

Let me put it in a different way. Techpowerup is an absolute spreader of tech-rumours and, very often, tech-bullshit-like news!

The red line is crossed and this is my last post. I'll Never visit Techpowerup again.

k bye.

 

[Wirko]

>>...according to industry leaker @Jukanlosreve...

Wow! Just Wow! Who is that Magician Guy? I mean @Jukanlosreve...

Let me put it in a different way. Techpowerup is an absolute spreader of tech-rumours and, very often, tech-bullshit-like news!

The red line is crossed and this is my last post. I'll Never visit Techpowerup again.

Well, I guess you can find other sites with an interest in bleeding-edge technology, which only discuss the past or the certain future, but never the uncertain future.

 

[R0H1T]

When was the last time I've seen news on Samsung Foundry having success?

Exynos, if you can call it as that?

 

[Pizdarenkowitch]

(CFET) are only coming around 2030 if we're lucky.

What about B2B Tunnel-Fet?

 

[Quicks]

Exynos is horseshit anyways.
Strange that they don't have buyers for 5nm & 7nm. Surely many don't need a smaller process then that.

Hopefully the get their act together because having TSMC supply basically the entire world is never a good idea.

 

[john_]

improving yields on existing nodes

Probably that's what they have to do. I mean, 3nm and 5nm, when did those became bad? I understand that 1.4nm might look like 4 times better than 3nm, but they keep chasing a more and more difficult target all the time, like a pole vault athlete who, every time he misses a jump, he puts the bar higher, hoping to win the 1st place.

 

[alwayssts]

Probably that's what they have to do. I mean, 3nm and 5nm, when did those became bad? I understand that 1.4nm might look like 4 times better than 3nm, but they keep chasing a more and more difficult target all the time, like a pole vault athlete that, every time he misses a jump, he puts the bar higher, hoping to win the 1st place.

Right. I feel they'll get it together for something similar to what TSMC/Intel have with BSPD, maybe slightly less perf, but only a year or so behind them. Maybe actually similar to Intel tbqh, given their yields.

They truly do need to stop putting the cart before the horse, but OTOH I also understand wanting to hit that target everyone else is attempting to yield right now before they take all the business.

Beyond that truly is silly (expense versus benefit). So what they're doing, while perhaps bad form before; up to this point, and would be in the future, actually kind of does make sense in this specific instance.

People also forget it's not like TSMC is going to be truly mass-producing 2nm (for consumer products outside of Apple/Qualcomm) for quite some time, let-alone even 3nm in the US.

Samsung *does* need to catch up, and sooner would be better than later, but unlike some that think it has to be done by EOY or early next for 2027 devices, I think if they can do it for 2028 stuff they'll be okay.

They would still be in the thick of it for most customers (especially those that don't want to spend more on TSMC/Intel's BSPD and/or limited wafers), and likely would be for at least a couple or more years after imo.

And even then, many won't transition past that offering (just like many didn't transition past other older processes for different reasons).

If they can compete with Intel/TSMC/Rapidus/whomever at that approx level and by that approx time I think they'll be fine. It just-so-happens TSMC was ahead and pushed them harder/faster than able until now.

At some point it's just going to be all about comparative pricing and volume, imo. And in the grand scheme of things, outside tech enthusiasts, it's not all that far away regardless if one, two, or even three years.

 

[kondamin]

Probably that's what they have to do. I mean, 3nm and 5nm, when did those became bad? I understand that 1.4nm might look like 4 times better than 3nm, but they keep chasing a more and more difficult target all the time, like a pole vault athlete who, every time he misses a jump, he puts the bar higher, hoping to win the 1st place.

They are doing it to them selves, they are a gigantic horizontal company and they could push the industry to use more capacity by offering products with more powerful chips and more storage and ram than the competition
which would force those to up their game too.

for some reason they are happy to stick to idiotic small amounts of ram and storage on the devices they sell millions of

 

[mb194dc]

Ugh, this means we are going to be supply constrained for pretty much every product launch for the foreseeable future

hopefully they will fix the yield issue(s)

Unlikely because demand for data centre chips likely goes to near zero. Already way too much supply.

Massive over capacity probably the big concern for 2026.

 

[Philaphlous]

How is SS doing with 3/4/5nm nodes? I haven't really seen anything regarding their foundry work except for this article. If they're competitive at 3/4nm why does the 1.4nm matter much at this point??? Nobody is mass producing 1.4nm or even 2nm to my knowledge... We'll eventually get there but not right now. Makes more sense to capitalize on current tech to remain competitive...

It seemed like back in the 14-28nm nodes that even without a die shrink there seemed to be quite a wide margin of efficiency gains with just changes in architecture...maybe that'll be the case in the 2-3nm nodes too?

 

[sephiroth117]

Nintendo could have used 5 or 7nm lol for the switch 2 if there was no strong demand, apparently they are on 8nm for it...
I suppose Nvidia could not offer something affordable enough besides there 8nm rumored T239 anyways

 

[Jermelescu]

Even regular stuff is affected by supply constraints. See: the 5070's supply issues.


k bye.

The GeForce 5070 is on cutting edge tech for nothing, you have older and arguably better stuff that do the same thing, except a few perks.

 

[kondamin]

The GeForce 5070 is on cutting edge tech for nothing, you have older and arguably better stuff that do the same thing, except a few perks.

3 year old "node" but yeah, those chips have no business being on the most in demand node.
Mainstream graphics could well be made on 7nm.
3 Graphics card vendors isn't enough.