TSMC operates at full capacity, AMD aims for AI chips reportedly seeks CoWoS-like supply chain.

In 2023, NVIDIA led the global AI chip development, and in 2024, the global demand for AI chips is expected to continue to surge due to the expansion of end-user applications such as PCs and mobile phones.

Meanwhile, AMD has not stopped in AI chip development either, with the expected MI300 products poised to heat up the global AI business opportunities. However, the key to supply lies in advanced packaging, and AMD will seek outsourced semiconductor assembly and test (OSAT) service providers to offer support similar to CoWoS.

According to Taiwan’s Commercial Times, TSMC’s CoWoS capacity has long been fully loaded, and even if it expands production this year, it will mainly be reserved for NVIDIA. Market sources pointed out that TSMC will continue to increase CoWoS capacity to support AMD’s demand, but it takes six to nine months to establish a new production line. Therefore, it is expected that AMD will seek cooperation with other companies with CoWoS-like packaging capabilities. ASE, Amkor, Powertech, and KYEC are the first batch of potential partners.

TSMC has been outsourcing part of its CoWoS operations for some time, mainly targeting small-volume, high-performance chips. TSMC maintains in-house production of the CoW, while the back-end WoS is handed over to test and assembly houses to improve production efficiency and flexibility. This model will continue in the future 3D IC generation.

ASE and Amkor both received WoS orders last year. ASE has strengthened the development of advanced packaging technology and has a complete solution for the entire CoWoS process. ASE previously stated that it sees the strong potential of AI and expects related revenue to double in 2024.

According to reports citing market sources, the monthly production capacity of the ASE Group’s 2.5D packaging is about 2,000 to 2,500 pieces. Some experts believe that test and assembly houses will maintain the business model of TSMC or UMC providing the interposer. Therefore, in 2024, a significant increase in CoWoS production capacity is expected.

KYEC is responsible for testing Nvidia AI chips and is expected to benefit from AMD’s search for CoWoS-like capacity. Nvidia is currently KYEC’s second-largest customer.

KYEC’s testing of Nvidia A100 and H100 chips is mainly in the final test (FT), with a market share of up to 70%. KYEC provides comprehensive IC burn-in testing, has self-developed burn-in equipment, and has been in the industry for more than a decade, accumulating many patents and technologies.

AMD stated at the end of 2023 that AI chip revenue could reach US$2 billion in 2024, excluding other HPC chips. AMD pointed out that the annual compound growth rate of the AI chip market in the next four years will reach 70%, and it is estimated that it will reach US$400 billion in 2027.

(Image: AMD)

Major Cloud Service Providers (CSPs) continue to see an increase in demand for AI servers over the next two years. The latest projections of TrendForce indicate a global shipment of approximately 1.18 million AI servers in 2023, with a year-on-year growth of 34.5%. The trend is expected to persist into the following year, with an estimated annual growth of around 40.2%, constituting over 12% of the total server shipments.

NVIDIA, with its key products including AI-accelerating GPU and the AI server reference architecture HGX, currently holds the highest market share in the AI sector. However, it is crucial to monitor CSPs developing their own chips and, in the case of Chinese companies restricted by U.S. sanctions, expanding investments in self-developed ASICs and general-purpose AI chips.

According to TrendForce data, AI servers equipped with NVIDIA GPUs accounted for approximately 65.1% this year, projected to decrease to 63.5% next year. In contrast, servers featuring AMD and CSP self-developed chips are expected to increase to 8.2% and 25.4%, respectively, in the coming year.

Another critical application, HBM (High Bandwidth Memory), is primarily supplied by major vendors Samsung, SK Hynix, and Micron, with market shares of approximately 47.5%, 47.5%, and 5.0%, respectively, this year. As the price difference between HBM and DDR4/DDR5 is 5 to 8 times, this is expected to contribute to a staggering 172% year-on-year revenue growth in the HBM market in 2024.

Currently, the three major manufacturers are expected to complete HBM3e verification in the first quarter of 2024. However, the results of each manufacturer’s HBM3e verification will determine the final allocation of procurement weight for NVIDIA among HBM suppliers in 2024. As the verifications are still underway, the market share for HBM in 2024 remain to be observed.

