Stock LLM Wiki

報告_MS_Computex2026重點整理_Agentic_AI_TSMC產能_20260602

更新 2026-06-03

PDF 原檔:260602_ms_key-computex-takeaways_original.pdf

PDF 原檔:260602_ms_key-computex-takeaways_original.pdf

原始內容

M June 2, 2026 04:22 PM GMT

Greater China Technology Semiconductors | Asia Pacific

Key Computex Takeaways: Agenetic AI, TSMC capacity and MediaTek's AI PC chips

Agentic AI and CPU servers dominated the Computex keynote discussions. Nvidia indicated that TSMC's capacity can support robust growth for Nvidia in 2027.

Agentic AI compute key focus: In their keynotes, Nvidia, Arm and Qualcomm emphasized Arm-based server CPUs to support future agentic AI demand, echoing our report, Global Tech: Agentic AI - The Surge Begins (11 May 2026).

Comments on Nvidia Vera CPU: Ahead of Computex, we published NVIDIA's Vera CPU and Rubin GPU seen as the main show (27 May 2026). At today's investor luncheon, our Nvidia analyst Joe Moore asked CEO Jensen Huang the first question about the US$20bn Vera CPU revenue assumption. Management indicated that demand spans both AI server head nodes and stand-alone CPU servers. Given its materially better performance than the x86 (see Exhibit 5 ), Nvidia can command higher values for Vera, excl. memory sales. Our supply chain checks suggest 2.5mn4mn units are in preparation. Rubin has already entered fab production with no apparent delays. On CPO (optics) vs. copper, Nvidia indicated that it will adopt CPO only when necessary and rely on copper as long as possible.

Sufficient TSMC capacity: At the 2024 and 2025 Computex events, we asked Jensen Huang about TSMC's pricing and value. This year, we asked whether Nvidia can secure sufficient TSMC capacity for 2027 and whether allocation could affect market share. We also asked - even more importantly - how Nvidia could secure incremental allocation versus competitors.

Nvidia indicated it has sufficient TSMC capacity to support robust growth in 2027. The company has convinced TSMC and other suppliers that demand remains strong. Even as supply expands, demand will likely still exceed supply in 2027. Management also emphasized that the AI semiconductor market is growing and does not represent a zero-sum or pure market share competition.

Agenetic AI at the edge - do we need a local LLM/agent: Nvidia announced its RTX Sparks AI PC WoA SoC in its keynote, jointly designed with MediaTek. We flagged this 9 months ago in our MediaTek report, saying ' Nvidia/MediaTek WoA AI PC keeps being pushed out - now to 2H26. For N1X, we think the new timing of product launch would be next year's Computex in June (vs. the CES in January), given that Microsoft's new AI Windows system will be more of a 2H26 rollout. Moreover, the new x86 SoC will fuse an Intel CPU + Nvidia RTX GPU chiplet via NVLink ". We do not see this launch as a surprise. Investors are asking about the volume of RTX Sparks in 2026 and whether edge AI compute is necessary for Agentic AI.

Idea

Morgan Stanley Taiwan Limited+ Charlie Chan Equity Analyst Charlie.Chan@morganstanley.com +886 2 2730-1725
Daniel Yen, CFA Equity Analyst Daniel.Yen@morganstanley.com Morgan Stanley Asia Limited+ +886 2 2730-2863
Daisy Dai, CFA Equity Analyst Daisy.Dai@morganstanley.com +852 2848-7310
Tiffany Yeh Equity Analyst Tiffany.Yeh@morganstanley.com +886 2 7712-3032
Lucas Wang Research Associate Lucas.Wang@morganstanley.com Morgan Stanley Asia Limited+ +886 2 2730-2875
260602_ms_key-computex-takeaways_001

Morgan Stanley does and seeks to do business with companies covered in Morgan Stanley Research. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of Morgan Stanley Research. Investors should consider Morgan Stanley Research as only a single factor in making their investment decision.

For analyst certification and other important disclosures, refer to the Disclosure Section, located at the end of this report.

+= Analysts employed by non-U.S. affiliates are not registered with FINRA, may not be associated persons of the member and may not be subject to FINRA restrictions on communications with a subject company, public appearances and trading securities held by a research analyst account.

aaiole wewtirealmellel tyniil

CAIlull <.

for N1X

(9208

M

Custom Bulft with MediaTok

Key takeaways from Computex Day 1 + implications for Greater China semis

Potential volume of Nvidia RTX Spark:

Our checks with PC brands at this year's Computex suggest AI PCs with N1X will need to price at US$2,899, while N1 models will be priced at US$1,799. We therefore estimate ~5mn-8mn units of N1X/N1 SoC shipments in 2026, contributing 5-10% of MediaTek's 2026 EPS, assuming an average US$40 royalty fee per chip.

