Once again a post where I had to laugh hysterically. I’m a bit puzzled. I was a super semicap bull in late 2024/ early 2025 and was rather surprised by the lack of traction given the very apparent surge in demand into last year.
Something postponed is not something abandoned and thus have significantly increased my semicap exposure since October. Iirc asml basically said during their investor’s day that they were sitting on news and waiting for the new year for further announcements. Should make for a very good read on the whole situation.
I’m very interested if we’ll see a race to high-NA now. Intel has that EUV trauma from Intel 10nm++++++++++, which, as far as I understand, is at least partially to blame on them being late to EUV besides that whole cobalt drama fueled by infinite arrogance.
Intel and Samsung have got their test machines, don’t see tsmc delaying much further considering the traction Samsung and Intel are apparently regaining.
One thing I’m very curious about is given the incredible state of the memory sector while the Chinese have now been gaining ground, what crazy manufacturing and product technologies will now be investigated by the big three on the off-chance they work? if there ever was a time for experiments, it is now.
Phenomenal call on the 18A yield issues way ahead of everyone. The Qualcomm-Samsung relationship as a capacity barometer is genius, I hadn't thought of using DX procurement decisionsas a signal for foundry health. The panic thesis makes sense given how memory and packaging sucked up all attention while logic tightness built quietly. One thing i'm curious about is whether Intels 18AP fixes are fundamentaly architectural or just more time for yield learnings to propagate, because the PDK issues you flagged last year sounded pretty deep.
Excellent fucking article! no supply train checking, no speaking right in social medias! this is why so many REAL rumours in X, especially source from korean medias. for my exprience in fab, foundry job is keeping improvement progress, the key is know how experience, which all need time to fine tune, unless steal from competitor!
Samsung looks like a winner on multiple front here. Memory shortage - check. No TSMC capacity = basically captive customers for foundry - check. Apple will have to raise prices for iphones by $150-200 because of dram & wafer pricing = higher margins or sales for Galaxy - check.
Samsung, in a surprising feat of abnormally-competent management, has Memory and Electronics do deals with each other at arms' length. It seems unlikely they would subsidize the phone business to gain material market share vs Apple (and I'm not sure this would work). I do think on the face of it, this makes sense though
Or used to be till AMD hoisted itself on that petard and gave up.
Anyway coming back to the topic on hand. I owned some KLAC and forgot to pay attention so was scratching my head a bit as to why it’s up so much. Go look it up. It’s beating all the big semi cap names for the past 5 years that I checked against and heck even MU (which I own and talked about in a past comment that you buy when beaten up).
My theory when I bought KLAC was look here is a semi cap name that is very important but few seem to talk about it or understand it. As things keep getting more complex, gotta pay the “yield tax”.
Samsung Electronics isn’t really a pure play, they have many interlinked businesses. Also foreign, family controlled chaebols are not everyone’s cup of tea to invest in.
My only concern is leading edge consumer electronics is about to take a nose punch otherwise agree. Even with consumer downturn you are probably right on semicap
Once again a post where I had to laugh hysterically. I’m a bit puzzled. I was a super semicap bull in late 2024/ early 2025 and was rather surprised by the lack of traction given the very apparent surge in demand into last year.
Something postponed is not something abandoned and thus have significantly increased my semicap exposure since October. Iirc asml basically said during their investor’s day that they were sitting on news and waiting for the new year for further announcements. Should make for a very good read on the whole situation.
I’m very interested if we’ll see a race to high-NA now. Intel has that EUV trauma from Intel 10nm++++++++++, which, as far as I understand, is at least partially to blame on them being late to EUV besides that whole cobalt drama fueled by infinite arrogance.
Intel and Samsung have got their test machines, don’t see tsmc delaying much further considering the traction Samsung and Intel are apparently regaining.
One thing I’m very curious about is given the incredible state of the memory sector while the Chinese have now been gaining ground, what crazy manufacturing and product technologies will now be investigated by the big three on the off-chance they work? if there ever was a time for experiments, it is now.
18A-P 1.0PDK… Do you smell that? What’s that smell? “Opportunity”. What? No. M-O-N-E-Y. I smell money!
Question is panic buy or panic sell?
Phenomenal call on the 18A yield issues way ahead of everyone. The Qualcomm-Samsung relationship as a capacity barometer is genius, I hadn't thought of using DX procurement decisionsas a signal for foundry health. The panic thesis makes sense given how memory and packaging sucked up all attention while logic tightness built quietly. One thing i'm curious about is whether Intels 18AP fixes are fundamentaly architectural or just more time for yield learnings to propagate, because the PDK issues you flagged last year sounded pretty deep.
Excellent fucking article! no supply train checking, no speaking right in social medias! this is why so many REAL rumours in X, especially source from korean medias. for my exprience in fab, foundry job is keeping improvement progress, the key is know how experience, which all need time to fine tune, unless steal from competitor!
Samsung looks like a winner on multiple front here. Memory shortage - check. No TSMC capacity = basically captive customers for foundry - check. Apple will have to raise prices for iphones by $150-200 because of dram & wafer pricing = higher margins or sales for Galaxy - check.
Samsung, in a surprising feat of abnormally-competent management, has Memory and Electronics do deals with each other at arms' length. It seems unlikely they would subsidize the phone business to gain material market share vs Apple (and I'm not sure this would work). I do think on the face of it, this makes sense though
But but … “Real men own fabs.”
Or used to be till AMD hoisted itself on that petard and gave up.
Anyway coming back to the topic on hand. I owned some KLAC and forgot to pay attention so was scratching my head a bit as to why it’s up so much. Go look it up. It’s beating all the big semi cap names for the past 5 years that I checked against and heck even MU (which I own and talked about in a past comment that you buy when beaten up).
My theory when I bought KLAC was look here is a semi cap name that is very important but few seem to talk about it or understand it. As things keep getting more complex, gotta pay the “yield tax”.
So if logic is so constrained and Samsung is improving, why not directly bet on Samsung?
Samsung Electronics isn’t really a pure play, they have many interlinked businesses. Also foreign, family controlled chaebols are not everyone’s cup of tea to invest in.
My only concern is leading edge consumer electronics is about to take a nose punch otherwise agree. Even with consumer downturn you are probably right on semicap
Thank you so much for this article. Do you like AEM in Singapore as well?