Read more

• [News] Memory Titans Vie for Control in HBM Tech, Who Will Shape the Next-Gen?
• [Insights] NVIDIA Maintains Confidence Despite Export Restrictions, Continues Unveiling New Products

(Photo credit: NVIDIA)

As the Japanese government injects substantial funds to revitalize its semiconductor industry, the dynamics of the Japanese semiconductor industry have been a global focus. TrendForce, during this year’s SEMICON Japan, organized its first overseas industry-focused information seminar, delving into the global semiconductor, optoelectronics, and electric vehicle industries, with a particular focus on the dynamics and strategies of the Japanese market and companies. The event attracted over a hundred participants from Japanese technology industry.

The seminar, opened by TrendForce CEO Kevin Lin, under the theme of “the era of challenges,” served as a commentary on the future development of the technology industry in the coming years. Lin pointed out that global technological industries, influenced by geopolitical factors, are experiencing a trend of supply chain restructuring. He also highlighted China’s expansion in the semiconductor, electric vehicle, and downstream supply chains, reshaping the global supply chain landscape—an aspect requiring global attention.

During the seminar, TrendForce’s Senior Research Vice President, Ken Kuo, presented an analysis of the global memory and AI server market. He noted that after a year and a half of adjustments, prices in the DRAM and NAND markets started to rise across the board in the fourth quarter, driven primarily by robust growth in AI. This trend is expected to continue into the next year.

Beyond AI servers, the introduction of technologies such as Microsoft’s Copilot, as well as AI PCs and AI smartphones, is poised to be a growth driver next year. In terms of AI chip shipments, NVIDIA is projected to maintain its dominance, with an estimated 1.5 million units shipped this year and an anticipated 100% growth next year.

The semiconductor foundry market is expected to recover in 2024.

The recovery of the semiconductor market in 2024 was a major focus for participants. TrendForce’s analyst Joanne Chiao mentioned that as supply chain inventory pressures gradually ease, the semiconductor foundry industry is expected to experience a recovery in 2024, driven by TSMC’s advanced processes and inventory replenishment momentum, with a projected growth of 7%.

In light of geopolitics, , semiconductor foundry supply chains are undergoing restructuring. In 2023, Taiwan is expected to account for approximately 46% of global semiconductor foundry capacity, followed by China at 26%, South Korea at 12%, the United States at 6%, and Japan at 2%. With the drive from subsidy policies in China and the United States to increase local production capacity, by 2027, Taiwan and South Korea’s production capacity shares are expected to converge to 41% and 10%, respectively.

Meanwhile, Japan is actively implementing subsidy policies to support local company Rapidus and attract Taiwan’s TSMC and PSMC to establish facilities, aiming to secure a place in the semiconductor foundry market.

The introduction of Apple Watch with Micro LED is expected in 2026, with estimated display costs 2.5-3 times higher than OLED.

TrendForce’s Senior Research Vice President, Eric Chiou, analyzed Apple’s progress in adopting new display technologies during the display technology session. He mentioned that the next-generation Apple Watch panel would use Micro LED as the display technology, with a size larger than the current Apple Watch Ultra at 2.12 inches.

The product will have two key suppliers: German LED giant ams OSRAM, which will exclusively supply Micro LED chips smaller than 10x10um, and South Korean panel manufacturer LG Display, responsible for the chip mass transfer engineering in addition to providing LTPO glass backplates.

Chiou pointed out that the adoption of small-sized chips inherently helps compress costs. Considering Apple’s strong bargaining power in the supply chain, he estimated that when the product is launched in 2026, the cost of the Micro LED display panel could be controlled below $120, equivalent to 2.5 to 3 times the current price of OLED panels—a reasonable range for a new technology.

Moreover, with Apple’s outstanding ability to integrate new technologies and specifications, there is an expectation of achieving million-unit-level shipments in the first year of launch, injecting abundant vitality into the demand for Micro LED chips and the overall industry’s development.

China’s EV expansion brings impact to the global automotive industry.

In 2023, China became the world’s primary exporter of automobiles, prompting the global automotive industry to recognize that competition with Chinese automakers will extend from the domestic market to the global market. TrendForce analyst Caroline Chen highlighted in her speech that the most significant threat to international automakers is China’s advantage in EV( including BEV, PHV, FCV).

She emphasized that due to China’s early development of EVs, it has established a complete supply chain, particularly in the proactive development of power battery production capacity and upstream materials. EVs account for over a quarter of China’s passenger car exports.

On the other hands, with a nearly 60% market share in the Southeast Asian market, Chinese automakers gradually threaten Japanese automakers’ long-term dominance in the Southeast Asian automotive market.

