Summary

• AMD's stock price dropped 35% last year, but its value aligns with the current price; hold the stock for AI growth potential.

• AMD benefits from Intel's strategic challenges and Nvidia's market creation, positioning itself as a low-cost producer in the Data Center segment.

• Margins are pressured by AI competition and R&D investments, but AMD's flexible fabless strategy and market share gains over Intel offer long-term potential.

• ARM technology poses a risk, but AMD's x86 focus and potential ARM adoption provide adaptability in a shifting market landscape.

Overview

Advanced Micro Devices, Inc. is one of the most prominent semiconductor companies in the world. It has been the main competitor of Intel (INTC) in the desktop and laptop market; it has been the leading provider of microchips in the gaming consoles Xbox from Microsoft and PSP from Sony. Lately, it is the only company competing with Nvidia (NVDA) in the hot GPU for AI market. As you can see in Figure 1, the stock price has decreased by 35% in the last 12 months.

Figure 1: Seeking Alpha

AMD has always been a technological follower and a lower-price competitor. This has happened in the desktop and laptop market chasing Intel, which is happening right now with Nvidia and the Data Center GPU market. The company is going to benefit from AI euphoria and take advantage of the new wave of AI PCs, but it is exposed to relevant risks. However, technological competition in those markets is intensifying and is stressing margins.

Like most competitors in the industry, the company is getting high valuation multiples that I consider out of reach in relation to its business fundamentals.

A clear Data Center opportunity

The two most lucrative markets for AMD are the Data Center and the Client segment. Both markets have their challenges, but in general, they are compelling big business opportunities. The Data Center segment is dominated by NVIDIA, and here, the risk is to invest high but be an insignificant player. This segment grew 94% year over year to $12.6 billion in 2024, driven by EPYC CPUs and Instint GPUs. However, Nvidia remains the dominant player in AI GPUs, with an 80% market share. Nvidia’s dominance is related to its leadership in GPU technology and CUDA software, as the company is trying to create an AI platform with network effects.

Nvidia has opened a very lucrative new market, and for me, it is not a competitor. I see it as a facilitator that is opening a vast opportunity for AMD: I think AMD will gain market share in this expansive market as a low-cost producer, especially when requirements are not so demanding as to necessitate Nvidia’s GPU. Sometimes, the opportunity will come from a lack of availability in Nvidia’s inventory.

The only companies that can overshadow AMD’s performance are those that personalize chips for hyperscalers: Broadcom (AVGO) or Marvell Technology (MRVL). I don’t think it will hurt much of AMD’s business. Broadcom GPU business in Q4 was $300 million, and I estimate AMD business in the last quarter is $800 million. Personalized business doesn’t scale and also lacks a common software platform, like ROCm.

Anyway, growth rates will be high, around 40%, a little bit lower than management has stated in the mid-term. This segment has the highest margin, is double the Client and Gaming segment, and is growing the fastest (93.6% vs 51.7% in the Client segment). However, this explosive growth won’t last forever; I estimate that the high AI growth will last four or five more years. Later, it will moderate its growth.

Taking advantage of Intel’s weakness

The other segment, Client, is dominated by desktops and laptops, and the main competitor has been Intel, see Figure 2. However, Intel is facing some strategic challenges. Once an integrated semiconductor player dominates the design and manufacturing of the CPU market with the x86 architecture, the manufacturing capabilities are now a drag on the design, and it is unable to adapt to market requirements. AMD is a design company that is gaining market share without the constraints of manufacturing facilities.

Figure 2: Mercury Research

Intel's weakness is allowing AMD to increase prices, positioning it as an innovation leader in the market. The price of mobile and desktop Ryzen processors will increase by 14% in 2024.

However, AMD is facing competition from other players like Qualcomm, which supports ARM’s architecture, which is the design win for energy-efficient and smaller devices. Those characteristics are critical for new AI devices.

Margins are hard to keep in disruptive times

Margins in the microprocessor industry are healthy, but we can observe some deterioration in practically all the players except Nvidia, which we could consider a special case (Figure 3). The table uses Cash Margin, which involves adjusting net income for non-cash items such as amortization and depreciation, stock-based compensation, and deferred income tax.

