On August 27, AI server giant Supermicro was accused of accounting violations, inadequate disclosure of related party transactions, and evading sanctions by selling products to Russia by short-seller Hindenburg Research.

In addition, Supermicro announced on August 28 that it would delay the release of its annual report, potentially facing order withdrawals. Industry sources also believe this news presents a chance for Supermicro’s competitor Dell to gain market share.

Besides Dell, a report from Commercial Times also points out that Hewlett Packard Enterprise (HPE) could benefit from the shift in orders, potentially boosting shipments for its Taiwanese supply chain partners such as Wistron, Inventec, Quanta, and Foxconn.

The report from Commercial Times also cite sources, suggesting that this shift could provide Gigabyte, which is actively promoting its liquid-cooled products for NVIDIA’s H200 series, with  opportunities in the second half of the year.

Wistron, as a key supplier of motherboard and GPU accelerator cards for NVIDIA’s Hopper and Blackwell GPU, is not only a major supplier for Supermicro’s server motherboards but also for Dell. Its clients include HPE and Lenovo as well, which makes the company one of the primary beneficiaries.

Similarly, Inventec, one of the server motherboard suppliers, is also expected to benefit if the shift in orders boosts Dell, HPE, and Lenovo.

Moreover, one of Supermicro’s largest clients, CoreWeave, is transitioning to become a cloud computing service provider specializing in GPU-accelerated computing.

This shift has increased demand for GPU-accelerated computing and liquid cooling solutions. Reportedly, it’s believed that Gigabyte, which holds orders from CoreWeave, could be one of the biggest beneficiaries of the upcoming order shift.

Read more

• [News] Supermicro Faces Order Withdrawal Crisis, Potentially Benefiting Dell
• [News] Quick Couplings’ Buying Surge Reportedly Leads to AI Server Component Shortage Brought up by Supermicro

(Photo credit: Supermicro)

In late March, NVIDIA CEO Jensen Huang personally delivered the first high-end DGX H200 AI server to customer OpenAI. According to a report from Commercial Times, following this, the upstream chips for the H200 AI-GPU entered mass production in late Q2, with large-scale deliveries expected from Q3 onwards.

Downstream companies, including Inventec, Quanta (QCT), Hewlett Packard Enterprise (HPE), Supermicro, and Lenovo, have reportedly listed the H200 products as ready for shipment, with deliveries anticipated to begin in the second half of the year.

The same report, citing sources, indicates that current pending orders are still largely focused on the H100 in the HGX architecture, with the H200’s share remaining limited. The H200 shipments expected in Q3 will primarily be NVIDIA’s DGX H200. As for the B100, there is already some visibility, with shipments expected in the first half of next year.

Despite the CoWoS production capacity catching up and a significant easing in the supply of AI GPUs, due to the allocation system, the delivery time for the main H100 GPU shipments from various end-system partners can still reach up to 20 weeks.

However, major Taiwanese manufacturers such as Quanta, Inventec, Wistron, Gigabyte, and ASUS have seen a substantial boost in their overall server operations, driven by AI server business in the first half of the year. They are reportedly optimistic about AI server shipments in the second half of the year, expecting strong demand to continue, making the server business effectively free from any off-season throughout the year.

Among them, Wistron holds a significant advantage in the production and supply of H100 series substrates and the subsequent B100 series GPU modules and substrates. Starting from the second quarter, Wistron’s AI server-related business has shown high visibility, providing strong support for its overall operations.

On the other hand, the H20 series, an AI chip tailored for China due to U.S. chip restrictions, has also seen demand in the Chinese market. As NVIDIA recently secured a rush order for the H20 series, Taiwanese companies, including Wistron and Inventec, are expected to benefit.

Read more

• [News] Taiwanese AI Server Supply Chain to Benefit from Apple’s Potential Collaboration with Baidu and Alibaba
• [News] Rumored Sharp Drop in H100 Server Black Market Prices in China Raises Concerns Over Market Stability

(Photo credit: NVIDIA)

According to TrendForce research, the global high-performance computing market reached approximately US$36.8 billion in 2021, growing 7.1% compared to 2020. The United States is still the largest market for high-performance computing in the world with an approximate 48% share, followed by China and Europe, with a combined share of approximately 35%. Segregated into application markets, high-performance computing is most widely used in scientific research, national defense/government affairs, and commercial applications, with market shares of 15%, 25%, and 50%, respectively. In terms of product type, software (including services) and hardware account for 58% and 42% of the market, respectively.

