TMTPOST -- The latest technology collaboration between Nvidia Corporation and Intel Corporation will not pose any threat to Taiwan Semiconductor Manufacturing Co. (TSMC)’s chip tech leadership, so artificial intelligence (AI) chips of the world’s top contract chip manufacturer will not be affected, suggested TF International Securities analyst Ming-Chi Kuo on Thursday.
Credit：TSMC
Kuo commented TSMC is likely to maintain its advantage at advanced nodes at least through 2030, and the Nvidia-Intel partnership is unlikely to change TSMC’s leading-edge position. As AI chips requires leading-edge advanced nodes, TSMC’s AI-chip orders should remain intact, said Kuo in a social media post.
Kuo acknowledged Nvidia’s investment in Intel could shift market share among competitors of both companies, such as Advanced Micro Devices, Inc. (AMD), which would in turn affect TSMC’s orders. But he noted both Nvidia and Intel are expected to remain key customers of TSMC, and the networking products using less advanced nodes contribute a smaller portion of its revenue. Therefore, he believed the overall risk for TSMC remains manageable.
Nvidia on Thursday announced it will invest $5 billion in Intel’s common stock, and two companies will also jointly develop multiple generations of custom data center and personal computer (PC) products, focusing on connecting both companies’ architectures using Nvidia NVLink, the high-speed interconnect for graphics processing unit (GPU) and CPU processors in accelerated systems.
Nvidia said it will collaborate with the chipmaker to create AI systems for data centers that combine Intel’s x86-based central processors with Nvidia’s GPUs and networking. Intel will build and deliver x86 system-on-chips (SOCs) that integrate Nvidia RTX GPU chiplets.
Kuo said the partnership could define and accelerate the AI PC Landscape. For Nvidia, developing its own Windows-on-ARM processors carries high uncertainty, while for Intel, establishing a competitive edge in GPUs is difficult. Given these facts, Kuo said teaming up (CPU + GPU) could create powerful synergies and advantages across the PC ecosystem.
The biggest question is if Nvidia will manufacture at Intel’s fabrications, Wolfe Research anlayst Chris Caso commented following the announcement of collaboration. Nvidia CEO Jensen Huang at a press conference on Thursday decline to answer if Nvidia will use Intel’s foundry, stating the announcement is about CPUs, not exactly how these chips will be manufactures.
The custom Nvidia CPUs made using Intel’s designs will largely be fabricated by TSMC using its silicon wafers, which will then be sent to Intel’s foundry to add key components known as packaging, which Intel is expected to make, the Wall Street Journal reported, citing people familiar with the matter. The report said what the Thursday’s deal doesn"t significantly address is Intel’s to become a world-leading contract chip maker—called a foundry in the industry.
