The emergence of generative AI, the arrival of the 10x speed AI generation, and the soaring demand for AI servers have propelled NVIDIA to seize the dominant position in the global AI landscape with its leading technology and economic scale advantage in combination with AI GPUs and powerful CUDA software technology. Consequently, NVIDIA further strengthens its control over the development of the server hardware industry chain. With all manufacturers vying to cooperate and join the supply chain list, the three words "seeking to soar" succinctly summarize the current competitive situation in the AI server market.
In early 2023, the explosive growth of generative AI applications, led by ChatGPT, enabled NVIDIA, which has invested heavily in AI R&D for many years, to secure global dominance in the AI landscape with its powerful GPU for parallel computing and high-speed transmission, as well as the CUDA software technology that accelerates GPU computing.
NVIDIA holds over 90% market share in AI GPUs, and its quarterly financial report as of the end of April 2024 once again astonished the market, with revenue reaching 26𝑏𝑖𝑙𝑙𝑖𝑜𝑛,𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑖𝑛𝑔𝑎𝑛1828 billion.
The server supply chain has expressed that the anticipated boom in electric vehicle opportunities has not materialized, and without AI servers taking over, the supply chain, which has relied for years on large orders released by international giants for PCs, mobile phones, and other products, can expect dull business operations in the coming years. NVIDIA CEO Jensen Huang's return to Taiwan to attend COMPUTEX is a parallel repetition of the 2023 performance, even more impassioned and practical.
Behind the scenes of Jensen Huang dining with Taiwanese manufacturers lies a deeper strategic significance: to solidify a comprehensive stakeholder strategy, ensuring the loyalty of the world's largest AI arsenal and expanding support from developers, businesses, and governments. Huang's sincerity is evidenced by the constantly updated guest list and repeated gratitude to partners, directly confirming the authenticity of the "true AI" industry chain.
While having dinner with Jensen Huang may not have been a significant event in the past, it has become a focal point of attention in the current pursuit of the "friends of Jensen Huang" list. The AI strategies and prospects of companies like Compal and Acer have also come under intensified scrutiny.
For the supply chain, being in the same frame as Jensen Huang or being included in its top 100 supply chain list reflects a sense of momentum, as being in the list guarantees orders. As the Chinese drama "In the Name of the People" once said, "You think you are eating lobster, but in fact, you are eating opportunities."
NVIDIA currently holds over 90% market share in AI GPUs, and the road ahead for competitors such as AMD, Intel, Meta, Google, Microsoft, and AWS, which have heavily invested in self-developed chips, will be arduous. Even within the next few years, it will be difficult for them to break free from their reliance on NVIDIA.
Taiwanese manufacturers are acutely aware of the industry's reality. At this stage, it is not about socializing, but about seeking opportunities. It is crucial to enter any cooperation list, as they hope to be "taken to new heights." At present, they are not concerned about NVIDIA's calculations.
Jensen Huang, dubbed "the trillion-dollar man" in 2023, saw his net worth skyrocket to three trillion dollars in 2024. During his speech in Taiwan in early June, he repeatedly thanked the Taiwanese supply chain, praising, "Without Taiwan, NVIDIA's vision would only be an unattainable dream."
Hardware manufacturing, from PCs to smartphones, is a strong suit of Taiwanese companies. They understand their role in the AI revolution, not as the leading actors, but as indispensable supporting characters. With NVIDIA becoming the world's largest AI arms dealer, the semiconductor, OEM, ODM, and component industries in Taiwan are following in NVIDIA's footsteps, establishing themselves as the most powerful arms depot.
As the delivery period for the H100 shortens, NVIDIA has begun to expand its supply. The introduction of the AI chips B100 and B200, and the GB200 with Grace CPU, built on the Blackwell architecture, will see TSMC, the exclusive supplier of NVIDIA's large orders, and SK Hynix, the largest supplier of HBM, experience even more fervent business performance in the second half of the year.
It is worth noting that TSMC is the most powerful force in NVIDIA's arsenal in Taiwan. Many semiconductor supply chain companies have been working with NVIDIA for many years. For example, the top load semiconductor testing interface manufacturers such as ASE and WNC are ramping up production for NVIDIA's expansion plans, and KYEC's share in NVIDIA's chip testing supply has also been steadily increasing. The Taiwanese semiconductor supply chain can expect significant growth in performance.
 

Managing Editor：朱原弘