On July 16, $Taiwan Semiconductor (TSM.US)$ It delivered an almost flawless quarterly report.
Data shows the company’s net profit for Q2 surged 77% year-over-year, with revenue, gross margin, operating margin, and Q3 revenue guidance all exceeding market expectations. Meanwhile, the share of advanced-node processes continued to rise, with AI-related orders remaining the biggest growth driver.
This result reaffirms that demand for AI computing power remains robust and global investment in AI infrastructure has not cooled down.
Yet the market responded very differently.Following the earnings release, Taiwan Semiconductor shares fell more than 2%. Instead of chasing the positive news, investors chose to take profits after the results were realized.
This reflects a new phase in AI investing: previously, the market asked whether AI demand would surge; now, it is questioning whether AI capital expenditures could become excessive and whether overcapacity might emerge in the future.
Just how strong was this earnings report?
Judging by the numbers alone, Taiwan Semiconductor remains at the heart of the AI wave. Over the past few years, the company has transformed from a traditional wafer foundry into a core manufacturing platform for AI infrastructure.
NVIDIA GPUs, AMD AI accelerators, and cloud providers’ custom ASICs all heavily rely on Taiwan Semiconductor’s advanced process nodes. Second-quarter growth primarily came from: continued ramp-up of the 3nm advanced node, rising demand for AI GPUs, and increased HPC (high-performance computing) orders. A...
Data shows the company’s net profit for Q2 surged 77% year-over-year, with revenue, gross margin, operating margin, and Q3 revenue guidance all exceeding market expectations. Meanwhile, the share of advanced-node processes continued to rise, with AI-related orders remaining the biggest growth driver.
This result reaffirms that demand for AI computing power remains robust and global investment in AI infrastructure has not cooled down.
Yet the market responded very differently.Following the earnings release, Taiwan Semiconductor shares fell more than 2%. Instead of chasing the positive news, investors chose to take profits after the results were realized.
This reflects a new phase in AI investing: previously, the market asked whether AI demand would surge; now, it is questioning whether AI capital expenditures could become excessive and whether overcapacity might emerge in the future.
Just how strong was this earnings report?
Judging by the numbers alone, Taiwan Semiconductor remains at the heart of the AI wave. Over the past few years, the company has transformed from a traditional wafer foundry into a core manufacturing platform for AI infrastructure.
NVIDIA GPUs, AMD AI accelerators, and cloud providers’ custom ASICs all heavily rely on Taiwan Semiconductor’s advanced process nodes. Second-quarter growth primarily came from: continued ramp-up of the 3nm advanced node, rising demand for AI GPUs, and increased HPC (high-performance computing) orders. A...
6
7
On July 14, the White House announced major news: the official launch of the 'Golden Eagle Initiative.'
This is not a traditional system upgrade or firewall, but rather a new AI-driven, nationally led cybersecurity information exchange center.
Crisis Looms: Why Was the 'Golden Eagle Initiative' Launched?
To understand the significance of the 'Golden Eagle Initiative,' one must first confront a harsh reality:AI development has already upended the traditional balance of cyber offense and defense.
In recent years, AI-enhanced cyberattacks have significantly lowered the barrier to entry. Cutting-edge large language models can not only assist in generating zero-day exploit tools but also bypass multi-factor authentication. Moreover, code generated by widely used AI-assisted programming tools (such as GitHub Copilot) often contains security vulnerabilities—effectively mass-producing security weaknesses while accelerating code output.
To address this AI-driven, imminent national information security threat, the Trump administration had already signed Executive Order 14409 on June 2, 2026, titled 'Advancing Innovation and Safety in Advanced Artificial Intelligence.'The executive order explicitly directed the federal government to establish an interagency AI vulnerability coordination platform.
After more than a month of preparation, the 'Golden Eagle Initiative' was launched. Spearheaded by the White House, it brings together the Department of the Treasury, the Department of Homeland Security, the Department of Defense, and key private-sector critical infrastructure giants. Its core objective is to leverage cutting-edge artificial intelligence technologies,...
This is not a traditional system upgrade or firewall, but rather a new AI-driven, nationally led cybersecurity information exchange center.
Crisis Looms: Why Was the 'Golden Eagle Initiative' Launched?
