Summary
Advanced Micro Devices, Inc.'s entrance into the generative AI industry with its new GPUs could lead to significant sales growth and capture a portion of the expanding market.
The company's upcoming MI300 series GPUs have advanced features that can accommodate larger AI models than competitors like Nvidia Corporation.
Potential headwinds include competition from Intel Corporation in 2025 and the impact of worsening Sino-American relations on AMD's business in China.
Las Vegas Hosts Annual CES Trade Show
David Becker
After the impressive performance of Nvidia Corporation (NVDA $NVIDIA Corp(NVDA)$ ) in Q2, questions are being asked about whether Advanced Micro Devices, Inc. (NASDAQ:AMD $Advanced Micro Devices(AMD)$ ) could also manage to significantly improve its performance and grow its revenues at a double-digit rate once its GPUs for artificial intelligence, or AI, enter the market later this year. Considering that the generative AI industry is expected to aggressively expand in the following decade, it’s likely that AMD would be able to expand its own total addressable market ("TAM") and capture a significant portion of this growing market. That’s why even though the company is likely to face several major headwinds, there are reasons to believe that AMD has everything going for it to create additional shareholder value and repeat the success of Nvidia once its latest products are released in the following months.
TAM Expansion Is On The Way
Earlier this month, AMD reported its Q2 earnings results which showed that even though the company’s revenues decreased by 18.2% Y/Y to $5.36 billion, they were still above the estimates by $40 million. At the same time, the gross margin remained at 46%, while non-GAAP EPS of $0.58 was above the expectations by $0.01.
While AMD’s performance was mixed at best in Q2, there are reasons to believe that the company would be able to significantly grow its sales in years to come once it enters the generative AI industry by launching its newest GPUs in the following quarters. If we take a look at the recent performance of AMD’s competitor Nvidia, we’ll see that the generative AI growth opportunities appear to be endless at this stage.
In Q2 alone, Nvidia’s revenues increased by 101.6% Y/Y to $13.51 billion and were above expectations by $2.43 billion as the company was able to make nearly 1000% profits on its H100 GPUs that continue to be in high demand to this day. What’s more is that Nvidia expects to generate $16 billion in revenues in Q3 alone, which accounts for ~60% of its annual revenues last fiscal year, as generate AI industry is forecasted to grow at a CAGR of over 40% and worth over $1 trillion in the next decade.
That’s why AMD’s upcoming entrance into the industry is likely to become a major growth opportunity for the company, especially since the AI cluster engagements increased by more than seven times recently. AMD will enter the generative AI industry with the launch of its upcoming MI300 series GPUs that use the CDNA architecture and are designed specifically for large language models. On top of that, the most advanced GPU from the series the MI300X has the ability to use up to 192 gigabytes of memory against Nvidia’s H100 120 gigabytes of memory, which indicates that AMD’s GPU can fit even bigger AI models than others. As for the launch date, AMD’s newest GPUs are expected to enter the market at the end of this year and the company will be able to ramp up their production in 2024.
Considering this, the only issue is that we would need to wait for more than half a year until there is financial evidence of the new growth.
Potential Headwinds To Consider
Since Intel Corporation (INTC) is expected to enter the generative AI race only in 2025, AMD would have more than enough time to capture a sizable portion of the market until new entrants appear on the horizon. However, there’s a risk that the potential market share that AMD would be able to capture will be smaller than expected. This is due to the fact that earlier this month Nvidia introduced a more powerful superchip platform under the name of GH200 that is expected to be launched a year from now.
In addition to that, a further worsening of Sino-American relations could have a profound impact on AMD’s business, as China was its second-biggest market last year. Last year, the new chip restrictions that were announced by the Biden administration prohibited Nvidia from selling its H100 GPUs directly to Chinese customers. Considering that the MI300 series chips are going to be as powerful as Nvidia’s latest GPUs, it’s likely that AMD won’t fully benefit from their upcoming launch in one of its biggest markets.
Considering all of this, there are questions about whether it’s worth it to acquire AMD’s shares given everything that’s happening with the company. Back in February, when AMD’s shares were trading below $80 per share, I made a discounted cash flow ("DCF") model which showed that the company’s fair value is $87.28 per share. Since that time, AMD has greatly appreciated as the improvement of the overall economy along with the excitement for the generative AI revolution helped the stock overcome that price target.
Since the company released additional earnings reports in recent months, I’ve decided to update my model, which can be seen below. Most of the assumptions in the model remained the same, while only the revenue forecast has been greatly changed. For FY23, the revenue assumptions remain in-line with the Street estimates. After that, they increase considerably in the following years thanks to the improvement of the PC and data center markets along with the expansion of TAM as a result of the company’s entrance into the generative AI industry which should help AMD exceed previous expectations.
AMD's DCF Model
AMD's DCF Model (Historical Data: Seeking Alpha, Assumptions: Author)
The updated model shows that AMD’s fair value is $93.55 per share, which is above the previous calculations but also below the current market price. As such, at this stage, it all comes down to whether the generative AI revolution would live up to its promise and have a profound impact on AMD’s business. If the demand for AMD’s new AI chips would be the same as for Nvidia’s chips, then my revenue assumptions could be too conservative and would require me to make an upward revision of the estimates which could result in a higher valuation for the company’s shares. After all, I expect a 20% revenue growth for AMD, which could be too conservative given that Nvidia is expected to increase its revenues by 100% Y/Y this fiscal year thanks to its entrance into the generative AI business.
AMD's DCF Model
AMD's DCF Model (Historical Data: Seeking Alpha, Assumptions: Author)
The Bottom Line
At the current price, it’s safe to say that AMD is a high-risk and high-reward play. By trading at a forward P/E of ~40x and expecting a Y/Y decrease in sales this year, it’s hard to justify a long position in the company, especially since it appears that there’s no margin of safety when purchasing its shares at the current levels.
However, given the impressive performance of Nvidia in recent months and how it expects to continue to grow at an aggressive rate due to the increased demand for its GPUs, there are reasons to believe that AMD could share the same fate once its own GPUs for AI are released later this year.
Considering that the generative AI industry is expected to grow at a double-digit rate for years to come, AMD’s potential ability to make chips that power such growth could help the business exceed expectations, make me revise my model even more to the upside, and help the shares rally. That’s why I’m sticking with my buy rating as there are reasons to believe that the market underestimates the growth opportunities that would appear once AMD starts to power the ongoing generative AI revolution. But until that happens, AMD’s stock would nevertheless remain a risky bet at the current price despite all the growth opportunities.
Source: seeking alpha
Comments
Great