External broker reports(aggregated public sources), CMB International

CMB International

External broker reports(aggregated public sources)

Incorporated in Hong Kong, supported by the mainland and targeting in global market, CMB International Capital Corporation Limited (“CMBI”) is an integrated financial institution providing comprehensive and professional services. As a wholly-owned subsidiary of China Merchants Bank ("CMB"), by leveraging on the Group’s synergy and coordination in domestic and overseas markets, as well as CMB’s strong resources and outstanding social reputation, CMBI and its subsidiaries have been actively implementing diversified business strategies and set up an overall layout of main business segments including Corporate Finance, Asset Management, Wealth Management, Global Markets and Structured Finance, etc. Adhering to the philosophy of “being clients oriented and increasing our own values by the way of creating value for clients”, CMBI endeavors to provide quality and integrated financial services to institutional, corporate and individual customers both domestic and overseas including Hong Kong IPO sponsorship and underwriting, share placement and rights issue for listed companies, financial advisors, bonds issuance services, asset management, direct investment, wealth management, equity sales and trading, etc.

Thematic (Sector/Industry)

<ul><li>On 10 October 2024, AMD (AMD US, NR) launched its latest AI accelerator, the Instinct MI325X, aimed at competing directly with H200 from Nvidia (NVDA US, NR) that will ship by 4Q24.</li><li>AMD also introduced MI350 that targets Nvidia’s Blackwell series during the event, but MI350 will not enter mass production until 2H25. Lisa Su, the CEO of AMD, projected the data center AI accelerator TAM to grow at 60%+ CAGR, reaching US$500bn by 2028, a significant increase from prior estimates (US$400bn by 2027E in early-2024).</li><li>The event did not offer too much surprise to investors. Some details we noticed include: 1) Meta Llama 405B model now runs exclusively on MI300X for all live traffic. This implies AMD gains better traction at Meta. </li></ul>Content is external broker report sourced from online content aggregator through publicly available sources and is displayed below for general informational purposes only. Refer full disclaimer below.

On 10 October 2024, AMD (AMD US, NR) launched its latest AI accelerator, the Instinct MI325X, aimed at competing directly with H200 from Nvidia (NVDA US, NR) that will ship by 4Q24.AMD also introduced MI350 that targets Nvidia’s Blackwell series during the event, but MI350 will not enter mass production until 2H25. Lisa Su, the CEO of AMD, projected the data center AI accelerator TAM to grow at 60%+ CAGR, reaching US$500bn by 2028, a significant increase from prior estimates (US$400bn by 2027E in early-2024).The event did not offer too much surprise to investors. Some details we noticed include: 1) Meta Llama 405B model now runs exclusively on MI300X for all live traffic. This implies AMD gains better traction at Meta. Content is external broker report sourced from online content aggregator through publicly available sources and is displayed below for general informational purposes only. Refer full disclaimer below.

Semiconductors - AMD “Advancing AI” event

Semiconductor & Technology Specialist, Ingenuity

William Keating

Semiconductor & Technology Specialist

Ingenuity

Our focus is exclusively on in-depth research & consulting in the broader semiconductor industry. We focus on the key players, namely Intel, Micron, Qualcomm, Apple, Samsung, NVIDIA, AMD, TSMC & Global Foundries - as well as some of the lesser known, but nonetheless critical, companies such as ASML (for EUV), Mellanox, Cavium, Xilinx etc. We specialise in their products, roadmaps, process technologies, competitive advantages & threats etc. We also help our clients understand the implications of new and emerging technologies, e.g. 3D X-Point, 3D NAND, Carbon Nanotubes, Autonomous Driving, Graphene, Silicon Photonics, Wafer-Level Optics, Artificial Intelligence, FPGA, GPU, ASIC etc. 

