🐦 推特SerenityV1 · 实盘不做把握 8/10
北京时间·🕒 发布 04-30 15:00:26时区推断·⟳ 抓取 06-03 19:00:02回填·⚡ 决策 06-17 05:39:14·132.7 天前
KOL推荐韩国HBM设备商Auros,以SIVE和AEHR作为业务类比
📄 原文Serenity↗ 原文
Samsung/SK Hynix Supplier Korean Stock Idea: It's kinda like $SIVE + CPO mixed with $AEHR but for memory (and less TAM)…
不做 SIVE.US
纯类比标的。KOL实际推荐的是韩国股票Auros(322310),仅借用SIVE的业务逻辑做比喻,对SIVE无实质利好或资金引流效应。
不做 AEHR.US
纯类比标的。KOL实际推荐的是韩国股票Auros(322310),仅借用AEHR的业务逻辑做比喻,对AEHR无实质利好或资金引流效应。
展开详情(成交明细 / 研究档案 / 原文全文 / 决策记录链)
研究档案
这条决策没有研究档案。
原文全文
展开原文 · 1,999 字
Samsung/SK Hynix Supplier Korean Stock Idea:
It's kinda like $SIVE + CPO mixed with $AEHR but for memory (and less TAM).
Auros (322310): ~$210m MC.
It's the SK Hynix and Samsung supplier for next-gen HBM.
This is a direct play for HBM4e hybrid bonding (every memory maker) + HBM4 (samsung) / metrology.
A comparison is $SIVE for CPO, where they weren't really around for HBM3 (photonics equivalent is like 800g) and have been developing the past many years. But with next-gen it's there for volume ramp.
It's the only Korean Domestic company for IR metrology. And product #1 is used for hybrid bonding.
So Samsung HBM4 lines use hybrid copper bonding in 2026, so they're trying to push overlay tool through qualification H2?
Sk Hynix actually just used MR-MUF (no hybrid bonding) for HBM4. But Trendforce reported today SK Hynix has completed 12-high hybrid bonding HBM validation and is raising yields for mass production?
So maybe SK Hynix are adopters too soon, which is bullish.
But at least for short-term it's probably Samsung first for volume if qualifications pass.
$KLA is a pure monopoly over this sort of IR metrology overlay that Korea (SK Hynix/Samsung) wants to break.
So Samsung they are likely using Auros to customize IR tools and keep domestic secrets.
Think of it as the Korean $INTC where they will use their domestic champion as much as possible if it actually works. And both will ramp up HBM4E -> HBM5 with this technology, so multiple years capex.
For product #2 (this is why I mention $AEHR, similar):
There's also the "AUROS Technology Inc. is conducting qualification testing of its ultra-precision thin-film thickness measurement equipment with a major domestic chipmaker".
This was dominated by $KLA and $AMAT before.
"Development began at its research center, followed by pilot manufacturing at a new facility from 2021, with about four years required to reach the current qualification stage"
Basically they've been developing many years
It's kinda like $SIVE + CPO mixed with $AEHR but for memory (and less TAM).
Auros (322310): ~$210m MC.
It's the SK Hynix and Samsung supplier for next-gen HBM.
This is a direct play for HBM4e hybrid bonding (every memory maker) + HBM4 (samsung) / metrology.
A comparison is $SIVE for CPO, where they weren't really around for HBM3 (photonics equivalent is like 800g) and have been developing the past many years. But with next-gen it's there for volume ramp.
It's the only Korean Domestic company for IR metrology. And product #1 is used for hybrid bonding.
So Samsung HBM4 lines use hybrid copper bonding in 2026, so they're trying to push overlay tool through qualification H2?
Sk Hynix actually just used MR-MUF (no hybrid bonding) for HBM4. But Trendforce reported today SK Hynix has completed 12-high hybrid bonding HBM validation and is raising yields for mass production?
So maybe SK Hynix are adopters too soon, which is bullish.
But at least for short-term it's probably Samsung first for volume if qualifications pass.
$KLA is a pure monopoly over this sort of IR metrology overlay that Korea (SK Hynix/Samsung) wants to break.
So Samsung they are likely using Auros to customize IR tools and keep domestic secrets.
Think of it as the Korean $INTC where they will use their domestic champion as much as possible if it actually works. And both will ramp up HBM4E -> HBM5 with this technology, so multiple years capex.
For product #2 (this is why I mention $AEHR, similar):
There's also the "AUROS Technology Inc. is conducting qualification testing of its ultra-precision thin-film thickness measurement equipment with a major domestic chipmaker".
This was dominated by $KLA and $AMAT before.
"Development began at its research center, followed by pilot manufacturing at a new facility from 2021, with about four years required to reach the current qualification stage"
Basically they've been developing many years
决策记录链 · 只追加
每次领取产生一条记录,重新决策追加一条并以 supersedes 指向被替代的记录(ADR 0014)。
这条决策早于决策账本(2026-10-08 起逐条追加),没有追加记录。