Jan 22 2020 - AMD: Fundamentals Trending Higher

Summary

AMD shares are still on a tear, reaching .60 at the time of writing. The business is currently valued at billion. I find this rally hard to believe.

The way I see it, AMD shares are prone to a negative catalyst, meaning a negative catalyst could cause a huge downward adjustment in the stock. However, fundamentals trend higher.

AMD's real opportunity is in the x86 CPU market. AMD is able to attack multiple growth vectors without much competition from Intel: notebook (Ryzen Mobile), desktop (Ryzen), and server (EPYC).

As of writing, Nvidia is valued at 5 billion on exclusively GPU. Intel is valued at 0 billion on CPU and memory. AMD is valued at billion while attacking both the CPU and GPU markets.

Reiterating HOLD, PT raised from to .

Taking A Step Back

While this is already common knowledge, AMD (NASDAQ:AMD) has had an absolutely phenomenal run over the last several years. Just when I believe the rally has lost fuel, the company always finds away to impress the market, and the stock moves higher. AMD has been the top performer in the S&P 500 back-to-back years. That itself is phenomenal.

ChartData by YCharts

The staggering rally in AMD shares has come off the back of AMD's strategy since Lisa Su took over as CEO. Let's look at what AMD has done since Lisa Su took over.

  • balance sheet rework
  • high performance computing
  • moving from GlobalFoundries to TSMC (NYSE:TSM)
  • datacenter push (GPU & CPU)
  • competitive "value" pricing
  • taken advantage of Intel's (NASDAQ:INTC) stumbling

Starting with AMD's rework of their balance sheet. The turnaround in the financials (balance sheet and cash flow statement in particular) has been extraordinary. AMD has gone from a business strapped with debt and near bankruptcy, with steady cash bleed to a financially sound and improving business. In 2014, AMD had a net cash deficit of .17 billion, with exploding debt and continued cash burn. When Lisa Su took over the helm of AMD, one of her top priorities was paying off the company's burdensome debt load. As of Q3, AMD is actually net cash positive, with a net cash surplus of 2 million. AMD now has the ability to pay off their entire debt load with their current cash on hand. With regards to cash flow, AMD burned through ~3 million in cash in 2014. Since then, AMD has continued to burn cash, but a lot of it has been going into paying off debt over the last few years. This year however, it looks like the business will near a breakeven year in terms of cash flow, with the company seeing FCF of 9 million in Q3 alone. Overall, the company has been fairly disciplined financially. AMD's improved financial footing should allow the company to focus on expanding their market share and focus on developing high quality products.

With regards to product strategy AMD has executed beautifully over the last several years. Other than de-risking the financials, management's largest focus has been on driving high-performance computing that can be competitive with Intel and Nvidia (NASDAQ:NVDA). In previous years, AMD (and ATI pre-acquisition) focused on carving out a slice of the low-end of the CPU and GPU markets. AMD's products were far less powerful, but consumers would buy the products solely based on the low-price points. When Dr. Su took over, AMD focused on overhauling previous low-performance architectures. Before Su took over as the head of AMD, the company built products on the Bulldozer architecture. When Su took over, she focused on developing high performance products, rather than budget products. So, AMD started working on a new CPU architecture called Zen, a new server processor called EPYC, and overhauled the company's graphics business. The company's focus on high-end computing drove costs (particularly R&D) higher, but allowed the company to begin technological improvements and market share grabs long-term. This revamped strategy combined with AMD's technological improvements over the last few years has led to market share expansion, a solid long-term strategy.

In addition, AMD has moved away from GlobalFoundries and moved towards TSMC for the manufacturing of the lower node process technologies. While GlobalFoundries is set to continue manufacturing AMD's 12nm and larger products through 2024, AMD is moving to TSMC for the 7nm and beyond nodes. Switching from GlobalFoundries to TSMC allows AMD to sidestep any production issues that competitors (Intel) may be facing. Since Intel operates its own fabs, and do not outsource production, they are now facing headwinds. AMD outsourcing to TSMC, whose whole business is semiconductor manufacturing is a smart strategic decision. While Intel will begin 7nm production en masse in 2021, AMD (via TSMC) will already be off to volume production on 5nm. This puts AMD perpetually ahead of Intel. This is important, as it is a core driver of my fundamental thesis on AMD.

