🔥#Aİ Is the GPU still going up? The real second wave of the market has already begun to erupt from the “chip back-end,” and last night the relevant US-listed stocks all surged together
Many people are still watching graphics cards; actually, the money is quietly flowing to more upstream and more hidden parts
AI capex is shifting from “buying GPUs” → to “building the entire semiconductor factory chain”
🚀 The industry is transmitting like this:
1️⃣ First stage: cloud companies疯狂 buying GPUs / HBM / storage / power
2️⃣ Second stage: TSMC, Samsung, Micron, CXMT, etc. begin expansion
3️⃣ Third stage: equipment → materials → packaging → testing—everything takes a big bite
👉 That is, the real main upswing has already shifted from “computing power” to “the people who build computing power”
⚙️ Next are the five most critical directions:
✔️ Advanced process equipment
✔️ HBM / storage equipment
✔️ Advanced packaging (Chiplet / CoWoS)
✔️ Ultra-high-purity materials
✔️ Domestic substitution chain
These five lines are the core main themes for the next 1–2 years
📦 One easily overlooked focus is packaging + testing—this round is actually seriously undervalued
Previously, the market thought it was a “low-margin labor segment,” but now with AI + HBM + advanced packaging coming in:
👉 It becomes a core positioning point close to front-end manufacturing
👉 Customer qualification + technical barriers directly raise the stakes
The logic has changed.
🌍 If you look at global leaders—ASML, AMAT, LAM, KLA, TEL, Advantest, Teradyne—these are certain core bets
If you look at the high-volatility directions—BESI, ASMPT, FormFactor, Ibiden, Shinko—as well as China’s supply chain: North Huachuang, ACM Research, Estun, SMIC, Changchuan, JF Micro, Avant, etc.
🇨🇳 In the long run, what matters more is:
👉 Improving domestic equipment
👉 Replacing domestic materials
👉 Catching up in advanced packaging
This line may be even longer than the AI cycle itself—but keep one point in mind:
The semiconductor chain is too long, and there are too many companies; ordinary people can easily pick the wrong track and chase the热点
👉 A simpler approach is actually to use a semiconductor ETF to capture the dividend of the whole industry chain
(e.g., SOXX / SMH / domestic 159516 / 588710, etc.)
What do you think this round’s semiconductor main theme will be strongest in: equipment, packaging, materials, or domestic substitution?
Click my profile to follow me—I’ll gauge the consensus and help you break down the directions where the next round of funds may flow🔥
Many people are still watching graphics cards; actually, the money is quietly flowing to more upstream and more hidden parts
AI capex is shifting from “buying GPUs” → to “building the entire semiconductor factory chain”
🚀 The industry is transmitting like this:
1️⃣ First stage: cloud companies疯狂 buying GPUs / HBM / storage / power
2️⃣ Second stage: TSMC, Samsung, Micron, CXMT, etc. begin expansion
3️⃣ Third stage: equipment → materials → packaging → testing—everything takes a big bite
👉 That is, the real main upswing has already shifted from “computing power” to “the people who build computing power”
⚙️ Next are the five most critical directions:
✔️ Advanced process equipment
✔️ HBM / storage equipment
✔️ Advanced packaging (Chiplet / CoWoS)
✔️ Ultra-high-purity materials
✔️ Domestic substitution chain
These five lines are the core main themes for the next 1–2 years
📦 One easily overlooked focus is packaging + testing—this round is actually seriously undervalued
Previously, the market thought it was a “low-margin labor segment,” but now with AI + HBM + advanced packaging coming in:
👉 It becomes a core positioning point close to front-end manufacturing
👉 Customer qualification + technical barriers directly raise the stakes
The logic has changed.
🌍 If you look at global leaders—ASML, AMAT, LAM, KLA, TEL, Advantest, Teradyne—these are certain core bets
If you look at the high-volatility directions—BESI, ASMPT, FormFactor, Ibiden, Shinko—as well as China’s supply chain: North Huachuang, ACM Research, Estun, SMIC, Changchuan, JF Micro, Avant, etc.
🇨🇳 In the long run, what matters more is:
👉 Improving domestic equipment
👉 Replacing domestic materials
👉 Catching up in advanced packaging
This line may be even longer than the AI cycle itself—but keep one point in mind:
The semiconductor chain is too long, and there are too many companies; ordinary people can easily pick the wrong track and chase the热点
👉 A simpler approach is actually to use a semiconductor ETF to capture the dividend of the whole industry chain
(e.g., SOXX / SMH / domestic 159516 / 588710, etc.)
What do you think this round’s semiconductor main theme will be strongest in: equipment, packaging, materials, or domestic substitution?
Click my profile to follow me—I’ll gauge the consensus and help you break down the directions where the next round of funds may flow🔥