To understand the significance of the 'Golden Eagle Initiative,' one must first confront a harsh reality:AI development has already upended the traditional balance of cyber offense and defense.
In recent years, AI-enhanced cyberattacks have significantly lowered the barrier to entry. Cutting-edge large language models can not only assist in generating zero-day exploit tools but also bypass multi-factor authentication. Moreover, code generated by widely used AI-assisted programming tools (such as GitHub Copilot) often contains security vulnerabilities—effectively mass-producing security weaknesses while accelerating code output.
To address this AI-driven, imminent national information security threat, the Trump administration had already signed Executive Order 14409 on June 2, 2026, titled 'Advancing Innovation and Safety in Advanced Artificial Intelligence.'The executive order explicitly directed the federal government to establish an interagency AI vulnerability coordination platform.
After more than a month of preparation, the 'Golden Eagle Initiative' was launched. Spearheaded by the White House, it brings together the Department of the Treasury, the Department of Homeland Security, the Department of Defense, and key private-sector critical infrastructure giants. Its core objective is to leverage cutting-edge artificial intelligence technologies,...
+3
27
2
118
Over the past two weeks, AI semiconductors have undergone a rapid and sharp de-leveraging. From the US market's $NVIDIA (NVDA.US)$ 、 $Broadcom (AVGO.US)$ 、 $Micron Technology (MU.US)$ , to Korea's $SK hynix (SKHY.US)$ 、 $Samsung Electronics (005930.KR)$ , the once most crowded AI hardware trade has seen a noticeable cooldown. As of this writing, the KOSPI has retraced more than 24%, and SOXX has declined over 13%.
This correction wasn’t triggered by a single negative catalyst but rather by the simultaneous release of multiple pressures: on one hand, the market has started reassessing the sustainability of AI-related capital expenditures; on the other, AI hardware trades had become excessively crowded, with valuations and institutional positioning both at elevated levels.
Meanwhile, earnings expectations for some semiconductor companies are beginning to diverge, prompting market concerns: has the AI hardware cycle already transitioned from a phase of high-speed growth into one of slowing profit growth?The current biggest point of contention centers on SK Hynix, a core player in AI memory.
A bullish-bearish divergence is forming around the AI memory cycle.
On July 13, leading Korean institutions successively downgraded SK Hynix’s earnings forecasts. KIS was the first to lower its 2026 and 2027 earnings estimates, and today Mirae Asset Securities also cut its Q2 2026 operating profit forecast by 12% and downgraded its DRAM...
This correction wasn’t triggered by a single negative catalyst but rather by the simultaneous release of multiple pressures: on one hand, the market has started reassessing the sustainability of AI-related capital expenditures; on the other, AI hardware trades had become excessively crowded, with valuations and institutional positioning both at elevated levels.
Meanwhile, earnings expectations for some semiconductor companies are beginning to diverge, prompting market concerns: has the AI hardware cycle already transitioned from a phase of high-speed growth into one of slowing profit growth?The current biggest point of contention centers on SK Hynix, a core player in AI memory.
A bullish-bearish divergence is forming around the AI memory cycle.
On July 13, leading Korean institutions successively downgraded SK Hynix’s earnings forecasts. KIS was the first to lower its 2026 and 2027 earnings estimates, and today Mirae Asset Securities also cut its Q2 2026 operating profit forecast by 12% and downgraded its DRAM...
124
30
256
Columns 投資者進化論｜從人手操作到AI助手，你的投資方式即將被重新定義
大家好，我是牛牛課堂 👋
今天想和大家聊一個正在發生的變化：AI，正在悄悄改變我們投資的方式。
過去十幾年，投資更多靠自己：
▎第一階段：靠自己摸索（2010-2020）
![]()
看新聞、翻財報、研究行業
![]()
自己畫K線、設定提醒
![]()
遇到市場變化，只能靠經驗和判斷
⚠️ 但問題也很明顯：
資訊太多，看不完
機會容易錯過
情緒容易影響決策
▎第二階段：工具幫你提升效率（2020-2025）
這幾年，投資工具越來越豐富：
智能選股、行情篩選
條件單、價格提醒
AI幫忙摘要新聞、解讀資訊
投資效率確實提高了不少。
⚠️ 但很多時候：
工具越來越多，卻需要自己一個個研究和組合
數據越來越豐富，但轉化成投資判斷仍需要時間
▎第三階段：AI真正融入投資（2025以後）
未來的投資助手，可能不只是回答你的問題。
它會更了解你的投資習慣：
你的關注方向、持倉情況、風險偏好。
當市場出現變化時，它可以主動幫你整理資訊、發現機會、提醒風險。
過去需要花30分鐘完成的信息整理，未來可能只需要一句話。
AI不是替你做投資決定，而是幫助每一位投資者，用更高效...