The margin deterioration is due to the intensity of product competition around AI, which causes companies to invest more in research and development. AMD has increased R&D expenses from 17-21% of revenue in 2021 and 2022 to 25-26% in the last two years. The company launched the MI325 AI accelerator last quarter. MI350 series is announced to be launched in the middle of the year, and the company keeps developing its developers’ software platform ROCm, trying to catch up with Nvidia’s CUDA platform.

Figure 3: Author

I foresee that margins will continue at that level during times of technological disruption and will improve later in more stable times. Data Center margins are higher, and I think AMD will keep gaining market share in the desktop and laptop market over Intel. It will give more scale, and as we have already seen, it will raise its prices. AMD will be the second-best player in each of the Data Center and Client markets, positioning the company in the 20% range of cash margin.

Valuation

Figure 4 shows the company's value drivers, considering a year as the last four quarters to capture the latest information.

Figure 4: Author

My revenue estimation is based on high growth in the Data Center segment, but not as fast as last year. It will be 40% until 2029, and then it will decrease to 25%. The Client segment will grow at a healthy 25% in 2025, 20% until 2031, and then 10%. The Gaming and Embedded segments will keep decreasing in 2025 and 2026 due to the inventory’s normalization, and later, they will grow slightly ahead of inflation, which is quite conservative. I estimate that revenue growth could grow five percentage points higher than my base case and eight percentage points lower, and I have modeled this risk as a normal distribution.

The cash margin will remain stable until 2027, a period of high AI innovation and investment in research and development. Later, it will grow slightly to 20% and 21% in 2032-2034. The improvement is due to lower research and development investment and a better position in the Client segment thanks to a market share gain over Intel, as we have discussed previously. Those cash margins can be up three percentage points over the base case or two percentage points under the base case, and I have also modeled it as a normal distribution.

Net working capital variation and CAPEX will remain at 6% of revenue and 2% of revenue, respectively. Cash flows will be discounted at an 11.3% WACC because the beta is 1.71. The risk-free rate is 4.5%. The company's leverage is insignificant. The perpetual growth rate is set at 3%.

Figure 5: Author

Based on Figure 5, I estimate a value of $123 per share, a 10% premium over its stock price. As I have declared in my previous articles, this premium doesn’t have a margin of safety enough to be a healthy investment. When I run the normal distribution of revenue growth and cash margin, the risk software Crystal Ball tells us that there is a 54% chance of getting a value higher than the current stock price. The revenue growth rate explains 86% of the variation of the value and the cash margin explains 14%.

In Figure 6, you can see that the multiple is the highest of the companies that we have talked about. The company has suffered a correction from the beginning of last year, and the current multiple is more reasonable. It is even higher than Nvidia’s (x59) and Broadcom’s (x57), which are two of the hottest stocks in the market. In the third tier, Qualcomm and Intel are facing structural problems, as we have mentioned.

Figure 6

The main risk is the ARM technology

AMD's main strategic priority is to use and promote the x86 architecture. It is collaborating with Intel through the x86 Ecosystem Advisory Group to strengthen the architecture and enhance compatibility across x86 platforms. This collaboration includes players like Broadcom, Dell, Google, HPE, and Microsoft.

Unlike Intel, AMD doesn’t manufacture its chips, and this fabless strategy allows it to be more flexible in adopting new architectures. This flexibility is what AMD can use in ARM technology. AMD holds a license and has stated its willingness to produce ARM chips but is not currently using them. The company has integrated ARM TrustZone security technology into its processors. They even developed ARM cores, but they discontinued these developments. I think AMD is not interested in ARM technology because, with x86 architecture, there is just one competitor, Intel, whose financial health is weak. AMD is gaining market share due to this weakness on Intel. Promoting ARM technology will increase the number of competitors like Qualcomm. This technological trend towards ARM is happening. For example, Qualcomm is gaining a 10% market share in PCs above $800.

My hypothesis is that the evolution will tend to reduce x86 architecture chips favoring ARM’s but AMD will have the flexibility to capture some of the market with ARM technology later on.

Figure 7: Counterpoint

Recently, Meta has announced the use of ARM technology in its data center servers, displacing the chips that used to be provided by Intel and AMD.

Conclusion

AMD’s stock price decreased 35% last year, but I estimate its value is still in line with the current stock price. For a buying opportunity, the stock price has to be reduced further. I recommend holding on to the stock because AMD is a clear winner in the AI space, with flexible capabilities to adapt and take advantage of AI expansion.