(#9) Computer Chips & Semiconductors: markets analysis and beyond the numbers

Hello,

Before I start, I would like to offer my appreciation for you, for reading this article. More than 3150 professionals like you, subscribed to this Linkedin Newsletter. I am grateful for your attention, comments & feedback in this project that started out of my curiosity for financial numbers, way back in 2016.

In this article, I analysed the computer chips & semiconductors and players like TSMC but also the players in GPUs manufacturing like Nvidia.

The purpose of this newsletter is to connect the story with the numbers, or sometimes the numbers with the actual story. It is, of course, my own view on the topics without any investment recommendation. I hope we can start conversations around the points below.

First, here's an interesting book: "Chip war. The fight for the world's most critical technology" by Chris Miller that you can order on amazon here https://www.amazon.com/Chip-War-Worlds-Critical-Technology/dp/1982172002

From the book presentation: "Economic historian Chris Miller explains how the semiconductor came to play a critical role in modern life and how the U.S. become dominant in chip design and manufacturing and applied this technology to military systems." Also, "You may be surprised to learn that microchips are the new oil—the scarce resource on which the modern world depends. Today, military, economic, and geopolitical power are built on a foundation of computer chips. Virtually everything—from missiles to microwaves, smartphones to the stock market—runs on chips."

Moreover, check this video below from Wall Street Journal on how chips that power AI work.

If you want to learn more about how GPUs (Graphical Processing Units) are driving the AI revolution check also this resource. As stated "Graphics Processing Units, or GPUs, have moved beyond their original role of rendering video game graphics and are now used in a variety of high-performance computing applications (HPC), from AI training to zooplankton classification.?" - https://www.visualcapitalist.com/sp/gpus-v-cpus/

NVIDIA points out here: https://blogs.nvidia.com/blog/why-gpus-are-great-for-ai/ that "GPUs have been called the rare Earth metals — even the gold — of artificial intelligence, because they’re foundational for today’s generative AI era."

As pointed out in this article https://www.visualcapitalist.com/sp/5-drivers-behind-the-growth-of-the-gpu-cloud-computing-market/ by Visual Capitalist and the graphic above, the AI generated the growth in production of GPUs, although initially designed by Nvidia for video games.

1. Semiconductors market

The semiconductors supply chain is complex and I invite you to check this interesting visual for more: https://www.visualcapitalist.com/sp/visualizing-the-global-semiconductor-supply-chain/

The Semiconductor Industry Association (SIA) points out that the "global semiconductor industry sales totaled $46.2 billion during the month of February 2024, an increase of 16.3% compared to the February 2023 total of $39.7 billion but a decrease of 3.1% from the January 2024 total of $47.6 billion."

Revenue in the Semiconductors market is projected to reach 613,10 bn $ in 2024 according to Statista while the compound annual growth rate of 6,3% will result to a market volume of 736,4 bn $ by 2027.

According to semi.org "Global shipments of silicon wafers are projected to decline 14% in 2023, to 12,512 million square inches (MSI) from the record high of 14,565 MSI in 2022 before bouncing back in 2024"

As mentioned by the source: "Silicon wafers are the fundamental building material for the majority of semiconductors, which are vital components of all electronic devices. The highly engineered thin disks are produced in diameters of up to 12 inches and serve as the substrate material on which most semiconductor devices, or chips, are fabricated." There is a continuous forecasted growth for the global shipments of silicon wafers.

The semiconductor sales worldwide almost doubled on a monthly basis since 2012.

In 2022, the artificial intelligence (AI) chip market reached 44.2 billion U.S. dollars. The estimation is for an increase to 53.5 billion dollars by the end of 2023 as the hype around AI applications continues to grow, with longer forecasts suggesting the market will approach 120 billion U.S. dollars in 2027.

In the '90, Japan, Europe and US dominated the chip production, but the forecast shows a rise for China, Taiwan and South Korea.

As represented by this Statista visual, USA lead the design of the components for semiconductors, while production is based in China & Taiwan.

The advanced semiconductor market share show Taiwan with a 68% market share in 2023 but also with a forecasted value for 2028 of 60%. Both South Korea and United States are expected to have higher market shares.

In Taiwan, TSMC (Taiwan Semiconductor Manufacturing Company) is the leader on semiconductor foundries with 16,7 billion $ in the first quarter of 2023 as represented below and around 70 bn. $ overall in 2023.

Unites States has already new projects for the next years of facilities for semiconductor manufacturing. There is a strong race towards developing the facilities as the development of AI will require higher and higher computing power.

2. TSMC analysis

According to CNBC, TSMC is the world’s largest contract semiconductor manufacturer, which makes chips for companies from Apple to Nvidia.

