ChatGPT’s debut has sparked a thrilling spec upgrade in the server market, which has breathed new life into the supply chain and unlocked unparalleled business opportunities. Amidst all this, the big winners look set to be the suppliers of ABF (Ajinomoto Build-up Film) substrates, who are poised to reap enormous benefits.

In the previous article, “AI Sparks a Revolution Up In the Cloud,” we explored how the surge in data volumes is driving the spec of AI servers as well as the cost issue that comes with it. This time around, we’ll take a closer look at the crucial GPU and CPU platforms, focusing on how they can transform the ABF substrate market.

NVIDIA’s Dual-Track AI Server Chip Strategy Fuels ABF Consumption

In response to the vast data demands of fast-evolving AI servers, NVIDIA is leading the pack in defining the industry-standard specs.

This contrasts with standard GPU servers, where one CPU backs 2 to 6 GPUs. Instead, NVIDIA’s AI servers, geared towards DL(Deep Learning) and ML(Machine Learning), typically support 2 CPUs and 4 to 8 GPUs, thus doubling the ABF substrate usage compared to conventional GPU servers.

NVIDIA has devised a dual-track chip strategy, tailoring their offerings for international and Chinese markets. The primary chip for ChatGPT is NVIDIA’s A100. However, for China, in line with U.S. export regulations, they’ve introduced the A800 chip, reducing interconnect speeds from 600GBps (as on the A100) to 400GBps.

Their latest H100 GPU chip, manufactured at TSMC’s 4nm process, boasts an AI training performance 9 times greater than its A100 predecessor and inferencing power that’s 30 times higher. To match the new H100, H800 was also released with an interconnect speed capped at 300GBps. Notably, Baidu’s pioneering AI model, Wenxin, employs the A800 chip.

To stay competitive globally in AI, Chinese manufacturers are expected to aim for the computational prowess on par with the H100 and A100 by integrating more A800 and H800 chips. This move will boost the overall ABF substrate consumption.

With the ChatBot boom, it is predicted a 38.4% YoY increase in 2023’s AI server shipments and a robust CAGR of 22% from 2022 to 2026 – significantly outpacing the typical single-digit server growth, according to TrendForce’s prediction.

AMD, Intel Server Platforms Drive ABF Substrate Demand

Meanwhile, examining AMD and Intel’s high-end server platforms, we can observe how spec upgrades are propelling ABF substrate consumption forward.

• AMD Zen 4:

Since 2019, AMD’s EPYC Zen 2 server processors have used Chiplet multi-chip packaging, which due to its higher conductivity and cooling demands, has consistently bolstered ABF substrate demand.

• Intel Eagle Stream:

Intel’s advanced Eagle Stream Sapphire Rapids platform boasts 40-50% higher computation speed than its predecessor, the Whitley, and supports PCIe5, which triggers a 20% uptick in substrate layers. This platform employs Intel’s 2.5D EMIB tech and Silicon Bridge, integrating various chips to minimize signal transmission time.

The Sapphire Rapids lineup includes SPR XCC and the more advanced SPR HBM, with the latter’s ABF substrate area being 30% larger than the previous generation’s. The incorporation of EMIB’s Silicon Bridge within the ABF substrate increases lamination complexity and reduces overall yield. Simply put, for every 1% increase in Eagle Stream’s server market penetration, ABF substrate demand is projected to rise by 2%.

As the upgrades for server-grade ABF substrates continue to advance, production complexity, layer count, and area all increase correspondingly. This implies that the average yield rate might decrease from 60-70% to 40-50%. Therefore, the actual ABF substrate capacity required for future server CPU platforms will likely be more than double that of previous generations.

ABF Substrate Suppliers Riding the Tide

By our estimates, the global ABF substrate market size is set to grow from $9.3 billion in 2023 to $15 billion in 2026 – a CAGR of 17%, underscoring the tremendous growth and ongoing investment potential in the ABF supply chain.

Currently, Taiwanese and Japanese manufacturers cover about 80% of the global ABF substrate capacity. Major players like Japan’s Ibiden, Shinko and AT&S, along with Taiwan’s Unimicron, Nan Ya, and Kinsus all consider expanding their ABF substrate production capabilities as a long-term strategy.

As we analyzed in another piece, “Chiplet Design: A Real Game-Changer for Substrates,” despite the recent economic headwinds, capacity expansion of ABF substrate can still be seen as a solid trend, which is secured by the robust growth of high-end servers. Hence, the ability to precisely forecast capacity needs and simultaneously improve production yields will be the key to competitiveness for all substrate suppliers.

