According to the news from Commercial Times, in a recent press conference, the four major American cloud service providers (CSPs) collectively expressed their intention to expand their investment in AI application services. Simultaneously, they are continuing to enhance their cloud infrastructure. Apple has also initiated its foray into AI development, and both Intel and AMD have emphasized the robust demand for AI servers. These developments are expected to provide a significant boost to the post-market prospects of Taiwan’s AI server supply chain.

Industry insiders have highlighted the ongoing growth of the AI spillover effect, benefiting various sectors ranging from GPU modules, substrates, cooling systems, power supplies, chassis, and rails, to PCB manufacturers.

The American CSP players, including Microsoft, Google, Meta, and Amazon, which recently released their financial reports, have demonstrated growth in their cloud computing and AI-related service segments in their latest quarterly performance reports. Microsoft, Google, and Amazon are particularly competitive in the cloud services arena, and all have expressed optimistic outlooks for future operations.

The direct beneficiaries among Taiwan’s cloud data center suppliers are those in Tier 1, who are poised to reap positive effects on their average selling prices (ASP) and gross margins, driven by the strong demand for AI servers from these CSP giants in the latter half of the year.

Among them, the ODM manufacturers with over six years of collaboration with NVIDIA in multi-GPU architecture AI high-performance computing/cloud computing, including Quanta, Wistron, Wistron, Inventec, Foxconn, and Gigabyte, are expected to see operational benefits further reflected in the latter half of the year. Foxconn and Inventec are the main suppliers of GPU modules and GPU substrates, respectively, and are likely to witness noticeable shipment growth starting in the third quarter.

Furthermore, AI servers not only incorporate multiple GPU modules but also exhibit improvements in aspects such as chassis height, weight, and thermal design power (TDP) compared to standard servers. As a result, cooling solution providers like Asia Vital Components, Auras Technology, and SUNON; power supply companies such as Delta Electronics and Lite-On Technology; chassis manufacturers Chenbro; rail industry players like King Slide, and PCB/CCL manufacturers such as EMC, GCE are also poised to benefit from the increasing demand for AI servers.

(Source: https://ctee.com.tw/news/tech/915830.html)

Last week, major power semiconductor manufacturer Infineon announced plans to invest up to 5 billion euros over the next five years to construct the world’s largest 8-inch SiC power wafer factory in Kulim, Malaysia. This expansion will raise the total investment in the Kulim plant from 2 billion euros to 7 billion euros.

Interestingly, in February of this year, Wolfspeed announced its own plans to build what is touted as the world’s largest 8-inch SiC factory in the Saarland region of Germany. Infineon’s significant investment in the Malaysian 8-inch SiC factory sets the stage for potential competition with Wolfspeed, sparking an impending battle for Silicon Carbide production capacity.

In fact, driven by the rapid growth of industries like electric vehicles, the space for SiC power devices is expanding, attracting both Chinese companies and international enterprises to ramp up production.

According to statistics from TrendForce, aside from Wolfspeed, the first half of this year saw numerous companies, including STMicroelectronics, Mitsubishi Electric, Rohm, Soitec, and ON Semiconductor, expanding their production capacities. STMicroelectronics, for instance, announced a $4 billion investment in January to expand 12-inch wafer production. In June, they partnered with San’an Optoelectronics to establish a joint venture for 8-inch SiC device manufacturing, with an estimated total investment of around $3.2 billion.

On the Chinese front, there have been seven expansion projects related to Silicon Carbide. CRRC is investing 11.12 billion yuan to establish a project for the industrialization of medium and low-voltage power devices. YASC is also planning to construct a Compound Semiconductor power device production project, encompassing epitaxial growth, wafer manufacturing, packaging, and testing lines. Upon completion, the facility will have an annual production capacity of 360,000 6-inch SiC wafers and 61 million power device modules.

Additionally, BYD plans to invest 200 million yuan to establish a SiC epitaxial trial production and mass production project at its automotive production base in Shenzhen. The expansion will add 6,000 SiC epitaxial wafers per year, bringing the total capacity to 18,000 wafers per year.

(Photo credit: Tesla)

According to the latest report from TrendForce, the primary factors influencing the global market share of notebook CPUs in 2024 can be categorized into “Architectural Design” and “Economic Factors.”

“Architectural Design” as a long-term factor affecting market share:

(1) Both AMD (AMD 3D V-Cache) and Intel (Intel Foveros Direct) may potentially integrate 3D packaging technology into notebook computers in the future.

(2) Apple’s M-series processors, using the Arm core architecture, as well as Intel processors, have adopted a big/little core hybrid design. AMD might also introduce this in the Ryzen 8000 series.

(3) Despite further advancements in processor technology by 2024, the notebook computer market remains highly sensitive to the cost for IT equipment.

“Economic Factors” as more immediate influencers of market share:

(1) Until 2024, a return to lower interest rates in the global economic environment could favor corporate expansion of capital expenditure. This could result in increased procurement of business-oriented notebook models, potentially allowing Intel to further expand its CPU market share beyond 70% in the business sector.

(2) Concerns about economic prospects among global citizens until 2024 could have significant negative implications for the consumer notebook computer market. With the restart of physical economic activities, the demand for consumer-oriented notebook models has declined from the high levels seen during the pandemic. Consequently, the consumer market demand outlook for 2024 remains uncertain. For AMD, which relies more on consumer market demand, changes in market share may be harder to predict compared to Intel.

