Missing Out NVIDIA in AI Trends? AMD's Growing Prominence is Generating More and More Attention

TSMC's Q4 2023 earnings results showed that:

1) Its revenue for the quarter reached approximately $20.08 billion, largely in line with the historical high achieved in the same period of the previous year. This was driven by the continued surge in demand for AI chips, and surpassed the estimated range of $18.8 billion to $19.6 billion provided by TSMC's executives. Net profit data also exceeded market expectations.

2) In addition, the world's leading semiconductor foundry announced a capital expenditure plan for 2024 that exceeds expectations, which is estimated to be between $28 billion to $32 billion, demonstrating the company's confidence in the industry's fundamentals.

3) More importantly, TSMC's optimistic outlook for AI chips has resulted in an encouraging 2024 performance guidance, with a projected revenue growth of approximately 20% to 25%. Except for Q1, revenue for every remaining quarter is expected to increase.

As the leading semiconductor foundry, TSMC's impressive performance and positive outlook have given investors who are chasing AI chip stocks more confidence in both the industry's prospects and the strong demand for AI chips. According to a research report by Goldman Sachs published on Thursday, TSMC's executives believe that the manufacturing of AI chips could account for "over 10%" of the company's total revenue within the next five years, which will have a positive impact on both Nvidia and AMD. Mizuho desk-based analyst Jordan Klein wrote on Thursday,

In December of last year, AMD announced the release of two new AI chips in its MI300 series, claiming that their performance outstrips Nvidia's H100 and have already secured orders from big names such as Meta, Microsoft, Oracle, as well as server suppliers like Dell, HP, Lenovo, and Supermicro. Specifically, the APU (Accelerated Processing Units) product - MI300A - which integrates CPU+GPU, is almost equal to the H100 in AI computing power, and has significant advantages in high-performance computing, memory capacity, and bandwidth. Meanwhile, the pure GPU-designed product - MI300X - has certain advantages over the H100 in all of the aforementioned areas.

The new GPU is expected to help the company expand its market share and drive revenue growth in 2024. According to the company, MI300 chip revenue for the quarter is expected to reach $400 million, and is projected to exceed $2 billion for the full year in 2024. The company predicts that the MI300 chip will become its fastest-selling product ever, reaching $1 billion in sales. At the same time, analysts have pointed out that the company has already begun building an AI software ecosystem for AMD, and that the MI300 chip and its supporting software system are among the few products that can compete directly with Nvidia's chips and systems, highlighting their potential for growth and competitiveness.

Bloomberg data shows that despite frequent upward revisions of AMD's target price by analysts over the past month, the company's stock price has still outpaced analysts' expectations. According to Bloomberg data tracking the stock over the past month, more than 80% of the 21 analysts covering the stock have given it a "buy" or similar ratinng, with an average target price of $165. The highest target price comes from Warren Lau at Aletheia Capital, who has set a target price of $205.

Barclays' Tom O'Malley and team have increased their price target for AMD from $120 to $200. They stated, "With supply constraints, customers are often using the entire NVDA platform in order to get priority shipments of accelerators". As other chip manufacturers such as AMD gain market share, 2024 will be the "the year AI begins to open up" . They further added that a second source of supply is needed for the software ecosystem to overcome current difficulties.

KeyBanc analyst raised their price target from $170 to $195. Cowen & Co. analyst Matthew Ramsay also raised his price target from $130 to $185. Ramsay said that the company's MI300 chip is focused on AI applications and is "poised for material growth/upside".