Nvidia (NASDAQ:NVDA) has dominated the generative artificial intelligence sector, with its stock price soaring 180% this year. This impressive leap has positioned it as the world’s most valuable company, boasting a market cap of nearly $3.5 trillion. Yet, trading at a steep valuation of around 48 times the anticipated FY2025 earnings, future gains for Nvidia may be challenging. On the other hand, AMD offers a more enticing opportunity. With its stock (NASDAQ:AMD) trading at a more reasonable 28 times forward earnings, AMD stands to benefit from the ongoing growth in AI while providing better value to investors. As the AI revolution unfolds, selecting the right companies will be crucial for continued profitability, suggesting that it may be time to reassess Nvidia and consider AMD’s potential. The appeal lies not just in AMD’s valuation, but also in trends indicating a shift toward cost-effectiveness and changes in model training processes that could favor AMD.
AMD’s Performance Over Time
Over the last four years, AMD stock has occasionally performed well, climbing from around $90 at the start of 2021 to peaks exceeding $200 earlier this year. The path has not been smooth, with annual returns fluctuating significantly compared to the S&P 500. AMD saw returns of 57% in 2021, dropped by 55% in 2022, and rebounded with 128% in 2023. In contrast, the Trefis High Quality Portfolio, featuring 30 carefully selected stocks, has outperformed the S&P 500 each year in that same period, illustrating a more stable growth pattern.
Nvidia’s AI Training Challenges
Significant investments have been poured into AI model development, which are increasingly complex. Training these expansive models requires substantial computing power, and Nvidia has capitalized immensely as its GPUs are considered the most efficient for this purpose. Its revenue trajectory is impressive, projected to jump from $27 billion in FY’23 to nearly $130 billion in FY’25. However, as these models grow, improvements in performance may not keep pace with increased investments, leading customers to reassess their spending plans.
Additionally, the accessibility of high-quality data for training large language models could become a bottleneck. Much of the desirable content is already ingrained in existing AI models, limiting opportunities for further enhancement. This may prompt a transition toward smaller, specialized models tailored for specific tasks, potentially reducing demand for Nvidia’s advanced GPUs and suggesting that the rapid growth seen by Nvidia may have peaked.
A Shift Towards Inference Could Favor AMD
As AI development evolves, the market may gradually pivot from model training to inference—the process of generating outputs from trained models. This part of AI is less demanding on computational resources, potentially allowing AMD to compete effectively. Nvidia reports that around 40% of its data center revenue comes from inference, but increasing competition may challenge this segment. MLCommons has found that AMD’s MI300X is quite competitive against Nvidia’s H100 GPU in inference benchmarks, resulting in AMD winning significant orders from hyperscalers for AI inference workloads. For instance, IBM plans to include AMD’s MI300X AI chip in its Cloud services in the first half of 2025.
Cost Pressure on Nvidia’s Margins
Economics also play a role, as customers of AI chips are becoming increasingly price conscious. The initial AI boom led enterprises to invest heavily due to apprehension about missing out, which bolstered Nvidia’s pricing power, achieving net margins over 50% recently. However, as the focus shifts toward returns, businesses will seek efficient spending on AI, possibly benefiting AMD and emerging competitors like Intel that offer more affordable alternatives. Nvidia’s GPUs, priced above $25,000 each, are prompting customers to explore cheaper options. For example, Oracle recently chose AMD’s accelerated computing chips for its supercluster designed for high-performance AI tasks, following tests that confirmed the competitiveness of AMD’s GPUs.
AMD’s Competitive Valuation
As previously highlighted, AMD’s stock trades at approximately 28 times projected FY’25 earnings, compared to Nvidia’s steep 48 times multiple. This differential makes AMD a more sensible investment, particularly given its potential for earnings growth driven by increased CPU market share and the broader expansion of the AI sector. AMD may be growing more slowly than Nvidia, but its performance remains impressive where it counts. In Q3, AMD exceeded expected earnings and revenues, with its data center business reporting $3.5 billion in sales, marking a 118% annual increase. Such developments position AMD favorably in the current AI landscape.
| Returns | Nov 2024 MTD [1] |
2024 YTD [1] |
2017-24 Total [2] |
| AMD Return | -5% | -7% | 1112% |
| S&P 500 Return | 5% | 26% | 167% |
| Trefis Reinforced Value Portfolio | 9% | 25% | 828% |
[1] Returns as of 11/27/2024
[2] Cumulative total returns since the end of 2016
Invest with Trefis Market-Beating Portfolios
See all Trefis Price Estimates
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