According to the company website "In 2022, TSMC led the foundry to start 3nm FinFET (N3) technology high volume production. TSMC’s 3nm process is the industry’s most advanced semiconductor technology offering best power, performance, and area (PPA), and is a full-node advance from its 5nm generation." Moreover, "TSMC has always insisted on building a strong, in-house R&D capability. As a global semiconductor technology leader, TSMC provides the most advanced and comprehensive portfolio of dedicated foundry process technologies."

The company states that the most important technology is "High Performance Computing (HPC) technology innovation enables broad artificial intelligence (AI) and 5G applications"

An article in Statista, points "in the fourth quarter of 2023, Taiwan Semiconductor Manufacturing Company (TSMC) recorded a market share of 61.2 percent in the global semiconductor foundry market." The revenue is close to 68 bn $.

An article in CNBC fro April 10 pointed out that "TSMC said March revenue came in at 195.2 billion new Taiwan dollars ($6.1 billion), up 34.3% year-on-year — marking the fastest pace of growth since November 2022." https://www.cnbc.com/2024/04/10/tsmc-posts-fastest-monthly-revenue-growth-since-2022-on-ai-chip-boom.html

The revenues of TSMC increase with 43% in 2022, but droped with 5% in 2023. TSMC's revenues doubled in the last 5 years. In 2023, the structure of revenues, as seen below, shows that more than half of the revenues were on technologies like 3nm process, 5nm process and 7nm process.

The HPC technology that is used for AI was 43% of the total revenues, and technology for smartphones was 38% of the total revenues. As pointed out by the company, the highest growth was for the automotive platform for their technology.

The profitability of the company shows 54% gross profit margin and 43% operating profit margin in 2023. The value remained stabled in the last 10 years.

In the same time, the Research & Development to Revenue ratio was 8% in 2023, and is constant over time, which shows the involvement of the company to continue to innovate in this sector. 72% of the operating expenses in 2023 were for research and development.

The net profit of TSMC increased in 2022 with more than 72% and decreased with 17% in 2023.

It would be interested to see the next quarterly earnings report of the company in relation to the revenue, but also the operating and net profit values.

3.NVIDIA analysis

Why is Nvidia so important for AI revolution? Check this short video for some explanations.

According tu Euronews, "Nvidia is the clear leader of the semiconductor race, controlling approximately 80% of the artificial intelligence chip sector". Also, Statista points out that: "As a pioneer and leader in AI, Nvidia chips are being used to train and run various large language models, including the one developed by OpenAI. ChatGPT was trained using thousands of Nvidia graphics processing units (GPUs), linked together in an advanced AI supercomputing data center belonging to Microsoft." Check also this article in The Economist: https://www.economist.com/the-economist-explains/2024/02/27/why-do-nvidias-chips-dominate-the-ai-market

The Data Center quarterly revenue reached 18,4 bln $ in Q4 of 2023.

This results were also seen by investors and the price of the share of Nvidia also increased over the last year with more than 220%. The other two competitors, Intel and AMD have also been in investors' focus, with AMD having a 77,7% increase in the stock price over the last year.

Because of this strong results, Nvidia entered the top 3 worldwide companies by market capitalisation. As of March 26, 2024, NVIDIA is the third global company by market capitalisation with 2,37 trillion $.

The Data center processors for analytics and AI was 78% of Nvidia's revenue, while the weight of GPUs for computers dropped to 17,1%.

NVIDIA's revenue has increased in 2023 with 126% (plus 33,9 bln $) to a record of 60,9 bln $.

The value of revenues increased with 126% between 2021 to 2023, while the operating profit increased from 10 bn $ to 33 bn $ (+228%).

The growth was mainly because of an increase in the Compute & Networking segment from 15 bln $ to 47,4 bln $. In 2022, the Compute & Networking segment had 56% of the total revenues while Graphics had 44%. In 2023, the weight in the total sales for Compute & Networking segment reached 78%.

In the same time, the Gross Profit Margin increased from 57% in 2022 to 73% in 2023 with a Gross Profit of 44,3 bln $. The operating profit margin skyrocketed from 15,7% in 2022 to 54,1% in 2023. The liquidity indicators show a very high current ratio which might signal also that the company is prepared to better use the available cash for further production development.

The company invested heavily in research and development in the last 7 year from 1,46 bn$ in 2017 to 8,6 bn $ in the last fiscal year.

The R&D to Revenue ratio decreased from 27% in 2022 to 14% in 2023. So, think about it, more than a quarter of the revenues in 2022 were directed by the company to research & development. In 2023 the R&D to Op Expenses Ratio increased from 66& in the previous year to 77%.

According to Statista, Microsoft and Meta have each purchased a high number of H100 graphics processing units (GPUs) from Nvidia. "In 2023, it was estimated that both companies had received 150,000 H100 GPUs, way ahead of the likes of Google and Amazon in terms of order numbers."