She believes that as Chinese automakers expand into the international market, Japanese automakers should not only accelerate the development of new energy vehicles but also leverage their long-accumulated brand value and well-established maintenance systems as core competitive advantages. Additionally, maintaining leadership positions in semiconductor and chemical materials is a strategy for sustained investment to consolidate their influence in the automotive industry.

On the 28th, Amazon unveiled two AWS-designed chips, Graviton4, a CPU propelling its AWS cloud services, and the second-gen AI chip Trainium2, tailored for large language models. Both chips boast substantial performance upgrades. With a positive market outlook, Amazon is intensifying its competition with Microsoft and Google for dominance in the AI cloud market. The demand for in-house chips is surging, leading to increased orders for key players like the wafer foundry TSMC and the silicon design and production services company ALCHIP, reported by UDN News.

According to reports, Amazon AWS CEO Adam Selipsky presented the fourth AWS-Designed custom CPU chip, Graviton4, at the AWS re:Invent 2023 in Las Vegas. It claims a 30% improvement in computing performance compared to the current Graviton3, with a 75% increase in memory bandwidth. Computers equipped with this processor are slated to go live in the coming months.

Trainium2, the second-gen chip for AI system training, boasts a computing speed three times faster than its predecessor and doubled energy efficiency. Selipsky announced that AWS will commence offering this new training chip next year.

AWS is accelerating the development of chips, maintaining its lead over Microsoft Azure and Google Cloud platforms. Amazon reports that over 50,000 AWS customers are currently utilizing Graviton chips.

Notably, Amazon’s in-house chip development heavily relies on the Taiwan supply chain, TSMC and ALchip. To produce Amazon’s chips, Alchip primarily provides application-specific integrated circuit (ASIC) design services, and TSMC manufactures with advanced processes.

TSMC consistently refrains from commenting on products for individual customers. Analysts estimate that TSMC has recently indirectly secured numerous orders from Cloud Service Providers (CSPs), mainly through ASIC design service providers assisting CSP giants in launching new in-house AI chips. This is expected to significantly contribute to TSMC’s high utilization for the 5nm family.

In recent years, TSMC has introduced successive technologies such as N4, N4P, N4X, and N5A to strengthen its 5nm family. The N4P, announced at 2023 Technology Symposium, is projected to drive increased demand from 2024 onwards. The expected uptick in demand is mainly attributed to AI, network, and automotive products.

(Image: Amazon)

Recent reports from the IC design industry suggest that TSMC, the leading semiconductor foundry, is contemplating a slight price concession for certain mature processes next year, marking a return after three years. Despite its reputation for firm pricing, TSMC’s willingness to make concessions is seen as a response to a decrease in capacity utilization. According to UDN News, this shift may indicate the broader trend of semiconductor foundries facing pricing pressures due to lower capacity utilization.

Known for its stable pricing with minimal fluctuations, TSMC typically offers single-digit percentage annual concessions to clients. The reported concession for specific mature processes is estimated to be around 2%. TSMC, however, declined to comment on these pricing adjustments.

Several IC design companies have confirmed ongoing negotiations with TSMC regarding price concessions for the upcoming year. One disclosed that TSMC’s concession method involves settling after the completion of a full quarter’s production, offsetting the next quarter’s mask costs. This approach allows for low single-digit percentage concessions in the following quarters.

Industry sources suggest that other semiconductor foundries have already taken significant measures, such as direct price reductions on large orders and providing additional free wafer allocations, aiming to boost capacity utilization. Chinese chipmakers initiated price reductions earlier and more aggressively than their Taiwanese counterparts, maintaining TSMC’s relatively firm pricing.

The news of TSMC considering concessions for certain mature processes, while not a direct price reduction, holds indicative significance. It is likely to exert pricing pressure on other industry players with mature processes before the peak season arrives in the latter half of next year.

During the semiconductor shortage in recent years, TSMC initially refrained from raising prices. As a result, its pricing remained relatively lower, even the lowest, compared to other industry players who significantly increased their prices. TSMC reportedly canceled concessions in 2021 and 2022 and initiated a rare price increase at the beginning of 2023, rumored to be in the range of 3% to 6%.

However, with the semiconductor market reversing, the supply chain has been gradually adjusting inventory since the second half of 2022. In the first half of this year, TSMC reportedly introduced an “increase quantity feedback plan,” offering additional mature process wafer allocations for orders reaching a certain quantity.

Although TSMC relies on advanced processes for over 50% of its revenue, with mature processes not being its primary focus, they remain a market consideration.

(Image: TSMC)