Goldman Sachs evaluates the AI chip industry chain: NVIDIA and three other companies rated as buy, adopting a "barbell" strategy to layout the semiconductor sector
Goldman Sachs released a research report covering several U.S. digital semiconductor and EDA software companies for the first time, pointing out significant investment opportunities in AI-related capital expenditures. Analysts rated NVIDIA, Broadcom, Cadence Design Systems, and Synopsys as "buy," with target prices of $185, $315, $380, and $620, respectively. The report emphasizes that AI infrastructure capital expenditures have surpassed $350 billion. Although it will take time to realize profits, early signs of revenue growth have supported continued investment. Analysts believe the semiconductor market is undergoing rapid restructuring, and a "barbell" investment strategy will focus on performance and cost advantages
According to the Zhitong Finance APP, Goldman Sachs recently released a research report covering several American digital semiconductor and electronic design automation (EDA) software companies for the first time, pointing out significant investment opportunities among commercial and custom chips and EDA suppliers with the most sustainable capital expenditures related to artificial intelligence.
The analyst team, led by James Schneider, gave buy ratings to four companies: NVIDIA (NVDA.US) with a target price of $185, Broadcom (AVGO.US) with a target price of $315, Cadence Design Systems (CDNS.US) with a target price of $380, and Synopsys with a target price of $620. At the same time, they maintained a neutral rating for AMD (AMD.US) with a target price of $140, Arm (ARM.US) with a target price of $160, and Marvell Technology (MRVL.US) with a target price of $75.
The report emphasizes that current capital expenditures on artificial intelligence infrastructure have exceeded $350 billion. Although profitability will take time to realize, early signs of revenue growth and progress in cost optimization have provided support for sustained investment. Analysts believe that the artificial intelligence investment cycle is at a critical transformation period, and the semiconductor market and technology leadership landscape are undergoing rapid restructuring.
The balance of demand for cutting-edge model training and low-cost inference will give rise to a "barbell" investment strategy: focusing on leading companies in performance and software ecosystems while also paying attention to segment leaders with cost advantages.
Specifically in the chip design segment, as the industry evolves from traditional client/server architectures to cloud computing and the generative AI era, system complexity and multi-chip integration demands have significantly increased, continuously expanding the pool of chip designers and customers, which directly boosts the value of EDA software.
Analysts point out that in this round of technological transformation, while the custom chip sector is still in its early development stage with limited market share, economies of scale will help leading companies gain an advantage; however, the commercialization process will still take time, and performance and ecosystem breadth will remain core premium factors.
The report particularly mentions geopolitical impacts, believing that technology-related risks remain a core variable in the semiconductor industry, but sovereign infrastructure demands from China and other regions are creating new opportunities, partially hedging against geopolitical uncertainties. The Goldman Sachs team maintains an optimistic outlook on the long-term development prospects of artificial intelligence, believing that the current stage of technological iteration and capital investment will create sustained growth space for companies with technological barriers and ecosystem advantages