Semiconductor Analysis & Research

Tesla first began to equip its Model S with self-driving technology in October 2014 and announced the launch of its "Autopilot" software to take advantage of those sensors exactly one year later. In separating the sensors from the software, and making that software incrementally available via over-the-air (OTA) updates, the company heralded an entirely new approach to autonomous driving that has since come to be both derided and admired in equal measure. Tesla's autopilot capability was widely regarded as ground-breaking when it was first introduced and many would argue that it was the catalyst for its peers and competitors to rapidly accelerate their autonomous driving investments and roadmaps.Since that time, Tesla's self-driving roadmap has seen many twists and turns. Their partnership with MobilEye, upon whose chips their initial autopilot roll-out critically depended, ended abruptly and acrimoniously in the wake of a fatal crash involving one of their Model S vehicles in May 2016. Their backup plan, the development of a custom ASIC chip to replicate MobilEye's functionality, was still in its infancy, forcing Tesla to turn to NVIDIA for its DRIVE PX platform. That same fatal crash was the impetus for a total redesign of their sensor suite. The new hardware platform was launched in October 2016 with the announcement that all future Tesla cars would be fully equipped with the hardware &amp; sensors required for full self-driving capability.As automobiles equipped with the new self-driving hardware rolled off the production lines, they were missing the software to make them work, meaning that they lacked basic advanced driving features such as automatic emergency braking (AEB), adaptive cruise control (ACC) etc. Tesla promised to deliver those capabilities incrementally in the following months, but inevitably, delays ensued and reliability concerns arose when they were initially released. Only in recent months has Tesla's "Enhanced Autopilot" begun to approach the same level of capability as its original launch version. Tesla CEO Elon Musk's commitment to demonstrating fully autonomous driving from US coast-to-coast by the end of 2017 failed to materialize, although he has recently stated that it should happen within 3-6 months. Meanwhile, the rapidly growing availability of new safety features such as AEB, the launch of competing driver assistance features such as General Motors (GM)"Super Cruise" and the publicly announced plans for driverless taxi pilots by both Waymo and GM have eroded Tesla's leadership. Indeed, one research firm, Navigant, doesn't even rate Tesla in its latest top ten leaders list.   In this Insight, we review the history of autonomous driving at Tesla, tracing its roots from the original autopilot launch to the present day. We examine the technology behind Tesla's self-driving platform and highlight their steadfast refusal to incorporate LIDAR in their sensor suite as every other competitor has done. Finally, in the wake of another fatal Tesla crash, we ask the very important question, how safe is Tesla Autopilot?

Tesla first began to equip its Model S with self-driving technology in October 2014 and announced the launch of its "Autopilot" software to take advantage of those sensors exactly one year later. In separating the sensors from the software, and making that software incrementally available via over-the-air (OTA) updates, the company heralded an entirely new approach to autonomous driving that has since come to be both derided and admired in equal measure. Tesla's autopilot capability was widely regarded as ground-breaking when it was first introduced and many would argue that it was the catalyst for its peers and competitors to rapidly accelerate their autonomous driving investments and roadmaps.Since that time, Tesla's self-driving roadmap has seen many twists and turns. Their partnership with MobilEye, upon whose chips their initial autopilot roll-out critically depended, ended abruptly and acrimoniously in the wake of a fatal crash involving one of their Model S vehicles in May 2016. Their backup plan, the development of a custom ASIC chip to replicate MobilEye's functionality, was still in its infancy, forcing Tesla to turn to NVIDIA for its DRIVE PX platform. That same fatal crash was the impetus for a total redesign of their sensor suite. The new hardware platform was launched in October 2016 with the announcement that all future Tesla cars would be fully equipped with the hardware &amp; sensors required for full self-driving capability.As automobiles equipped with the new self-driving hardware rolled off the production lines, they were missing the software to make them work, meaning that they lacked basic advanced driving features such as automatic emergency braking (AEB), adaptive cruise control (ACC) etc. Tesla promised to deliver those capabilities incrementally in the following months, but inevitably, delays ensued and reliability concerns arose when they were initially released. Only in recent months has Tesla's "Enhanced Autopilot" begun to approach the same level of capability as its original launch version. Tesla CEO Elon Musk's commitment to demonstrating fully autonomous driving from US coast-to-coast by the end of 2017 failed to materialize, although he has recently stated that it should happen within 3-6 months. Meanwhile, the rapidly growing availability of new safety features such as AEB, the launch of competing driver assistance features such as General Motors (GM)"Super Cruise" and the publicly announced plans for driverless taxi pilots by both Waymo and GM have eroded Tesla's leadership. Indeed, one research firm, Navigant, doesn't even rate Tesla in its latest top ten leaders list.   In this Insight, we review the history of autonomous driving at Tesla, tracing its roots from the original autopilot launch to the present day. We examine the technology behind Tesla's self-driving platform and highlight their steadfast refusal to incorporate LIDAR in their sensor suite as every other competitor has done. Finally, in the wake of another fatal Tesla crash, we ask the very important question, how safe is Tesla Autopilot?  