今天想和大家聊一個正在發生的變化：AI，正在悄悄改變我們投資的方式。
過去十幾年，投資更多靠自己：
▎第一階段：靠自己摸索（2010-2020）
⚠️ 但問題也很明顯：
資訊太多，看不完
機會容易錯過
情緒容易影響決策
▎第二階段：工具幫你提升效率（2020-2025）
這幾年，投資工具越來越豐富：
智能選股、行情篩選
條件單、價格提醒
AI幫忙摘要新聞、解讀資訊
投資效率確實提高了不少。
⚠️ 但很多時候：
工具越來越多，卻需要自己一個個研究和組合
數據越來越豐富，但轉化成投資判斷仍需要時間
▎第三階段：AI真正融入投資（2025以後）
未來的投資助手，可能不只是回答你的問題。
它會更了解你的投資習慣：
你的關注方向、持倉情況、風險偏好。
當市場出現變化時，它可以主動幫你整理資訊、發現機會、提醒風險。
過去需要花30分鐘完成的信息整理，未來可能只需要一句話。
AI不是替你做投資決定，而是幫助每一位投資者，用更高效...
富途AI產品升級發佈會｜下一個升浪 由你掌握
Jul 15 15:00
14
1
4
如果說過去十年半導體行業的鐵律是「摩爾定律」，那麼當下產業鏈最核心的關鍵詞無疑是「先進封裝」。
隨著AI大模型參數從百億級狂飆至萬億級別，單純依賴工藝微縮來提升算力的路徑已逼近物理極限。要在單顆AI晶片中同時容納海量計算單元與高帶寬記憶體（HBM），傳統2D封裝已顯得力不從心。在此背景下，HBM+CoWoS的黃金組合，幾乎已成為全球頭部AI晶片廠商無可替代的必選項。
從英偉達的Blackwell架構GPU，到AMD的MI系列加速器，再到雲端巨頭自研的ASIC訓練晶片——當前的產業共識是：誰掌握了CoWoS產能，誰才能在AI算力軍備競賽中真正確立優勢。
然而，在這場如火如荼的AI基礎設施擴張中，一個過往極易被市場忽視的環節，正悄然成為限制產能放量的新瓶頸：封裝設備。
作為全球封裝設備的領軍企業， $ASMPT (00522.HK)$ 正精準卡位這一產業升級的核心節點。這家公司正在從一家傳統封裝設備商，蛻變為AI先進製造鏈不可或缺的基礎設施供應商，年內股價已飆升逾140%。
摩根士丹利最新研報指出，隨著OSAT...
隨著AI大模型參數從百億級狂飆至萬億級別，單純依賴工藝微縮來提升算力的路徑已逼近物理極限。要在單顆AI晶片中同時容納海量計算單元與高帶寬記憶體（HBM），傳統2D封裝已顯得力不從心。在此背景下，HBM+CoWoS的黃金組合，幾乎已成為全球頭部AI晶片廠商無可替代的必選項。
從英偉達的Blackwell架構GPU，到AMD的MI系列加速器，再到雲端巨頭自研的ASIC訓練晶片——當前的產業共識是：誰掌握了CoWoS產能，誰才能在AI算力軍備競賽中真正確立優勢。
然而，在這場如火如荼的AI基礎設施擴張中，一個過往極易被市場忽視的環節，正悄然成為限制產能放量的新瓶頸：封裝設備。
作為全球封裝設備的領軍企業， $ASMPT (00522.HK)$ 正精準卡位這一產業升級的核心節點。這家公司正在從一家傳統封裝設備商，蛻變為AI先進製造鏈不可或缺的基礎設施供應商，年內股價已飆升逾140%。
摩根士丹利最新研報指出，隨著OSAT...
