Goldman Sachs Favors AMD Over Nvidia for AI Inference Market Growth

Goldman Sachs analyst James Schneider believes the inference market in AI will significantly increase over the next five years, driving a "barbell" approach to semiconductor investments. Schneider recommends Nvidia (NASDAQ:NVDA) over Advanced Micro Devices (NASDAQ:AMD) due to Nvidia's control of the full AI ecosystem, including advanced chips, networking hardware, and CUDA software. Nvidia's dominance in the AI market has contributed to its status as the world's most valuable company.

Goldman Sachs analyst James Schneider recently predicted a significant increase in the AI inference market over the next five years, leading to a "barbell" approach to semiconductor investments. Schneider recommends Nvidia (NASDAQ:NVDA) over Advanced Micro Devices (NASDAQ:AMD) due to Nvidia's control of the full AI ecosystem, including advanced chips, networking hardware, and CUDA software [1].

Nvidia, the dominant player in AI infrastructure, has seen its market capitalization reach $4 trillion, reflecting its near-monopoly in the AI chip market. Its H100 and A100 GPUs are the gold standard for training large language models, with Microsoft alone investing $10 billion to equip its Azure cloud with Nvidia's infrastructure [2].

Advanced Micro Devices (AMD), while still a distant second to Nvidia in GPUs, has carved out a niche in AI inference. AMD's ROCm software platform, while not as advanced as CUDA, is considered good enough for inference tasks. AMD's AI opportunity could just be getting started, with major cloud providers using its GPUs for search, recommendation engines, and generative AI tasks [1].

Broadcom, another contender, is focusing on AI networking and custom AI chips. Its AI networking revenue surged 70% last quarter, driven by the increasing demand for networking components in large AI clusters. Broadcom's custom AI chips can deliver better performance and power efficiency than off-the-shelf GPUs, with several hyperscalers planning to deploy millions of chip clusters by 2027 [1].

Nvidia's dominance in the AI market has contributed to its status as the world's most valuable company. However, its size could limit some of its upside. As such, Schneider believes Nvidia is well-positioned to outperform AMD and Broadcom in the coming years, given its control of the full AI ecosystem [1].

References:
[1] https://www.nasdaq.com/articles/prediction-these-2-ai-chip-stocks-will-outperform-nvidia-over-next-5-years
[2] https://www.ainvest.com/news/nvidia-4-trillion-valuation-ai-era-growth-engine-2507/