Read more:

• HBM Supply Leader SK Hynix’s Market Share to Exceed 50% in 2023 Due to Demand for AI Servers, Says TrendForce
• Shipments of AI Servers Will Climb at CAGR of 10.8% from 2022 to 2026, Says TrendForce

(Photo Credit: Google)

According to TrendForce research, corporate demand for digital transformation including artificial intelligence and high-performance computing has accelerated in recent years, which has led to increasing adoption of cloud computing. In order to improve service flexibility, the world’s major cloud service providers have gradually introduced ARM-based servers. The penetration rate of ARM architecture in data center servers is expected to reach 22% by 2025.

In the past few years, ARM architecture processors have matured in the fields of mobile terminals and Internet of Things but progress in the server field has been relatively slow. However, companies have diversified cloud workloads in recent years and the market has begun to pay attention to the benefits ARM architecture processing can provide to data centers. TrendForce believes that ARM-based processors have three major advantages. First, they can support diverse and rapidly changing workloads and are more scalability and cost-effective. Second, ARM-based processors provide higher customization for different niche markets with a more flexible ecosystem. Third, physical footprint is relatively small which meets the needs of today’s micro data centers.

Influenced by geopolitics and the strengthening of data sovereignty in various countries, major cloud service providers and telecom operators are actively developing micro data centers which will further drive the penetration of ARM-based processors. At the same time, from the perspective of cloud service providers currently adopting ARM-based processors, Graviton, led by AWS, has the largest market scale and began encroaching gradually into the market in 2021. TrendForce also observed that AWS’s deployment of ARM-based processors in 2021 reached 15% of overall server deployment and will exceed 20% in 2022. This forces other major cloud service providers to keep up by initiating their own projects at various foundries. If testing is successful, these projects are expected to start mass introduction in 2025.

In addition, according to the Neoverse Platform plan previously released by ARM, its Platform Roadmap will also be one of the key drivers of penetration. This product line is set up to target ultra-large-scale data centers and edge computing infrastructure. However, it is worth mentioning, since x86 is still mainstream in the market and ARM-based server CPU suppliers only maintain small-batch production orders at this stage and primarily focus on ultra-large-scale data centers, introduction of ARM-based servers into enterprise data centers will be slow going. Thus, TrendForce believes that it will still be difficult for ARM-based servers to compete with x86-based servers before 2025.

According to TrendForce research, due to the vigorous stocking of various terminal applications causing a shortage of wafers in 2021, the global IC industry was severely undersupplied. This, coupled with spiking chip prices, boosted the 2021 revenue of the global top ten IC design companies to US$127.4 billion, or 48% YoY.

TrendForce further indicates three major disparities from the 2020 ranking. First, NVIDIA surpassed Broadcom to take the second position. Second, Taiwanese companies Novatek and Realtek rose to sixth and eighth place, respectively. Originally ranked tenth, Dialog was replaced at this position by Himax after Dialog was acquired by IDM giant Renesas.

Qualcomm continues its reign as number one in the world, primarily due to 51% and 63% growth YoY in sales of mobile phone SoC (System on Chip) and IoT chips, respectively. The addition of diversified development in its RF and automotive chip businesses was key to a 51% increase in revenue. NVIDIA implemented the integration of software and hardware, demonstrating its ambitions in creating a “comprehensive computing platform.” Driven by the annual growth of gaming graphics card and data center revenue at 64% and 59%, respectively, NVIDIA successfully climbed to second place. Broadcom benefited from the stable sales performance of network chips, broadband communication chips, and storage and bridging chips, with revenue growing 18% YoY. AMD’s computer and graphics revenue grew by 45% YoY due to strong sales of the Ryzen CPU and Radeon GPU and rising average selling price. Coupled with accelerating demand from cloud companies, the annual revenue of AMD’s enterprise, embedded, and semi-customized divisions increased by 113%, driving annual growth of total revenue to 68%.