+2
31
4
44
7月13日， $SK Hynix (000660.KR)$ 超過12%，跌破200萬韓元，較6月高點回撤約35%。就在前一個交易日，其美國ADR才剛剛完成上市首秀， $SK hynix-WI (SKHYV.US)$ 收漲12.8%。
一邊是創紀錄的美國上市和火爆認購，一邊是韓國本土股價繼續下跌。與此同時，韓國投資證券KIS又下調了海力士2026年、2027年的盈利預測。
表面看，這似乎是典型的「利好出盡」：ADR上市完成、盈利預測下調、存儲股見頂。
但如果仔細拆解KIS的預測，會發現這次下調釋放的信號并不是「海力士基本面惡化」，而是存儲行業的盈利模式正在發生變化：
海力士犧牲了一部分存儲價格暴漲階段的利潤彈性，換取未來三至五年更高的收入確定性。
因此，現在真正需要討論的不是「海力士二季度利潤有沒有少賺」，而是市場應該如何給一種利潤峰值略低、但持續時間更長的存儲商業模式估值。
本文將為牛友們剖析KIS這份研報說了什麼？今日為何大跌？股價到底了嗎？
KIS這份研報說了什麼？
韓國投資證券KIS預計，SK海力士第二季度營收將達80.9萬...
一邊是創紀錄的美國上市和火爆認購，一邊是韓國本土股價繼續下跌。與此同時，韓國投資證券KIS又下調了海力士2026年、2027年的盈利預測。
表面看，這似乎是典型的「利好出盡」：ADR上市完成、盈利預測下調、存儲股見頂。
但如果仔細拆解KIS的預測，會發現這次下調釋放的信號并不是「海力士基本面惡化」，而是存儲行業的盈利模式正在發生變化：
海力士犧牲了一部分存儲價格暴漲階段的利潤彈性，換取未來三至五年更高的收入確定性。
因此，現在真正需要討論的不是「海力士二季度利潤有沒有少賺」，而是市場應該如何給一種利潤峰值略低、但持續時間更長的存儲商業模式估值。
本文將為牛友們剖析KIS這份研報說了什麼？今日為何大跌？股價到底了嗎？
KIS這份研報說了什麼？
韓國投資證券KIS預計，SK海力士第二季度營收將達80.9萬...
214
57
1074
記憶體半導體大廠鎧俠控股在經歷暴漲後，股價自高點一度回撤約30%，目前已浮現止跌回穩的跡象。近期不僅有逢低買盤湧入，這波漲勢也順勢帶動了鎧俠的供應商概念股。
圍繞著AI記憶體的投資熱潮是否將延續？除了鎧俠，本文也將帶您檢視支撐美國美光科技與韓國SK海力士等AI記憶體供應鏈的優質日股。
5檔鎧俠相關日股
根據彭博等數據，鎧俠的供應鏈企業包含了 $Quest (2332.JP)$ 、 $Kanto Denka Kogyo (4047.JP)$ 、 $Japan Material (6055.JP)$ 、 $Fixstars (3687.JP)$ 以及 $HIRATA (6258.JP)$ 。其中，關東電化工業今年以來的漲幅更逼近了180%。
$Kanto Denka Kogyo (4047.JP)$ ：主力產品為半導體製造所需的特殊氣體及高純度氟系產品。隨著3D NAND的堆疊層數增加，成膜與蝕刻等加工難度也隨之攀升。這類製造過程所需的高純度材料，特別容易受惠於記憶體的高層化與產能擴張。該股今年以來的漲幅高達180%，已累積相當龐大的漲幅。雖然市場對其業績成長抱有...
