Despite the recent adjustments experienced by AI industry leaders like Nvidia amid the impact of DeepSeek, Morgan Stanley believes that DeepSeek, along with the Stargate project, will usher in a new wave of demand for the AI chip supply chain.

In their latest research report, Morgan Stanley analysts Charlie Chan, Erik W. Woodring, and others stated that with the global surge in demand for AI technology, particularly in China, key components and market players in the AI supply chain are poised for new growth opportunities.

The report predicts that the AI supply chain will undergo significant transformations as the Stargate project and China's demand for high-performance computing (HPC) continue to grow. It emphasizes that global cloud capital expenditure (capex) remains strong, indicating robust demand for AI semiconductors/hardware.

Morgan Stanley has also raised its 2025 cloud infrastructure spending growth forecast from 29% year-on-year to 32%.

Additionally, the analysts are particularly optimistic about the applications of GPUs and ASICs in the AI semiconductor sector, especially in the competition between Nvidia and AMD, with a preference for Nvidia.

The report also highlights that as AI technology advances, related supply chain components such as Blackwell servers and Hopper GPUs will experience new peaks in demand. These predictions are based not only on current market trends but also on technological advancements and changes in the global economic environment.

Stargate: A Capital Expenditure Revolution in AI

The Stargate project, a new driver of global AI capital expenditure, is attracting widespread attention in the industry. The project is expected to attract up to $500 billion in investments over the next four years, with $100 billion already deployed. This massive investment scale signals a surge in demand for AI hardware, particularly in the high-performance computing (HPC) sector.

Morgan Stanley analysts noted this project could imply a rapidly growing market demand for high-performance AI hardware. The Stargate project is not only driving demand for Nvidia and its ODM partners but also injecting new vitality into the entire AI supply chain. The successful implementation of this project could become a significant milestone in the AI field, accelerating the development of related technologies and applications.

DeepSeek: A New Engine for AI Demand in China

DeepSeek applications in China are becoming a new engine driving the growth of AI demand. Morgan Stanley's report emphasizes that thanks to DeepSeek, demand for China's H20 GPUs is rebounding.

Analysts learned from the Taiwanese semiconductor supply chain that approximately 750,000 units of Hopper chips are being tested by KYEC in the first quarter of 2025, with the majority being H20. This trend indicates that the demand for AI technology in China is growing rapidly, presenting new opportunities for the global AI supply chain.

The promotion of DeepSeek applications is not only accelerating the development of AI technology in China but also providing new momentum for the global AI supply chain. As China continues to advance in the AI field, its demand for high-performance computing is expected to grow steadily, further driving the development of the global AI supply chain.

The report suggests that this demand growth in China could have a profound impact on the global AI hardware market, prompting strategic adjustments and resource allocation among related companies and supply chain participants.

Synergistic Effects of Stargate and DeepSeek

The report points out that the combination of the Stargate project and DeepSeek applications heralds a new phase of development for the AI supply chain. The synergistic effects of these two factors could accelerate the global adoption and application of AI technology, driving innovation and market expansion among related companies and supply chain participants.

The successful implementation of the Stargate project and DeepSeek applications may trigger a new wave of investment in the AI sector, attracting more capital and resources into the research and development of AI technologies and applications. This will not only propel the rapid development of AI technology but also bring new opportunities and challenges to the global AI supply chain.

Technological Challenges and Market Opportunities

Despite the optimistic market outlook, the report also highlights technological challenges, particularly the assembly issues with Blackwell NVL72. Analysts mentioned that due to ongoing NVL72 assembly bottlenecks, more B200 chips are being used in HGX (8-GPU) servers. This challenge could impact supply chain efficiency and costs, but analysts believe that with technological advancements, this issue will be resolved.

In terms of market opportunities, the report discusses the competition between AI GPUs and ASICs. Analysts noted that while ASICs may have advantages in certain applications, GPUs still hold a significant position in the AI field. As technology continues to evolve and market demands shift, the competition between AI GPUs and ASICs will intensify, creating new opportunities for related companies and supply chain participants.