In the post-pandemic era, AMD, Arm/Apple, and Intel are pursuing distinct technological competition strategies to capture market share in the personal computing market.

AMD:

(1) The Socket AM5 platform is poised to aid AMD CPUs in achieving substantial performance and efficiency gains.

(2) The AMD Ryzen 7040 incorporates an artificial intelligence engine to emphasize AI computing performance’s importance in the thin and light notebook market.

Arm/Apple:

(1) The M2 Ultra processor heralds Apple’s complete transition of personal computing products to the Arm core. Apple Mac computer products will no longer be sold with Intel processor.

(2) The Apple M-series processors, built on the Arm core architecture, facilitate a “fanless design” to maintain MacBook’s slim profile. This feature highlights its irreplaceable positioning in the portable notebook computer market, emphasizing portability.

Intel:

(1) With the waning trend of the “hybrid work mode,” Intel is optimistic about diversified development in the post-pandemic era for desktop computer products. This includes microcomputers, micro workstations, and general workstations. Due to the characteristic of continuous operation for 24 hours, desktop computers still possess unique attributes that cannot be replaced by notebook computers.

(Photo credit: Intel)

DRAM Spot Market:

Continuing from the previous week, the influx of used chips that were stripped from decommissioned server DRAM modules is still causing spot prices of DDR4 products to drop. As for DDR5 products, there is no such issue affecting their supply. However, module houses are holding a high level of inventory for DDR5 products because they stocked up aggressively in 2Q23, while the actual demand remains in a slump. Hence, spot prices on the whole have kept falling. The average spot price of the mainstream chips (i.e., DDR4 1Gx8 2666 MT/s) dropped by 0.74% from US$1.479 last week to US$1.468 this week.

NAND Flash Spot Market:

The spot market has been lingering at sluggish transactions since July under feeble inquiries between market participants. Fortunately, the reduction of spot prices has somewhat mitigated over the past two weeks due to suppliers’ continuous intention in an on-going production cut and price increment. Certain components have not seen a further drop in prices after a consolidation at the low end, though the level of transactions remains at a rather insignificant extent, and the continuity of purchases will require further observations. 512Gb TLC wafer spots have risen by 1.21% this week, arriving at US$1.421.

Looking at the impact of AI server development on the PCB industry, mainstream AI servers, compared to general servers, incorporate 4 to 8 GPUs. Due to the need for high-frequency and high-speed data transmission, the number of PCB layers increases, and there’s an upgrade in the adoption of CCL grade as well. This surge in GPU integration drives the AI server PCB output value to surpass that of general servers by several times. However, this advancement also brings about higher technological barriers, presenting an opportunity for high-tech PCB manufacturers to benefit.

TrendForce’s perspective: 

• The increased value of AI server PCBs primarily comes from GPU boards.

Taking the NVIDIA DGX A100 as an example, its PCB can be divided into CPU boards, GPU boards, and accessory boards. The overall value of the PCB is about 5 to 6 times higher than that of a general server, with approximately 94% of the incremental value attributed to the GPU boards. This is mainly due to the fact that general servers typically do not include GPUs, while the NVIDIA DGX A100 is equipped with 8 GPUs.

Further analysis reveals that CPU boards, which consist of CPU boards, CPU mainboards, and functional accessory boards, make up about 20% of the overall AI server PCB value. On the other hand, GPU boards, including GPU boards, NV Switch, OAM (OCP Accelerator Module), and UBB (Unit Baseboard), account for around 79% of the total AI server PCB value. Accessory boards, composed of components such as power supplies, HDD, and cooling systems, contribute to only about 1% of the overall AI server PCB value.

• The technological barriers of AI servers are rising, leading to a decrease in the number of suppliers.

Since AI servers require multiple card interconnections with more extensive and denser wiring compared to general servers, and AI GPUs have more pins and an increased number of memory chips, GPU board assemblies may reach 20 layers or more. With the increase in the number of layers, the yield rate decreases.

Additionally, due to the demand for high-frequency and high-speed transmission, CCL materials have evolved from Low Loss grade to Ultra Low Loss grade. As the technological barriers rise, the number of manufacturers capable of entering the AI server supply chain also decreases.

Currently, the suppliers for CPU boards in AI servers include Ibiden, AT&S, Shinko, and Unimicron, while the mainboard PCB suppliers consist of GCE and Tripod. For GPU boards, Ibiden serves as the supplier, and for OAM PCBs, Unimicron and Zhending are the suppliers, with GCE, ACCL, and Tripod currently undergoing certification. The CCL suppliers include EMC. For UBB PCBs, the suppliers are GCE, WUS, and ACCL, with TUC and Panasonic being the CCL suppliers.

Regarding ABF boards, Taiwanese manufacturers have not yet obtained orders for NVIDIA AI GPUs. The main reason for this is the limited production volume of NVIDIA AI GPUs, with an estimated output of only about 1.5 million units in 2023. Additionally, Ibiden’s yield rate for ABF boards with 16 layers or more is approximately 10% to 20% higher than that of Taiwanese manufacturers. However, with TSMC’s continuous expansion of CoWoS capacity, it is expected that the production volume of NVIDIA AI GPUs will reach over 2.7 million units in 2024, and Taiwanese ABF board manufacturers are likely to gain a low single-digit percentage market share.

(Photo credit: Google)