圍繞著AI記憶體的投資熱潮是否將延續？除了鎧俠，本文也將帶您檢視支撐美國美光科技與韓國SK海力士等AI記憶體供應鏈的優質日股。
5檔鎧俠相關日股
根據彭博等數據，鎧俠的供應鏈企業包含了 $Quest (2332.JP)$ 、 $Kanto Denka Kogyo (4047.JP)$ 、 $Japan Material (6055.JP)$ 、 $Fixstars (3687.JP)$ 以及 $HIRATA (6258.JP)$ 。其中，關東電化工業今年以來的漲幅更逼近了180%。
$Kanto Denka Kogyo (4047.JP)$ ：主力產品為半導體製造所需的特殊氣體及高純度氟系產品。隨著3D NAND的堆疊層數增加，成膜與蝕刻等加工難度也隨之攀升。這類製造過程所需的高純度材料，特別容易受惠於記憶體的高層化與產能擴張。該股今年以來的漲幅高達180%，已累積相當龐大的漲幅。雖然市場對其業績成長抱有...
11
35
随著全球AI算力軍備競賽進入白熱化階段，市場對「AI基礎設施投資見頂」的擔憂正被強勢打破。
面對外界關於算力過剩的質疑， $Meta Platforms (META.US)$ 首席執行官扎克伯格日前公開予以否認。同時，Meta拋出激進的擴張藍圖——目標至2027年將數據中心算力翻倍至14GW，約為今年部署規模的兩倍，德銀測算僅2027年資本開支或將突破2000億美元超市場此前預期，徹底穩定了市場對終端需求的信心。
在龐大算力需求的牽引下，記憶體晶片巨頭正全面吹響擴產號角。 $Micron Technology (MU.US)$ 不僅將美國本土製造的投資承諾由2000億美元大幅提升至2500億美元，更斥資30億美元強化半導體供應鏈，包含向環球晶圓注資5億美元進行戰略融資。與此同時， $SK hynix (SKHY.US)$ 也明確將於2026年下半年重啟大連廠二期擴建，新增V8 NAND產能。
值得注意的是，產能擴張的背後，是全球半導體供應鏈的加速重構。盧特尼克近期透露，正積極遊說三星與SK海力士赴美建廠，旨在進一步強化美國本土晶片供應鏈的韌性。
綜...
面對外界關於算力過剩的質疑， $Meta Platforms (META.US)$ 首席執行官扎克伯格日前公開予以否認。同時，Meta拋出激進的擴張藍圖——目標至2027年將數據中心算力翻倍至14GW，約為今年部署規模的兩倍，德銀測算僅2027年資本開支或將突破2000億美元超市場此前預期，徹底穩定了市場對終端需求的信心。
在龐大算力需求的牽引下，記憶體晶片巨頭正全面吹響擴產號角。 $Micron Technology (MU.US)$ 不僅將美國本土製造的投資承諾由2000億美元大幅提升至2500億美元，更斥資30億美元強化半導體供應鏈，包含向環球晶圓注資5億美元進行戰略融資。與此同時， $SK hynix (SKHY.US)$ 也明確將於2026年下半年重啟大連廠二期擴建，新增V8 NAND產能。
值得注意的是，產能擴張的背後，是全球半導體供應鏈的加速重構。盧特尼克近期透露，正積極遊說三星與SK海力士赴美建廠，旨在進一步強化美國本土晶片供應鏈的韌性。
綜...
+5
28
5
104
Columns SK海力士赴美上市：下一個台積電時刻？估值折價或迎來重估
$SK hynix (SKHY.US)$ 將於本周五正式登陸美國資本市場。市場的關注焦點已超越單一的上市交易，轉向其背後可能引發的估值體系重塑。
業內人士指出， $SK Hynix (000660.KR)$此次赴美上市的路徑，宛如當年台積電逐步提升海外投資者認知的翻版——透過接軌全球資本市場，公司價值有望從傳統的「韓國半導體股」，重新定價為「全球AI基礎設施核心資產」。
近年來，憑藉 HBM（高頻寬記憶體）的技術突破，SK海力士已躍升為AI浪潮下最核心的受益者，在 AI GPU 供應鏈中扮演著不可或缺的角色。然而，受限於韓國資本市場環境，包含三星電子與SK海力士在內的韓國半導體巨頭，長期以來皆面臨明顯的「韓股估值折價」困境。
韓國半導體巨頭：業績全球化，估值卻仍停留在本土市場
從估值角度觀察，韓國記憶體企業與海外同業存在著顯著差距。儘管同樣受惠於AI記憶體需求的爆發，SK海力士目前在韓國市場的交易估值，仍大幅低於美股同業。
可以看到，作為AI存儲核心受益者， $SK Hynix (000660.KR)$目前Forward P/E僅約5.5倍，仍低於同樣受益...