In terms of Taiwanese firms, MediaTek’s strategy of focusing on mobile phone SoC has produced miraculous results. Benefiting from an increase in 5G penetration, the sales performance of MediaTek’s mobile phone product portfolio surged by 93% and the company has committed to increasing the proportion of high-end product portfolios, resulting in 61% annual revenue growth. Novatek’s two major product lines of SoC and display driver IC have both grown significantly. Due to improved product specifications, increased shipments, and beneficial pricing gains, revenue grew by 79% YoY, the highest among the top ten. Realtek has been driven by strong demand for Netcom and commercial notebook products, while the performance of audio and Bluetooth chips remains quite stable, conferring an annual revenue growth of 43%. Himax joins the top ten ranking for the first time in 2021. Due to significant annual revenue growth in large-sized and medium/small-sized driver IC of 65% and 87%, respectively, and the successful introduction of driver IC into automotive panels, total revenue exceeded US$1.5 billion, or 74% YoY.

Looking forward to 2022, after AMD completes the acquisition of Xilinx, other players will fill out the rankings. In the broader picture, intensifying demand for high-specification products such as high-performance computing, Netcom, high-speed transmission, servers, automotive, and industrial applications will create good business opportunities for IC design companies and drive overall revenue growth. However, terminal system manufacturers face the correction of component mismatch issues. In addition, growing foundry costs, intensifying geopolitical conflicts, and rising inflation will all be detrimental to global economic growth and may impact an already weakened consumer electronics market. These are the challenges IC design companies face in 2022 and by what means can product sales momentum be maintained within existing production capacity, R&D efficacy strengthened, and chip specifications upgraded, will become the primary focus of development in 2022.

The semicondustor market in 3Q21 is red hot with total revenue of the global top 10 IC design (fabless) companies reaching US$33.7 billion or 45% growth YoY, according to TrendForce’s latest investigations. In addition to the Taiwanese companies MediaTek, Novatek, and Realtek already on the list, Himax comes in at number ten, bringing the total number of Taiwanese companies on the top 10 list to 4.

Qualcomm has been buoyed by continuing robust demand for 5G mobile phones form major mobile phone manufacturers with further revenue growth from its processor and radio frequency front end (RFFE) departments. Qualcomm’s IoT department benefited from strong demand in the consumer electronics, edge networking, and industrial sectors, posting revenue growth of 66% YoY, highest among Qualcomm departments. In turn, this drove Qualcomm’s total 3Q21 revenue to US$7.7 billion, 56% growth YoY, and ranking first in the world.

Second ranked NVidia, is still benefiting from gaming graphics card and data center revenue as the annual revenue growth for these two primary product departments reached 53% and 48%, respectively. In addition, professional design visualization solutions only accounted for 8% of total revenue. However, due to enduringly strong demand for mining and customers actively deploying the RTX series of high-performance graphics cards, NVidia’s product department revenue grew 148% YoY with overall revenue increasing by 55% to US$6.6 billion.

Third ranked Broadcom’s main revenue stream came from their network chip, broadband communication chip and storage and bridge chip businesses. Driven by post-COVID hybrid working models, companies are accelerating migration to the cloud, increasing demand for Broadcom chips, and driving revenue growth to US$5.4 billion or 17% YoY. AMD’s Ryzen, Radeon, and EPYC series of products in the fields of games, data centers, and servers performed well, driving total revenue to US$4.3 billion, 54% growth YoY, and fifth place overall.

In terms of Taiwanese companies, MediaTek continues to expand its global 5G rollout and, benefiting from optimization of product portfolio composition, product line specification enhancement, increase in sales volume, increases in pricing, and other factors, revenue of MediaTek’s mobile phone product line increased 72% YoY. Annual revenue of other product lines also posted double digit growth with total revenue in the 3Q21 reaching US$4.7 billion or 43% YoY, a fourth place ranking. Novatek continues to focus on its two primary product lines of system-on-chip and panel driver chips. The proportion of its OLED panel driver chip shipments has increased, product ASP has risen, and shipments have been smooth with 3Q21 revenue reaching US$1.4 billion or 84% YoY. In addition, Realtek’s revenue surpassed Xilinx to take the eighth position due to higher priced Netcom chips in 3Q21. Himax also saw significant growth in its three main product lines of TVs, monitors, and notebooks due to large-size driver chips. Revenue from large-size driver chips increased 111% YoY, driving total revenue to exceed the US$400 million mark, a 75% increase, and enough to squeeze onto this year’s ranking.

Overall, 3Q21 revenue for major IC design (fabless) companies has generally reached historic levels. Rankings for the top 7 companies remained the same as in 2Q21 with change coming in ranks 8 to 10. Looking forward to 4Q21, TrendForce believes Taiwanese IC design (fabless) companies will generally lean conservative. In addition to the electronics industry moving into the traditional off-season, a slowing of demand for consumer applications and customer-end materials supply issues reducing procurement will make continued revenue growth a challenge. In addition to consumer electronic products, global industry leaders are focused on the positive development of server and data center products to maintain an expected revenue growth trend.