Since high-performance computing can support data analysis, machine learning (ML), network security, scientific research, etc., it plays a key role in military fields such as nuclear warhead design and missile explosion simulations. Therefore, there are relatively few players occupying key positions in the value chain. Primary suppliers are Fujitsu, HPE, Lenovo, and IBM. These four manufacturers account for a market share of approximately 73.5% globally.

In addition, the continuous development of smart cities, smart transportation, self-driving cars, the metaverse, and space exploration and travel programs launched by Space X, Blue Origin, and Virgin Galactic will increase the demand for high-performance computing focused on R&D and testing along the two major axes of simulation and big data processing and analysis. The global high-performance computing market is expected to reach US$39.7 billion in 2022, with a growth rate of 7.3%. The CAGR (Compound Annual Growth Rate) of the global high-performance computing market from 2022 to 2027 will be 7.4%.

In view of this, the global high-performance computing market is growing steadily but not by much. The reason is that many of the aforementioned commercial application terminals are still in the growth stage, so high-performance computing technologies and solutions adopted by cloud service providers are limited to local deployment This enables HPC servers to scale on-premises or in the cloud and provides dedicated storage systems and software to drive innovation, thereby accelerating the development of hybrid HPC solutions.

In terms of end-use, the high-performance computing market is segmented into BFSI (Banking, Financial Services and Insurance), manufacturing, healthcare, retail, transportation, gaming, entertainment media, education & research, and government & defense. High-performance computing’s highest revenue share was derived from the government and defense market in 2021, primarily due to related agencies actively adopting cutting-edge and advanced IT solutions to improve computing efficiency. At present, government agencies in the United States, China, Japan, South Korea, as well as European countries have successively adopted high-performance computing systems to support digitization projects and contribute to economic development. Therefore, in 2021, the global scale of the on-premise high-performance computing server market was US$14.8 billion, of which Supercomputer, Divisional, Departmental, and Workgroup accounted for 46.6%, 18.9%, 25%, and 9.5% of the market, respectively. The global on-premise high-performance computing server market in 2022 is expected to reach US$16.7 billion with Supercomputer and Divisional growing by 11.5% and 15.2% compared with 2021.

（Image credit: Pixabay）

The new normal ushered in by the pandemic will not only become the driving force of digital transformation but will also continue to drive the server market in 2022, according to TrendForce’s investigations. It is worth noting that potential unmet demand in 2021 and the risk of future server component shortages will become medium and long-term variables that influence the market. Analyzing the shipment volume of completed servers, a growth rate of approximately 4-5% in completed server shipments is expected next year with primary shipment dynamics remaining concentrated in North American data centers with an annual growth rate of approximately 13-14%. From the supply chain perspective, the ODM Direct business model has gradually replaced the business model of the traditional server market, giving cloud service providers the ability to respond quickly to market changes. However, based on the unpredictability of the market, TrendForce assumes two forecasts for server growth trends. One, the supply situation of key components is effectively improved. Two, the supply situation of key components is exacerbated.

TrendForce states, based on the current situation as materials issues ease quarter by quarter, the annual growth rate of server shipments in 2022 will reach 4~5%. There are three primary factors driving market momentum. First, the introduction of the Intel Sapphire Rapids and AMD Genoa platforms into the market may once again stimulate the replacement of enterprise client servers and infrastructure construction in data centers. Second, the market generally believes that transformational needs generated by the pandemic in 2022, such as shifts in working paradigms and the new normal, will continue to drive the cloud market. Furthermore, international tensions have led to geopolitical uncertainty, which in turn has encouraged countries to tighten their control over data sovereignty and prompting the emergence of small-scale data centers in specific geographic locations.

Actual shipment volume of completed servers in 2022 depends on improvement of supply chain issues

Based on the two aforementioned assumptions, if the pandemic is effectively controlled next year, and international logistics, satisfaction of materials demand, and other factors either return to normal or fare better than expected, server companies will be able to increase their shipping capabilities and the annual growth rate of shipments in the overall server market will be able to reach 5-6% while the annual growth rate of ODM-Direct will approach 15%, up from the original forecast 13%. However, if the pandemic intensifies next year, the overall global economy will continue under that dark cloud which will greatly affect the willingness of companies to invest. In that case, the estimated annual growth rate of server shipments will fall to only 3-4%. In addition, the growth momentum of North American data centers will also be affected leading to an annual growth rate of ODM-Direct of only 10%, approximately.

