The total “AI investment” of the four US Big Tech companies is 30 trillion yen! The leading supplier is NVIDIA! Pay attention to the few companies that follow suit, and those Japanese companies too?

ビットバレー投資家 wrote a column · May 10, 2024 17:34

An important indicator for predicting the sustainability of the AI boomThere is one. It is the capital investment amount of the four US Big Tech companies that are the protagonists of AI investment. Based on recent analysts' predictions,The total capital investment amount of the Big Tech companies in 2025 is expected to rise to 200 billion dollars (about 31 trillion yen)that's it. Where will the four companies' huge capital investment funds go?

(The 4 Big Tech companies are $Microsoft (MSFT.US)$ 、 $Alphabet-A (GOOGL.US)$ 、 $Meta Platforms (META.US)$ 、 $Amazon (AMZN.US)$ )

At this financial results presentationThe four Big Tech companies explained that the reason for the increase in capital investment is to increase investment in AII did it. The king of AI semiconductorsIt's easy to guess that NVIDIA will benefit. In the list of the top supplier companies of the four Big Tech companies, NVIDIA ranked first, and it was shown that the guess was reasonable.

On the other hand,Similar to NVIDIA, companies that are all on the top supplier company list of the 4 companiesKazushaI was there。I hope these companies can also enjoy the benefits of the AI investment boom. Additionally, from the perspective of supply chain flows, suppliers of these companies, starting with NVIDIA, will also benefit from the capital investment of the four Big Tech companies. In addition to American semiconductor companies, those Japanese companies are also on the list.

The total “AI investment” of the four US Big Tech companies is 30 trillion yen! The leading supplier is NVIDIA! Pay attention to the few companies that follow s...

List of the top suppliers of the four Big Tech companies
Using the supplier list (*) of companies provided by Bloomberg, we have compiled a list of top supplier companies for each of the four Big Tech companies.

(*Summarized by Bloomberg based on materials from each company. Since the data reference dates for each company do not necessarily match, the list of top purchasing companies is reference data only. (On the other hand, the data standards are generally from the latter half of 2023 to the most recent 2024, and it can be said that they reflect relatively current conditions.)

First of all, of the top supplier companies of the Big Tech companies, all $NVIDIA (NVDA.US)$Is in the leadIt has become. Infrastructure investment for AI is being expanded so that 4 companies compete, and it can be said that it has also appeared on the supplier list.

Next,Similar to NVIDIA, what was on the list of all 4 companies, SK Hynix and $Advanced Micro Devices (AMD.US)$that's it. SK Hynix, a major Korean semiconductor company, has the largest global market share in the high-performance semiconductor memory market used for AI driving, and the company's orders for AI semiconductor memory have almost been filled until 2025. Competitors like Samsung Electronics $Micron Technology (MU.US)$The reason why AI semiconductor memory production is being strengthened in a hurry is because demand is extremely strong.

AMD, which is a competitor of NVIDIA, launched the “MI300X,” which rivals the AI semiconductor “H100” made by NVIDIA, in 2023/12. Since NVIDIA products are difficult to obtain (slightly mitigated now) and relatively expensive, the Big Tech 4 companies are also purchasing AMD products. The movement to reduce excessive dependence on NVIDIA is probably in the background.

Next, out of the 4 companies, the one that is among the top 3 supplier companies is Hon Hai Precision Industry, a major OEM company. Of a major semiconductor company $Broadcom (AVGO.US)$and electronic component manufacturer's $Celestica (CLS.US)$is on the list of 2 companies. At the financial results presentation in March, Broadcom raised the forecast for the ratio of semiconductor sales for the 2024 fiscal year to 35% from the previous 25% for AI.

List of top supplier companies of NVIDIA, AMD, and SK Hynix
Three companies, NVIDIA, AMD, and SK Hynix, are all on the list of top supplier companies of the four Big Tech companies. If the 3 companies receive orders from the 4 Big Tech companies, they will place orders with their own suppliers. Therefore, it seems that the top supplier companies of the three companies can indirectly expect the benefits of the expansion of capital investment by the four Big Tech companies.

If you check the list of top supplier companies of the 3 companies, in the case of NVIDIA and AMD, the largest semiconductor contract manufacturing companies in the world $Taiwan Semiconductor (TSM.US)$is in the lead. Since NVIDIA and AMD are fabless, this is because TSMC, which is mainly responsible for their production, is responsible. Therefore, it was decided to check TSMC's top supplier companies as well.

If you compare the above list with the list of Big Tech companies (large customers of major semiconductor companies), you can see that the supply chain in the semiconductor industry is increasing the number of companies involved as you go upstream, and many Japanese companies are also included. For example, $Tokyo Electron (8035.JP)$ Ya $Advantest (6857.JP)$ 、 $Screen Holdings (7735.JP)$ 、 $Lasertec (6920.JP)$ Semiconductor companies representing Japan and major semiconductor materials companies, etc. $Shin-Etsu Chemical (4063.JP)$that's it. The expansion of capital investment by the Big Tech companies will also benefit these Japanese companies.