For more information on reports and market data from TrendForce’s Department of Semiconductor Research, please click here, or email Ms. Latte Chung from the Sales Department at lattechung@trendforce.com

Nvidia hosted its fall GTC (GPU Technology Conference) in early November, during which the company shared details regarding the progress that it had made on products and services such as AI software, data centers, automotive applications, and healthcare. In particular, Nvidia’s foray into virtual worlds and digital twins, both of which are closely tied to the metaverse, garnered significant attention from the public. By leveraging diverse simulation tools that reflect real-life circumstances, Nvidia has extended the application of virtual worlds from the local scale to the planetary scale, thereby reflecting the metaverse’s pioneering qualities and developmental progress.

Along with the ongoing metaverse craze, Nvidia also released its Omniverse Avatar technology platform as well as its Omniverse Replicator, which is a “synthetic data-generation engine” according to the company. Both of these releases are based on the Nvidia Omniverse, a platform that specializes in virtual collaboration. Whereas the Omniverse Avatar platform enables the creation of interactive virtual characters through synergies among voice AI technology, machine vision, and NLP (natural language processing), the Omniverse Replicator constructs more realistic, lifelike virtual worlds by training DNN (deep neural networks) using such synthetic data as velocity, depth, and weather conditions.

Digital twin-based virtual factories are starting to show the first hints of the metaverse

The metaverse value chain primarily revolves around commonly seen infrastructural backbones formed by telecommunications and cloud/edge computing. The virtual space that is then built on top of this infrastructure comprises HMI (human machine interface), decentralization, application creation, and user experiences. More specifically, HMI produces an AI-empowered immersive experience by combining multiple interactive technologies with an AR/VR base layer. At the moment, companies such as Nvida, Meta (formerly known as Facebook), Microsoft (including Xbox), and Vive are heavily invested in HMI development. Application creation, on the other hand, refers to mechanisms that make the metaverse more lively, reliable, diverse, and attractive. Some examples include graphical tools and cryptocurrency technologies. Representative groups focusing on this field include Roblox, IBM, Google AI, Epic, and Unity.

Regarding the content of Nvidia’s presentation during GTC apart from the Omniverse Avatar and Replicator, the company also released CloudXR, Showroom, and other Omniverse-based tools used for optimizing immersive experiences. As well, Nvidia also released the Modulus neural network model, which is accelerates the build-out of digital twins. These releases, in turn, demonstrates Nvidia’s competency and leadership in creating AI-driven software tools for the metaverse value chain. With regards to real-life use cases, digital twins currently represent most of Nvidia’s applications. For instance, BMW and Nvidia have partnered to construct a digital twin-based factory via the Omniverse platform capable of connecting ERP (enterprise resource management), shipment volume simulation, remote-controlled robots, production line simulation, etc. This partnership is indicative of promising early-stage growth of the metaverse.

Nvidia is extending its simulation application from factories to planets

While smart city development has remained one of the main use cases of simulation in recent years, Nvidia has further extended its simulation applications from use cases previously limited to singular offices or factory facilities. For instance, BIM (building information modeling) specialist Bentley Systems has teamed up with Nvidia to apply digital twins to public property management and maintenance. Ericsson, on the other hand, is utilizing Nvidia’s technology to construct a digital replica of an entire city for the purpose of checking 5G signal coverages, optimizing base station placement, and improving antenna designs. During the GTC, Nvidia unveiled the Earth-2 system, which is a supercomputer that generates a digital twin of planet earth for weather forecasts.

As a matter of fact, most products and services announced by Nvidia during GTC represent either a partial or entry-level application of the metaverse. However, as the post-pandemic new normal continues to drive up the demand for contactless and digital transformation applications, strengthening CPS (cyber physical systems) will remain one of the most significant trends in the market. As real-world environments become increasingly complex due to interactions among an increasing number of tools and use cases, Nvidia will aim to create a comprehensive framework for metaverse development through products/services based on more intelligent, comprehensive, and instant virtual worlds. Hence, TrendForce believes that Nvidia will need to address certain major challenges going forward, including lowering its tools’ usage barriers, strengthening its ecosystem, and attracting new users.

（Image credit: NVIDIA）