The resignation of NVIDIA CEO Jensen Huang on the 12th shocked the global technology industry and also affected the changes in the AI chip industry landscape.
[Reporter Lin Yiru/Taipei Report] NVIDIA CEO Jen-Hsun Huang resigned on the 12th, shocking the global technology industry. Counterpoint Research, a research organization, put forward six key points of observation. It believes that NVIDIA's existing AI chip product route and supply chain cooperation are expected to remain stable in the short term. However, in the medium and long term, it is still necessary to observe the strategic direction of the new management team, the supply chain cooperation model, and the development of customized AI chips. Counterpoint predicts that server AI ASIC shipments will surpass GPUs in 2027. As large cloud service providers (CSPs) continue to develop their own chips, the AI computing market will develop in parallel with GPUs and customized ASICs.
Counterpoint pointed out that NVIDIA’s product planning from Blackwell to Vera Rubin has been gradually established, and the existing R&D team and supply chain cooperation relationships are expected to support product advancement in the next 18 to 24 months. The current NVIDIA ecosystem covers TSMC's advanced processes and CoWoS packaging, SK Hynix and Micron HBM memory, as well as server manufacturers such as Hon Hai, Quanta, and Wistron. Therefore, short-term product delivery and supply chain cooperation are not expected to be immediately affected by changes in management. However, if the pace of future product updates is adjusted, it will still affect suppliers' capacity planning, R&D investment, and capital expenditures.
Counterpoint further analyzed that NVIDIA's competitive advantages are no longer limited to GPUs, but extend to multiple levels such as energy, chips, infrastructure, AI models and applications, and form integrated advantages through the CUDA software ecosystem. As the company enters a new management stage, whether the new team can continue its cross-technology platform integration capabilities and maintain its sovereign AI and cloud infrastructure partner investment plans will be important factors affecting future growth. In addition, NVIDIA has adopted a highly flat management structure in the past. After the redistribution of organizational responsibilities in the future, the retention of core technical talents, the stability of the management team and the efficiency of cross-department decision-making will also become the focus of observation.
On the other hand, the development of customized AI chips may also accelerate. As large cloud players such as Google TPU and AWS Trainium continue to develop their own AI chips, the AI computing market is gradually forming a parallel pattern of GPUs and ASICs. Brady Wang, deputy director of research at Counterpoint, predicts that server AI ASIC shipments will surpass GPUs in 2027, and changes in NVIDIA's leadership may also prompt some customers to re-evaluate their supplier portfolio. However, actual purchasing decisions still depend on computing performance, cost, software ecosystem and return on investment. In addition, power supply and data center construction progress will continue to affect AI infrastructure investment and deployment.
Neil Shah, co-founder and vice president of research at Counterpoint Research, believes that NVIDIA’s leadership change has brought new observation variables to the AI industry, but it does not mean that existing technology development and market demand will immediately change. In the short term, product delivery, supply chain cooperation and customer demand will still be the main driving factors. In the medium and long term, we must pay attention to the new management team strategy, AI chip supplier competition, customized ASIC penetration, and the evolution of global regulatory policies. As the AI industry shifts from rapid expansion to placing more emphasis on large-scale deployment, cost efficiency and investment returns, companies' technology integration capabilities and operational execution efficiency will become the key to competition in the next stage.