Divergence rate from target stock prices of major companies that can be expected to benefit
Of the stocks listed in the above two tables, the divergence rate between target stock prices (Bloomberg aggregation) and target stock prices for the most recent month has been summarized for major companies excluding stocks listed in Korea and Taiwan. With the exception of AMD in US stocks, target stock prices have been raised in the last month. AMD's target stock price was lowered, but the divergence rate from the target stock price increased.

Performance trends of major companies that can be expected to benefit (US listed companies)

● $NVIDIA (NVDA.US)$
◆Earnings trends(Financial results announced on 2/21, next on 5/22)
・Sales for the fiscal year ending November/January (4Q) increased more than three times and reached a record high. Due to investments in AI by big tech companies, NVIDIA's performance is growing “twice as much in games.”
・Sales guidance for February-April (1Q) was 24 billion dollars ± 2%, which greatly exceeded market expectations (21.9 billion dollars) (Bloomberg tally, same as below). Demand for the company's AI semiconductors continued to be strong, and it was shown.
◆It has a market share of over 80% in the global AI semiconductor market. The company's products are dominant for generative AI training, but in the February financial results, it was found that the company's products are also used a lot for AI inference. It was suggested that there is a possibility that the company will maintain its advantage even when the introduction of AI progresses one step further.

● $Advanced Micro Devices (AMD.US)$
◆Earnings trends(Financial results announced on 4/30)
・Results for the fiscal year ending January-March (1Q) exceeded market expectations, and 2Q sales guidance was generally in line with market expectations.
・The company revised its GPU sales forecast for the data center division upward from the previous 3.5 billion dollars to 4 billion dollars. However, it fell short of the market forecast of about 6 billion dollars. NVIDIA also initially had production problems, and there was a period where supply could not keep up with demand, but it seems that AMD is also facing similar problems in the early stages of production. The upward revision of sales forecasts suggests that production issues will gradually improve.
◆Similar to NVIDIA, it handles AI semiconductors. The benefits from the Big Tech companies' investment in AI semiconductors are not as good as NVIDIA, but an increase in business results can be expected in the medium term.

● $Taiwan Semiconductor (TSM.US)$
◆Earnings trends(Financial results announced on 4/18)
・Sales for the fiscal year ending January-March (1Q) had the highest growth in over a year, and exceeded market expectations. The 2Q sales guidance also exceeded market expectations.
・The management stated that AI-related demand continues to increase rapidly. At the same time, it was emphasized that “TSMC is an important entity in realizing AI applications,” and it was revealed that “almost all AI innovations are cooperating with TSMC to tackle insatiable AI-related demand for energy-efficient computing power.”
・Sales for April, which were announced on 5/10, were up 60% compared to the same month last year, and sales were strong.
◆It is the world's largest contract semiconductor manufacturer. Since NVIDIA and AMD, which handle AI semiconductors, are fabless, semiconductor contract manufacturing companies such as TSMC, Samsung Electronics, and Global Foundry's are responsible for actual production. TSMC, which is the largest company in the industry and is leading with cutting-edge technology, will benefit the most.

● $Micron Technology (MU.US)$
◆Earnings trends(Financial results announced on March 21)
・The sales forecast for the March-May period (3Q) greatly exceeded market expectations.
・The management team commented, “I believe Micron is one of the biggest beneficiaries in the semiconductor industry of multi-year business opportunities brought about by AI.”
◆We work on semiconductor memories. In the AI semiconductor memory market, it is in 3rd place after SK Hynix, which has the largest market share in the world, and Samsung Electronics, which is in 2nd place. Mass production of the “HBM (broadband memory) 3E” used in NVIDIA's next-generation AI semiconductor “H200” began in February this year.

● $Broadcom (AVGO.US)$
◆Earnings trends(Financial results announced on March 7)
・Sales for the November-January fiscal year ended (1Q) were above market expectations. Sales in the semiconductor sector fell short of market expectations. The full-year sales forecast was on par with previous forecasts.
・The company raised the forecast for the ratio of semiconductor sales for the full year to 35% from the previous 25% for AI. The management team commented that demand for AI data centers for the company's network products is strong, and there are strong inquiries from major cloud computing providers (big tech companies) for custom AI semiconductors.
◆It does not handle AI semiconductors like NVIDIA and AMD, but it is involved in networking parts supply and custom chip design work used to learn large-scale language models for AI.

● $Applied Materials (AMAT.US)$
◆Performance trends(Financial results announced on 2/16, next 5/16)
・Sales for November-January (1Q) were higher than the market. Sales guidance for the second quarter showed a bullish outlook that exceeded market expectations.
・The management team expressed that “wide bandwidth memory (HMB, AI semiconductor memory) is expected to grow at an average annual growth rate of 50% over the next few years, and it will be a major business opportunity for our company.” The market share expansion in the DRAM market is expected to continue in the future, and it was indicated that in 2024, revenue from HBM packaging is expected to quadruple the previous fiscal year.
◆A major semiconductor manufacturing equipment company. Samsung Electronics, TSMC, Micron, etc. are the main customers, and they will be able to indirectly enjoy the benefits of expanding investment for AI. As indicated by the management team, strong demand for AI semiconductor memory will lead to future business growth.

Created on 2024/5/10 Market Analyst Amelia
Source: Company materials and created by MooMoo Securities from Bloomberg