Analyst Jay Goldberg: Advanced Micro Devices and Broadcom could become the new favorites in AI chip investments, while Nvidia is facing pressure on market share and valuation.

TaiwanBusiness05/01 06:00
Analyst Jay Goldberg: Advanced Micro Devices and Broadcom could become the new favorites in AI chip investments, while Nvidia is facing pressure on market share and valuation.

Seaport Research Partners analyst Jay Goldberg has rated NVIDIA as a "sell" with a target price of $100 amidst a booming AI chip market, noting that its valuation already aligns with market expectations and faces pressure from customers creating their own chips. On the other hand, he is optimistic about AMD and Broadcom, giving them "buy" ratings with target prices of $100 and $230, respectively, due to their strategic moves and strong execution in the AI and data center markets, potentially making them more attractive investments.

Key Updates

05/01 06:00

Analyst Jay Goldberg: Advanced Micro Devices and Broadcom could become the new favorites in AI chip investments, while Nvidia is facing pressure on market share and valuation.

Seaport Research Partners analyst Jay Goldberg has rated NVIDIA as a "sell" with a target price of $100 amidst a booming AI chip market, noting that its valuation already aligns with market expectations and faces pressure from customers creating their own chips. On the other hand, he is optimistic about AMD and Broadcom, giving them "buy" ratings with target prices of $100 and $230, respectively, due to their strategic moves and strong execution in the AI and data center markets, potentially making them more attractive investments.

Nvidia: Valuation Already Reflects Positives, Faces Challenges from In-House Chip Development

As a leader in the AI chip market, Nvidia has long benefited from the explosive demand for generative AI and large language models. However, Seaport analyst Goldberg points out that Nvidia's stock price has fully reflected its leading position in the AI boom and warns that future demand may decrease. He rates Nvidia's stock as a "sell," setting a target price of $100, significantly lower than the current market price.

Goldberg specifically notes that the high deployment cost of Nvidia's high-end GPUs and the increasing demands from customers for returns on AI investments may limit its future growth potential. Additionally, as hyperscalers like Meta and Google accelerate the development of their own AI chips, Nvidia's market share is facing erosion. For example, Meta's recently launched MTIA chip has begun replacing Nvidia GPUs in some servers and plans to expand to AI training uses.

Broadcom: AI Custom Chip Driver, Potential Not Fully Valued by the Market

Unlike Nvidia's high profile, Broadcom's role in the AI field is more low-key but crucial. Goldberg rates Broadcom as a "buy," with a target price of $230, believing that its ability to assist tech giants in developing custom chips is the core driver of its future growth.

Broadcom currently provides custom chip design and manufacturing services for tech giants like Google, Apple, and Meta, with these chips mostly used for AI inference and data center computing. Goldberg points out that when Nvidia loses market share, it is often because these customers choose to design their own chips, and Broadcom is the biggest beneficiary of this trend.

Moreover, Broadcom's business structure is diverse, including stable revenue sources such as network chips, storage controllers, and enterprise software, in addition to AI chips. This gives it stronger risk resistance amid AI market fluctuations. Although Broadcom's stock price fell after reaching a $1 trillion market cap in early 2025, analysts believe its AI potential is not yet fully reflected in its stock price.

Advanced Micro Devices (AMD): Steady Execution and Data Center Market Share Growth Are Key

Advanced Micro Devices (AMD) also received a "buy" rating from Seaport, with a target price of $100. Goldberg states that although AMD's market share in the AI chip market is still far behind Nvidia, its continuous encroachment on Intel in the data center and PC markets demonstrates its execution strength and product competitiveness.

AMD is expected to launch the MI350 GPU series in the second half of 2025, which is anticipated to improve AI inference performance by 35 times compared to the previous generation. Although this series has not yet been released, Goldberg believes its product roadmap is clear and execution progress is steady, making it a key focus for investors.

Additionally, AMD is gradually expanding its collaboration with cloud service providers and continues to capture Intel's market share in the server CPU market. Seaport points out that although AMD's AI business is still in its early stages, its overall business performance is stable, with long-term growth potential.

AI Chip Market Competition Intensifies, In-House Development Becomes New Trend

Goldberg's report also reflects that the AI chip market is entering a new phase. In recent years, Nvidia has dominated the market due to its GPUs' strong performance in AI training and inference. However, as AI applications shift from training to inference and companies demand more in terms of cost and performance, more tech companies are choosing to develop their own chips to reduce reliance on a single supplier.

This trend poses a challenge to Nvidia but also creates opportunities for companies like Broadcom that provide custom chip solutions. Meanwhile, companies like AMD, with a complete product line and stable execution, are also poised to benefit from the expansion of data centers and AI applications.

References

People Also Ask...