Managing Director, Zero One, Zero One

Vincent Fernando, CFA

Managing Director, Zero One

Zero One

Founded in 2015, Zero One produces independent investment research and provides corporate finance advisory services with a focus on both listed and private technology companies. Founder Vincent Fernando has over 20 years of experience in US and Asia capital markets.

Independent Investment Research

Equity Bottom-Up

TSMC (Taiwan Semiconductor Manufacturing) - ADR / Equity Bottom-Up

TSMC (Taiwan Semiconductor Manufacturing) - ADR

<ul>
<li>Samsung's foundry business is losing more customers to TSMC, who is perceived as lower risk.</li>
<li>TSMC, UMC Monthly Sales: TSMC up strong but plateauing, UMC lackluster, confirming 2 big trends.</li>
<li>Memory Monitor: Nanya results miss; highlights weakness in non-AI demand.</li>
</ul>


Samsung's foundry business is losing more customers to TSMC, who is perceived as lower risk.
TSMC, UMC Monthly Sales: TSMC up strong but plateauing, UMC lackluster, confirming 2 big trends.
Memory Monitor: Nanya results miss; highlights weakness in non-AI demand.


Taiwan Tech Weekly: TSMC Taking More Samsung Customers; Hon Hai Unveils Major AI Projects

Technology / Semiconductor / Internet Research

Nicolas Baratte

TSMC (Taiwan Semiconductor Manufacturing) / Equity Bottom-Up

TSMC (Taiwan Semiconductor Manufacturing)

<ul>
<li>AI demand will remain strong but is plateauing in terms of YoY growth. The rest of end-demand remains weak (Consumer, PC, Auto, Industrials). TSMC and UMC monthly sales confirm this. </li>
<li>TSMC high growth plateau ~40% YoY: AI (Nvidia Broadcom, AMD), Intel, new smartphone chips. Comps will get harder early 2025. Valuations are high but not extreme: limited upside and downside.</li>
<li>UMC lackluster at ~6% YoY growth, reflecting lackluster end-markets from smartphone, Consumer Elec, PC, Auto &amp; Industrials. Average valuations; in need of an elusive catalyst.</li>
</ul>


AI demand will remain strong but is plateauing in terms of YoY growth. The rest of end-demand remains weak (Consumer, PC, Auto, Industrials). TSMC and UMC monthly sales confirm this. 
TSMC high growth plateau ~40% YoY: AI (Nvidia Broadcom, AMD), Intel, new smartphone chips. Comps will get harder early 2025. Valuations are high but not extreme: limited upside and downside.
UMC lackluster at ~6% YoY growth, reflecting lackluster end-markets from smartphone, Consumer Elec, PC, Auto &amp; Industrials. Average valuations; in need of an elusive catalyst.


TSMC, UMC Monthly Sales: TSMC up Strong but Plateauing, UMC Lackluster, Confirming 2 Big Trends

Taiwan Tech Focused Senior Analyst

Patrick Liao

<ul>
<li>For 2nm technology, <a href="https://www.smartkarma.com/entities/taiwan-semiconductor-manufacturing-company" target="_blank" data-smartkarma-citation="entities:11230:1" rel="noopener">Taiwan Semiconductor (TSMC) (2330 TT)</a> should start <a href="https://www.smartkarma.com/entities/apple-inc" target="_blank" data-smartkarma-citation="entities:666:1" rel="noopener">Apple (AAPL US)</a> iPhone 17 production in the middle of 2025, most likely.  </li>
<li>TSMC's competitors, <a href="https://www.smartkarma.com/entities/samsung-electronics-co-ltd" target="_blank" data-smartkarma-citation="entities:191:1" rel="noopener">Samsung Electronics (005930 KS)</a> and <a href="https://www.smartkarma.com/entities/intel-corp" target="_blank" data-smartkarma-citation="entities:16764:1" rel="noopener">Intel Corp (INTC US)</a> are running foundry business and keeping owning manufacturing business at the same time, which is contradictory to purely foundry business.</li>
<li>To making the advanced chip in the US, we consider the main main concern still the cost.</li>
</ul>