It is clear that for the moment, Nvidia dominates the AI chip market. It will be interesting how the other players will respond. Intel announced the launch of Gaudi 3 AI chip, a competitor Nvidia's H100 graphic processing unit (GPU).

As stated in this article in CNN Business: “The outlook for Nvidia is positive as AI chip competition from Intel, AMD, Meta and Microsoft could be months away while demand for Nvidia chips is only surging.” https://edition.cnn.com/2024/02/21/tech/nvidia-ai-sales-boom/index.html

Things to watch:

• how Intel or AMD will answer to Nvidia.
• evolution of demand by major companies, but also their involvement in the production. We should expect many other rising players.
• The growth of the production facilities in Western Countries.
• next quarterly results from Nvidia, Intel or AMD.

Before we go, check this video on how Nvidia powered ChatGPT

Thanks for reading and take care,

Costin Ciora, PhD.

P.S. Next week, I will present an analysis on the most popular sport on Earth. It will be about football, so don't forget to share & subscribe to this Linkedin Newsletter if you liked this edition.

References & interesting resources:

[1] https://blogs.nvidia.com/blog/why-gpus-are-great-for-ai/

[2] https://www.visualcapitalist.com/sp/gpus-v-cpus/

[3] https://www.visualcapitalist.com/nvidia-revenue-by-product-line/

[4] https://www.visualcapitalist.com/sp/5-drivers-behind-the-growth-of-the-gpu-cloud-computing-market/

[5] https://www.statista.com/statistics/1369267/intel-quarterly-revenue-by-segment/#:~:text=In%20the%20fourth%20quarter%20of,to%2014.16%20billion%20U.S.%20dollars.

[6] https://ir.amd.com/news-events/press-releases/detail/1180/amd-reports-fourth-quarter-and-full-year-2023-financial

[7] https://investor.nvidia.com/financial-info/sec-filings/default.aspx

[8] https://www.intc.com/news-events/press-releases/detail/1672/intel-reports-fourth-quarter-and-full-year-2023-financial

[9] https://www.visualcapitalist.com/top-10-semiconductor-companies-by-market-share/

[10] https://www.visualcapitalist.com/semiconductor-foundry-companies-ranked/

[11] https://www.google.com/finance/

[12] https://www.semiconductors.org/global-semiconductor-sales-increase-16-3-year-to-year-in-february/

[13] https://www.statista.com/outlook/tmo/semiconductors/worldwide

[14] https://www.statista.com/statistics/1350976/leading-tech-companies-worldwide-by-market-cap/

[15] https://www.statista.com/statistics/988048/nvidia-research-and-development-expenses/

[16] https://www.semi.org/en/products-services/market-data/materials/si-shipment-statistics

[17] https://www.statista.com/statistics/1446564/nvidia-h100-gpu-shipments-by-customer/

[18] https://www.tomshardware.com/tech-industry/nvidia-ai-and-hpc-gpu-sales-reportedly-approached-half-a-million-units-in-q3-thanks-to-meta-facebook

[19] https://www.statista.com/statistics/867223/worldwide-semiconductor-foundries-by-market-share/#:~:text=Semiconductor%20foundries%20revenue%20share%20worldwide%202019%2D2023%2C%20by%20quarter&text=In%20the%20fourth%20quarter%20of,11.3%20percent%20of%20the%20market.

[20] https://www.cnbc.com/2024/04/10/tsmc-posts-fastest-monthly-revenue-growth-since-2022-on-ai-chip-boom.html

[21] https://www.tsmc.com/english/dedicatedFoundry/technology/logic/l_3nm

[22]https://www.tsmc.com/english/dedicatedFoundry/technology/platform_HPC

[23] https://www.statista.com/statistics/1446564/nvidia-h100-gpu-shipments-by-customer/

[24] https://www.visualcapitalist.com/sp/visualizing-the-global-semiconductor-supply-chain/

[25] https://www.euronews.com/business/2024/04/10/intel-gives-nvidia-run-for-money-with-new-gaudi-chip

[26] https://www.statista.com/topics/6153/artificial-intelligence-ai-chips/#topicOverview

[27] https://www.statista.com/chart/31834/new-projects-in-semiconductor-manufacturing-in-the-united-states/

[28] https://www.statista.com/chart/27903/annual-revenue-of-the-leading-semiconductor-foundries/

[29] https://www.statista.com/chart/31371/distribution-of-global-semiconductor-fabricating-capacity/

[30] https://www.statista.com/chart/30041/global-semiconductor-wafer-fabrication-capacity-by-type-and-location/

[31] https://www.statista.com/chart/31585/semiconductor-vendors-annual-revenue/