Insights
South Korean media reported that Samsung is set to manufacture a new generation of Full Self-Driving (FSD) chips for Tesla’s Level 5 autonomous vehicles. These chips will be utilized in Tesla Hardware 5 (HW 5.0) onboard computers, with production expected to commence after 2025. The chips will be manufactured using Samsung’s 4nm process.
TrendForce’s analysis:
Samsung May Competing with TSMC for Tesla HW 5.0 Chips
In the early stages of Tesla’s autonomous driving technology, the company collaborated with Samsung for FSD chips used in various vehicle models, including Model 3, Model 5, Model X, and Model Y. However, in 2022, Tesla chose to work with TSMC, citing TSMC’s better yield performance in 4nm process technology at that time.
In response, Samsung has been actively improving its 3nm and 4nm process technologies within a short period. While Samsung’s 4nm process yield has reached 75%, it still slightly lags behind TSMC’s 80%. Despite this difference, given their previous collaborations, it is not ruled out that Tesla might place orders with both TSMC and Samsung this time. The main reason being Samsung’s plan to advance to the 2nm-level SF2 process technology in 2025 and further progress to the 1.4nm-level SF1.4 process technology in 2027, aligning its overall roadmap with TSMC’s. This advancement will assist Tesla in accelerating the production plan of its DOJO supercomputer, facilitating the transition to Level 5 autonomous driving.
(Photo credit: Tesla)
Insights
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:
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.
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)
Insights
China’s Automotive Price War Rages On: Some automakers have been gradually reclaiming outsourced orders for the battery, motor, electronic control system since May and June, shifting towards in-house production. Recently, they have asked suppliers to requote for second-half orders, with Samsung, Murata, Taiyo Yuden, PSA and Yageo actively vying for contracts.
Due to the more stringent certifications in the automakers’ supply chain compared to tier 1 suppliers, the majority of battery, motor, electronic control system MLCC suppliers still come from Taiwan, Japan, and Korea. Among them, Korean manufacturer Samsung has made significant progress in the Chinese automotive market this year. They have been actively providing sample for certifications and competitive pricing, securing a large share of orders and displacing Japanese manufacturers Murata and TDK, who had long held the lead.
Ongoing negotiations between automakers are expected to conclude with finalized orders by the end of August. According to the channel check from TrendForce, it appears that Samsung will maintain its leading position with a low-price strategy, while Murata, unwilling to be drawn into a price war reminiscent of consumer electronics, will remain conservative with pricing to secure a substantial market share. Taiyo Yuden, PSA and Yageo, though limited in automotive product offerings, have been proactive in their bidding efforts and have secured several orders.
(Photo credit: Yageo)
In-Depth Analyses
As the pandemic has eased, the global automotive market is picking up momentum, and it is estimated that the global shipments of automotive panels will exceed 200 million units in 2023. With the continuous demand for size enlargement and specification improvement in automotive panels, the adoption of TDDI architecture is becoming more prevalent, and it is expected that TDDI will gradually become the mainstream for automotive panels.
On the other hand, AMOLED panels have started to have opportunities for adoption in emerging electric vehicles and some high-end car models. However, their adoption has been slow due to potential issues with reliability, lifespan, and brightness. Currently, the overall penetration rate for AMOLED panels in the automotive sector is estimated to reach 6% by 2026.
Can Panel Manufacturers Replace Traditional Tier 1 Players and Directly Serve Automakers?
As traditional internal combustion engine vehicles transition to electric vehicles and the level of in-car electronics continues to rise, coupled with the development of autonomous driving technologies, the demand for automotive displays is constantly expanding. The integration of digital display panels with touch functionality is gradually becoming mainstream, and panel sizes are increasing, moving towards more integrated designs. Specifications such as resolution, wide viewing angles, and high refresh rates, as well as unique designs, are becoming focal points. Currently, display panel specifications are moving towards LTPS LCD panels, which offer larger sizes, superior display performance, and better energy efficiency.
Looking at the market conditions, after the outbreak of the pandemic in 2020, the demand for automotive panels declined, but it gradually recovered in 2021 and 2022. However, there is still an oversupply situation, and it is estimated that there will be a slight growth of 5.1% to reach 205 million units in 2023. In terms of shipment scale, China’s panel shipments maintain the best position with a share of over 40%, while Japanese panel manufacturers have been squeezed by Chinese counterparts, reducing their share to about 20%. Taiwan’s panel manufacturers account for approximately 21%, and Korean panel manufacturers represent 8%.
The traditional shipment model involves Tier 1 players contracting with car manufacturers for related validation, assembly, and supply chain management roles, and then subcontracting Tier 2 panel suppliers. With the transformation of the automotive industry and the semiconductor component shortages in the past few years, as well as the increased requirements for interior design in vehicles, car manufacturers are starting to seek better control over the supply chain. As a result, panel manufacturers may replace Tier 1 players and directly supply to automakers, and Tier 1 suppliers will face competition from panel manufacturers.
The Automotive TDDI Architecture Has Cost Advantages
In the early days, LCD automotive panels mainly used external touch solutions, with car-use DDI and independent touch ICs on the IC architecture. However, as panel sizes increased, the number of ICs used also increased, leading to higher costs. Therefore, the TDDI architecture became a new development direction.
TDDI is commonly used for panels up to 30 inches in size. A single TDDI solution can be used for 20-inch panels, while for 20-30 inch panels, a TDDI-cascade solution with approximately 2-3 TDDI-cascade architectures is often used. Panels larger than 30 inches use the LTDI (Local TDDI) structure.
New Display Technology Awaits Automotive Certification; Significant Growth Expected after 2025
AMOLED is mostly used in high-end car models or stylish new electric vehicles, but its rapid development is hindered by limitations in brightness, panel lifespan, and reliability. In comparison, LCDs with MiniLED BLU architecture offer similar display performance to AMOLED but at a more affordable price and with better safety, and they are expected to compete with AMOLED in the market.
For more information on this report or market data from TrendForce’s Department of Display Research, please click here, or email Ms. Grace Li from the Sales Department at graceli@trendforce.com
Insights
Tesla’s Shanghai factory has reportedly initiated layoffs among its battery assembly workforce. Industry sources suggest that the majority of the layoffs will affect employees in the first phase of battery assembly, with the reduction expected to exceed 50%. While most of the affected individuals will be offered compensation through negotiations, a small number will be reassigned to other positions. Additionally, the equipment in the first phase of battery assembly will either be dismantled or relocated.
From a production capacity standpoint, Tesla’s Shanghai factory currently operates at a capacity of approximately 100,000 vehicles per month. In order to maintain product scarcity and brand image, the output is expected to be controlled within the range of 75% to 85%.
According to TrendForce’s understanding, the layoffs in the first phase of battery assembly are expected to be related to US government policies. The US government has imposed restrictions on subsidizing batteries imported from China and requires the use of locally manufactured batteries. As a result, export orders for batteries from Tesla’s Shanghai factory have been cut, leading to excess production capacity. Tesla, known for its efficiency-driven corporate culture, is intolerant of resource wastage.
On another note, the reduction in capacity and production volume of the first phase of battery assembly by Tesla may indicate preparations for transferring some of the capacity to the United States. By completing the battery pack manufacturing process in the United States, Tesla aims to increase the proportion of the value chain related to battery production in the US, in order to qualify for the full subsidy of USD 7,500 per vehicle in the United States.
(Photo credit: Tesla)