業內人士指出， $SK Hynix (000660.KR)$此次赴美上市的路徑，宛如當年台積電逐步提升海外投資者認知的翻版——透過接軌全球資本市場，公司價值有望從傳統的「韓國半導體股」，重新定價為「全球AI基礎設施核心資產」。
近年來，憑藉 HBM（高頻寬記憶體）的技術突破，SK海力士已躍升為AI浪潮下最核心的受益者，在 AI GPU 供應鏈中扮演著不可或缺的角色。然而，受限於韓國資本市場環境，包含三星電子與SK海力士在內的韓國半導體巨頭，長期以來皆面臨明顯的「韓股估值折價」困境。
韓國半導體巨頭：業績全球化，估值卻仍停留在本土市場
從估值角度觀察，韓國記憶體企業與海外同業存在著顯著差距。儘管同樣受惠於AI記憶體需求的爆發，SK海力士目前在韓國市場的交易估值，仍大幅低於美股同業。
可以看到，作為AI存儲核心受益者， $SK Hynix (000660.KR)$目前Forward P/E僅約5.5倍，仍低於同樣受益...
41
8
138
Columns 恒生科技指數近期強勢反彈：是曇花一現還是反轉行情起點？
牛友們，如果你們最近打開港股賬戶，大概率會有一種久違的恍惚感，那個連跌了近九個月、讓人幾乎想關掉行情軟件的恒生科技指數，突然又活躍了過來。
自2025年10月起， $Hang Seng TECH Index (800700.HK)$ 整體運行偏弱，走勢持續下行，6月26日該指數盤中下探至4229.94點，創下年內階段低位，市場整體投資信心趨於低迷。在多數投資者維持謹慎觀望的背景之下，市場迎來修復行情；截至7月9日收盤，恒生科技指數收於4731.56點，相對前期年內低點反彈幅度近500點。
但興奮之餘，投資者可能更想知道：恒生科技指數的這輪反彈，究竟是熊轉牛的起點，還是熊市裡的一次技術性反抽？
跌出空間？
東吳證券發布的研報顯示，截至6月末，恒生科技指數滾動市盈率（PE-TTM）為21.90倍，處於指數成立以來23.60%的歷史分位數。換句話說，當前估值比歷史上約四分之三的時間都要便宜。
港股互聯網龍頭此前遭遇的，正是典型的「預期過度下修」，外賣補貼戰、AI投入侵蝕現金流、盈利預測接連下調，幾乎把所有能想到的利空都計價完畢。
當悲觀情緒達到極值，反彈的種子便已埋下...
自2025年10月起， $Hang Seng TECH Index (800700.HK)$ 整體運行偏弱，走勢持續下行，6月26日該指數盤中下探至4229.94點，創下年內階段低位，市場整體投資信心趨於低迷。在多數投資者維持謹慎觀望的背景之下，市場迎來修復行情；截至7月9日收盤，恒生科技指數收於4731.56點，相對前期年內低點反彈幅度近500點。
但興奮之餘，投資者可能更想知道：恒生科技指數的這輪反彈，究竟是熊轉牛的起點，還是熊市裡的一次技術性反抽？
跌出空間？
東吳證券發布的研報顯示，截至6月末，恒生科技指數滾動市盈率（PE-TTM）為21.90倍，處於指數成立以來23.60%的歷史分位數。換句話說，當前估值比歷史上約四分之三的時間都要便宜。
港股互聯網龍頭此前遭遇的，正是典型的「預期過度下修」，外賣補貼戰、AI投入侵蝕現金流、盈利預測接連下調，幾乎把所有能想到的利空都計價完畢。
當悲觀情緒達到極值，反彈的種子便已埋下...
+2
Hong Kong Market Compass | Capital rotation in play? Tech stocks rebound as the Hang Seng Tech Index
26
5
24
![[empty]](https://static.futunn.com/node_futunn_nnq/assets/images/folder.5c37692712.png)
![[error]](https://static.futunn.com/node_futunn_nnq/assets/images/no-network.991ae8055c.png)
![[loading]](data:image/svg+xml;base64,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)