As a whole and continuing under the influence of the two-year pandemic, the business trend of flexible deployment is irreversible. Regardless of overall economic changes, TrendForce expects double-digit growth in the demand for ODM-direct servers next year while overall server demand will also maintain a positive growth trajectory. However, continued attention should be focused on issues related to server order fulfillment in the broader market, including the fulfillment rate of key PMIC and LAN chip materials. At the same time, another major market variable will be whether Intel and AMD can introduce their two new platforms as scheduled next year and inject additional momentum into equipment replacement.

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

In light of rapid advancements in 5G, AI, cloud computing, and cloud services, TrendForce has observed that, since 2012, the widespread adoption of smartphones has brought about the emergence of certain new applications such as smartphone apps that leverage backend server support. Furthermore, enterprise cloud migration has noticeably accelerated after digital transformation activities began taking place in 2016. As a result, widespread data center deployments took center stage as the industry mainstream prior to 2020.

Moving into 2021, however, one of the many impacts of 5G commercialization on the market has to do with the regional deployment of infrastructures (i.e., small-scale data centers). This trend towards local deployment can be primarily attributed to the increasing stringency of personal data protection laws by governments worldwide that emphasize both the residency and the sovereignty of consumer data. Major public cloud companies are now beginning to establish cloud deployments closer to regions that represent sources of data generation, in order to deliver faster data analysis that is still in accordance with the law.

TrendForce’s observations on the build-out of data centers can be divided into the individual and enterprise aspects. With regards to the former, as connected devices become increasingly widespread and emerging content services more popular in the post-pandemic stay-at-home economy, consumer spending on streaming videos and online shopping began to experience a rapid spike, thereby contributing to a gradual increase in server build-outs for cloud services. Enterprises, on the other hand, have been seeking and deploying more flexible infrastructures in response to pandemic-induced uncertainties. Hence, certainly subscription services and hybrid cloud services have also been gaining momentum in the ongoing trend of cloud migration.

It should be noted that, due to recent geopolitical tensions, which intimately dictate the development of various countries’ industries and supply chains, global enterprises also face rapidly changing market needs as well as a high degree of uncertainties stemming from the pandemic. In turn, enterprise demand for cloud services has also seen a continued rise in the past two years. With respect to the adoption of AI and other emerging technologies, most enterprises prefer cloud services due to such services’ flexible cost structures. TrendForce’s latest investigations indicate that flexible pricing strategies and diverse services offered by data centers have directly propelled the demand from enterprises for cloud applications in the past two years. From the perspective of the server supply chain, these shifts have facilitated a gradual shift of the predominant business model in the server market from traditional server brands to ODM Direct.

In addition, data center-related technologies have also progressed significantly. As the way people work and live transforms, accompanied by the emergence of e-commerce and streaming media, enterprises have also become increasingly well-versed in cloud services and increasingly able to leverage related technologies. As such, the primary sources of competition in the cloud market will include not only infrastructures responsible for computing, storage, and networking, but also emerging technologies such as edge computing and software/hardware integration of related services by major operators. In particular, as 5G services successively kick into gear worldwide, the concept of “cloud-edge-local network” will begin to replace the current “cloud only” framework on a massive scale, thereby extending the relevant commercial opportunities from cloud services to hardware vendors. That is to say, in the future, cloud services will no longer be limited to the software front, as in-house hardware brands from CSPs are set to become the next battlefield while these companies compete to offer comprehensive services.

All of this raises the question of whether the build-out of data centers will involve more challenges and opportunities going forward. TrendForce believes that, in addition to factors such as telecommuting and e-commerce, data center demand from biomedical applications (for instance, the ramp-up of vaccinations) will also experience substantial growth, with the caveat that regulations governing the protection and collection of medical data will be even more stringent than those driving various countries’ data sovereignty endeavors. Hence, privacy and security pertaining to medical data will likely become not only a global pursuit, but also a significant challenge facing the application of data centers.

（Image credit: Unsplash）