Exhibit 1: New AI PC announced by Nvidia (N1X)

260602_ms_key-computex-takeaways_002

Source: Nvidia

Exhibit 3: RTX Spark (N1X) provided by MSI (available in 3Q26)

260602_ms_key-computex-takeaways_003

Source: MSI

Exhibit 2: MediaTek provides 20-core customized Grace CPU for N1X

260602_ms_key-computex-takeaways_004

Source: Nvidia

Exhibit 4: RTX Spark (N1X) provided by ASUS (available in 3Q26)

260602_ms_key-computex-takeaways_005

Source: ASUS

Mieula len vruviueo cu uule cuotultlzeu Ulaue uru

20-Core Grace CPU

Custom Built With MediaTek

Compilation

X86 CPU

• NVIDIA Vera

1.9X

M

Greater China semis in the supply chain:

88 NVIDIA Custom Olympus Core

With Spatial Multithreading

PCle Gen 6 CXL 3.1

164 MB L3 Cache

3.4 TB/s Core-to-Core

Bisection Bandwidth

Up to 1.5 TB LPDDR5X Memory

  • Nvidia AI GPU supply chain: TSMC (foundry), ASE (packaging), KYEC (testing), AllRing (equipment)
  • CPO supply chain: FOCI (FAU), Himax (WLO), ASE (optical insertion testing), Hon Precision (optical insertion tool)
  • CPU supply chain: KYEC (Vera and Google CPU test), Aspeed (general server BMC)
  • WoA AI PC: MediaTek (N1X and N1 chips), Macronix (NOR Flash for AI PC)

Exhibit 5: Vera performance could be 1.8x that of highestperformance x86 CPU

Exhibit 6: Vera CPU cores are not split across chiplets, enabling faster core-to-core connection

260602_ms_key-computex-takeaways_006

Source: Nvidia

260602_ms_key-computex-takeaways_007

Source: Nvidia

Python

M

Disclosure Section

The information and opinions in Morgan Stanley Research were prepared or are disseminated by Morgan Stanley Asia Limited (which accepts the responsibility for its contents) and/or Morgan Stanley Asia (Singapore) Pte. (Registration number 199206298Z) and/or Morgan Stanley Asia (Singapore) Securities Pte Ltd (Registration number 200008434H), regulated by the Monetary Authority of Singapore (which accepts legal responsibility for its contents and should be contacted with respect to any matters arising from, or in connection with, Morgan Stanley Research), and/or Morgan Stanley Taiwan Limited and/or Morgan Stanley & Co International plc, Seoul Branch, and/or Morgan Stanley Australia Limited (A.B.N. 67 003 734 576, holder of Australian financial services license No. 233742, which accepts responsibility for its contents), and/or Morgan Stanley Wealth Management Australia Pty Ltd (A.B.N. 19 009 145 555, holder of Australian financial services license No. 240813, which accepts responsibility for its contents), and/or Morgan Stanley India Company Private Limited having Corporate Identification No (CIN) U22990MH1998PTC115305, regulated by the Securities and Exchange Board of India ('SEBI') and holder of licenses as a Research Analyst (SEBI Registration No. INH000001105); Stock Broker (SEBI Stock Broker Registration No. INZ000244438), Merchant Banker (SEBI Registration No. INM000011203), and depository participant with National Securities Depository Limited (SEBI Registration No. IN-DP-NSDL-567-2021) having registered office at Altimus, Level 39 & 40, Pandurang Budhkar Marg, Worli, Mumbai 400018, India; Telephone no. +91-22-61181000; Compliance Officer Details: Mr. Tejarshi Hardas, Tel. No.: +91-22-61181000 or Email: tejarshi.hardas@morganstanley.com; Grievance officer details: Mr. Tejarshi Hardas, Tel. No.: +91-22-61181000 or Email: msic-compliance@morganstanley.com which accepts the responsibility for its contents and should be contacted with respect to any matters arising from, or in connection with, Morgan Stanley Research, and their affiliates (collectively, "Morgan Stanley"). Morgan Stanley India Company Private Limited (MSICPL) may use AI tools in providing research services. All recommendations contained herein are made by the duly qualified research analysts.

For important disclosures, stock price charts and equity rating histories regarding companies that are the subject of this report, please see the Morgan Stanley Research Disclosure Website at www.morganstanley.com/eqr/disclosures/webapp/generalresearch, or contact your investment representative or Morgan Stanley Research at 1585 Broadway, (Attention: Research Management), New York, NY, 10036 USA.

For valuation methodology and risks associated with any recommendation, rating or price target referenced in this research report, please contact the Client Support Team as follows: US/Canada +1 800 303-2495; Hong Kong +852 2848-5999; Latin America +1 718 754-5444 (U.S.); London +44 (0)20-7425-8169; Singapore +65 6834-6860; Sydney +61 (0)2-9770-1505; Tokyo +81 (0)3-6836-9000. Alternatively you may contact your investment representative or Morgan Stanley Research at 1585 Broadway, (Attention: Research Management), New York, NY 10036 USA.

圖片清單(已驗證 2026-07-02)

回補驗證:僅涵蓋已被 lib 頁嵌入的圖片,非全量驗證。

檔名 size 分類 親眼所見內容
260602_ms_key-computex-takeaways_006.png 47KB 真資料圖 長條圖比較 x86 CPU 與 NVIDIA Vera 效能:Compilation 項目 1X vs 1.7X、Python 項目 1X vs 1.9X(疑似會場簡報投影翻拍照)