For 2nm technology, Taiwan Semiconductor (TSMC) (2330 TT) should start Apple (AAPL US) iPhone 17 production in the middle of 2025, most likely.  
TSMC's competitors, Samsung Electronics (005930 KS) and Intel Corp (INTC US) are running foundry business and keeping owning manufacturing business at the same time, which is contradictory to purely foundry business.
To making the advanced chip in the US, we consider the main main concern still the cost.


TSMC (2330.TT; TSM.US): The Productivity and Profitability Are the Essential Competitiveness.

United States

NVIDIA Corporation designs, develops, and markets three dimensional (3D) graphics processors and related software.  The Company's products provide interactive 3D graphics to the mainstream personal computer market.

Price Gauge
52-week Range	0.00 - 0.00
Market Cap (USDm)	0.00
P/E	0.00
Dividend (Yield)	0.00

NVIDIA Corp (NVDA US)

Semiconductors - AMD “Advancing AI” event

On 10 October 2024, AMD (AMD US, NR) launched its latest AI accelerator, the Instinct MI325X, aimed at competing directly with H200 from Nvidia...

CMB International

Autonomous Driving- The Tesla Delusion

In the wake of Elon Musk's launch (relaunch?) of Tesla's RoboTaxi yesterday (which tanked the company's share price), you might be interested to...

William Keating

Taiwan Tech Weekly: TSMC Taking More Samsung Customers; Hon Hai Unveils Major AI Projects

Samsung's foundry business losing customers to TSMC due to perceived lower risk; TSMC and UMC sales trends confirm industry shifts, while Nanya...

Vincent Fernando, CFA

TSMC, UMC Monthly Sales: TSMC up Strong but Plateauing, UMC Lackluster, Confirming 2 Big Trends

AI demand remains strong but YoY growth is plateauing. The rest of Tech end-demand remains weak (Consumer, PC, Auto, Industrials). Sept. monthly...

Nicolas Baratte

TSMC (2330.TT; TSM.US): The Productivity and Profitability Are the Essential Competitiveness.

TSMC may start producing iPhone 17 with 2nm technology in mid-2025, while competitors Samsung and Intel face challenges with their mixed business...

Patrick Liao

Price Gauge

52-week Range

Market Cap (USDm)

P/E

Dividend (Yield)

Semiconductors - AMD “Advancing AI” event

On 10 October 2024, AMD (AMD US, NR) launched its latest AI accelerator, the Instinct MI325X, aimed at competing directly with H200 from Nvidia...

CMB International

Autonomous Driving- The Tesla Delusion

In the wake of Elon Musk's launch (relaunch?) of Tesla's RoboTaxi yesterday (which tanked the company's share price), you might be interested to...

William Keating

Taiwan Tech Weekly: TSMC Taking More Samsung Customers; Hon Hai Unveils Major AI Projects

​Samsung's foundry business losing customers to TSMC due to perceived lower risk; TSMC and UMC sales trends confirm industry shifts, while Nanya...

Vincent Fernando, CFA

TSMC, UMC Monthly Sales: TSMC up Strong but Plateauing, UMC Lackluster, Confirming 2 Big Trends

AI demand remains strong but YoY growth is plateauing. The rest of Tech end-demand remains weak (Consumer, PC, Auto, Industrials). Sept. monthly...

Nicolas Baratte

TSMC (2330.TT; TSM.US): The Productivity and Profitability Are the Essential Competitiveness.

​TSMC may start producing iPhone 17 with 2nm technology in mid-2025, while competitors Samsung and Intel face challenges with their mixed business...

Patrick Liao

Samsung's foundry business losing customers to TSMC due to perceived lower risk; TSMC and UMC sales trends confirm industry shifts, while Nanya...

TSMC may start producing iPhone 17 with 2nm technology in mid-2025, while competitors Samsung and Intel face challenges with their mixed business...