Asia and Global Semiconductor and Tech Research

Andrew Lu

Thematic (Sector/Industry)

<ul>
<li>Taiwan Dell said "AI GPU shortage issue has been resolved. The lead time has been cut from 40 weeks by the end of 2023 to 8-12 weeks or less now.</li>
<li>We see this news short term positive but medium term negative to AI GPU supply chain vendors but need more information to confirm if this is an inflection point.</li>
<li>1. Taiwan Dell got priority supply? 2. Earlier overbooking over? 3. Lead time cut by supply upside? 4. Cooling company said the same; 5. Will lead time continue to shrink?</li>
</ul>


Taiwan Dell said "AI GPU shortage issue has been resolved. The lead time has been cut from 40 weeks by the end of 2023 to 8-12 weeks or less now.
We see this news short term positive but medium term negative to AI GPU supply chain vendors but need more information to confirm if this is an inflection point.
1. Taiwan Dell got priority supply? 2. Earlier overbooking over? 3. Lead time cut by supply upside? 4. Cooling company said the same; 5. Will lead time continue to shrink?


Early Sign of AI GPU Market Inflection Point?

Equity Long-Short, Semiconductors, Fabricated Knowledge

Douglas O'Laughlin

Equity Long-Short, Semiconductors

Fabricated Knowledge

Learn more about the world's most important manufactured product. Meaningful insight, timely analysis, and an occasional investment idea.

Equity Bottom-Up

<ul>
<li>Nvidia CEO Jensen Huang has repeatedly emphasized that the data center is the new unit of compute.</li>
<li>While this concept seemed straightforward at first, the deeper implications became clearer after Nvidia's presentations at GTC and OFC 2024.</li>
<li>I only recently grasped exactly what is happening, and a simple reframing of the underlying principles that drove Moore’s Law makes the entire picture clearer.</li>
</ul>


Nvidia CEO Jensen Huang has repeatedly emphasized that the data center is the new unit of compute.
While this concept seemed straightforward at first, the deeper implications became clearer after Nvidia's presentations at GTC and OFC 2024.
I only recently grasped exactly what is happening, and a simple reframing of the underlying principles that drove Moore’s Law makes the entire picture clearer.


The Data Center is the New Compute Unit: Nvidia's Vision for System-Level Scaling

<ul>
<li>Earlier expectation on Micron will have rooms to raise becoming a reality, largely driven by over 30% q/q NAND bit price jump for storage business in 1Q24.</li>
<li>10ppts stronger 2Q24 sales guidance and 3-6ppts higher gross margin guidance are driving free cash flow improvement.</li>
<li>AI server, PC, and smartphone driving DRAM bit demand and HBM3e/HBM4 and 1-gamma EUV DRAM slowing bit supply growth both are positive except MOI remains 15% above 5 years average. </li>
</ul>


Earlier expectation on Micron will have rooms to raise becoming a reality, largely driven by over 30% q/q NAND bit price jump for storage business in 1Q24.
10ppts stronger 2Q24 sales guidance and 3-6ppts higher gross margin guidance are driving free cash flow improvement.
AI server, PC, and smartphone driving DRAM bit demand and HBM3e/HBM4 and 1-gamma EUV DRAM slowing bit supply growth both are positive except MOI remains 15% above 5 years average. 


The First Semi-Micron to Guide 2Q24; All Good News Except MOI

<ul>
<li>"It is so much cheaper to do simulations than real experiments, so much more flexible in testing...we can even do things which cannot be done in any lab… So it’s inevitable this trend will continue for some time." - Richard Hamming, <a href="https://www.amazon.com/Art-Doing-Science-Engineering-Learning/dp/1732265178" target="_blank" rel="noopener">The Art of Doing Science and Engineering</a>.</li>
<li>Richard Hamming said this to the President of Bell Labs in the 1950s. And I think this insight rings more accurate than ever. Simulations of reality are much more flexible and cheaper than observing and interacting with reality.</li>
<li>I will assert that LLMs and “AI” are effectively the logical conclusion of this trend. A large language model is trained on the largest amount of data possible, often the entire crawlable Internet<a href="https://kili-technology.com/large-language-models-llms/9-open-sourced-datasets-for-training-large-language-models" target="_blank" rel="noopener">, all the books ever written, and so much more</a>.</li>
</ul>


"It is so much cheaper to do simulations than real experiments, so much more flexible in testing...we can even do things which cannot be done in any lab… So it’s inevitable this trend will continue for some time." - Richard Hamming, The Art of Doing Science and Engineering.
Richard Hamming said this to the President of Bell Labs in the 1950s. And I think this insight rings more accurate than ever. Simulations of reality are much more flexible and cheaper than observing and interacting with reality.
I will assert that LLMs and “AI” are effectively the logical conclusion of this trend. A large language model is trained on the largest amount of data possible, often the entire crawlable Internet, all the books ever written, and so much more.


<ul>
<li>Post 35% price gain in a quarter, many good news we expected and other analysts expected already factored into the share price. We believe the risk/reward not attractive here.</li>
<li>TSMC guides stronger 1Q24 and 2024, helped by Nvidia and copy cat AI GPU/ASIC orders. We see greater opportunities for alternative copy cat solutions like AMD and Alchip.</li>
<li>We attribute Nvidia customers' sales discrepancy to AI server is cannibalizing non-AI server and Nvidia is allocating more GPUs to premium price paying customers in 1H24 and reverse in 2H24.</li>
</ul>


Post 35% price gain in a quarter, many good news we expected and other analysts expected already factored into the share price. We believe the risk/reward not attractive here.
TSMC guides stronger 1Q24 and 2024, helped by Nvidia and copy cat AI GPU/ASIC orders. We see greater opportunities for alternative copy cat solutions like AMD and Alchip.
We attribute Nvidia customers' sales discrepancy to AI server is cannibalizing non-AI server and Nvidia is allocating more GPUs to premium price paying customers in 1H24 and reverse in 2H24.


From Nvidia Results to See Supply Chain Stories?

United States

Advanced Micro Devices, Inc. (AMD) produces semiconductor products and devices. The Company offers products such as microprocessors, embedded microprocessors, chipsets, graphics, video and multimedia products and supplies it to third-party foundries, as well as provides assembling, testing, and packaging services. AMD serves customers worldwide.

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

Advanced Micro Devices (AMD US)

Early Sign of AI GPU Market Inflection Point?

Early Sign of AI GPU Market Inflection Point? We see this lead time cut news short term positive but long term negative to AI GPU supply chain...

Andrew Lu

The Data Center is the New Compute Unit: Nvidia's Vision for System-Level Scaling

Copper, Cooling, and Compute Density: The Pillars of Nvidia's Data Center Vision. Nvidia CEO Jensen Huang has repeatedly emphasized that the data...

Fabricated Knowledge

The First Semi-Micron to Guide 2Q24; All Good News Except MOI

The First Semi-Micron to Guide 2Q24; All Good News Except MOI; AI server, PC, and smartphone driving DRAM bit demand and HBM3e/HBM4 and 1-gamma EUV...

Andrew Lu

Jensen's World: Compressing Reality

Nvidia is pursuing a strategy of broadening its software distribution on one of the largest installed bases in the world.

Fabricated Knowledge

From Nvidia Results to See Supply Chain Stories?

From Nvidia Results to See Supply Chain Stories? Post 35% price gain in a quarter, many good news already factored into the share price. We...

Andrew Lu

Price Gauge

52-week Range

Market Cap (USDm)

P/E

Dividend (Yield)