49381769-15790550950946038.png(source: MarketRealist)

One of the most important growth drivers for AMD right now and going forward, is their datacenter business. Right now, EPYC is driving growth, and will be for the foreseeable future. That being said, AMD may have an interesting value proposition going with their Radeon Instinct AI GPU accelerators. When AMD launched their original Ryzen CPUs targeted at the PC (primarily desktop) market, they also attacked the server/datacenter market with a new product line and architecture, AMD EPYC. AMD hadn't seen volume production until just a few quarters ago. Prior to that, AMD was still in the sampling phase of development and deployment, with only a few customers deploying EPYC. Now however, as performance advantages over Intel come out as well as EPYC setting new records, it seems like AMD gaining massive share is inevitable.

49381769-1579057135968259.png(source)

Intel's Platinum series processors, per my understanding, are Intel's top-of-the-line server CPUs. AMD establishing a lead in multiple task loads against Intel (and by a hefty margin) establishes the company as the performance leader. AMD becoming more than just competitive in the datacenter bodes well for market share gains.

Since AMD came to market with Ryzen and Radeon, the company has focused on delivering on a "value" pricing strategy. Because AMD hasn't controlled the top-end of both of these markets, price has become an important determining factor in consumer interest for AMD products. AMD's strategy to this point has been to deliver comparable performance products, at marked down price points, yielding a robust price/performance ratio. This has however come at the expense of margins. While Nvidia and Intel are able to maintain ~63% gross margins, AMD has been stuck in the 40s for a while. As AMD opens up to the high-end of the GPU, server, and desktop/laptop CPU segments, they will be able to leverage ASPs higher, driving margins. This is the biggest stretch in my valuation (more on that later), the idea that AMD can break into and maintain the high-end of their markets with margins comparable to Intel and Nvidia. If AMD maintains a "value" pricing strategy, not beating the competition on raw performance, but on performance-per-dollar, these ~60%+ margins will be very difficult to achieve.

In the past, AMD has been bailed out by Intel. Intel has needed AMD to survive (just barely) in order to claim a lack of monopolization of the CPU industry. However, Intel's dominance in the CPU business (desktop, laptop, and server) has morphed into complacency, flat-footedness, and even price-gouging tendencies. For those that don't know, Intel assembles its CPU processors in-house. They do not outsource production of their chips to other companies (like TSMC, or GlobalFoundries as AMD does). These outside companies specific operation is semiconductor manufacturing. They do not have to deal with the entire semiconductor supply chain that a company like Intel has to. Eventually, somewhere, Intel messed up. Now, they are playing a game of catch-up as AMD races ahead of Intel in the lithography race. By the time Intel is working on 7nm, in 2021, TSMC will already be working on volume production of the 5nm node. This structural imbalance is an opportunity for AMD, an opportunity they have taken. AMD continues to take share at the high-end of the CPU range, with 7nm and 10nm offerings, while Intel attempts to stretch their 14nm technology as far as possible (this only leads to skyrocketing power consumption/TDP, something AMD was known for years ago). AMD has full-on taken advantage of Intel's stumbling, and is now seeing rapidly increasing market share as a result.

Valuation

Now, on to the most interesting part of the article (at least in my opinion), valuing AMD. First of all, we have to estimate the size of the markets AMD is attacking:

  • PC CPU (desktop/notebook)
  • GPU (desktop/notebook)
  • Datacenter (CPU/GPU)

This is AMD's estimated TAM (total addressable market) as of Q2'18. AMD has not publicly updated these estimates since Q2'18:

49381769-15793117578215368.pngRight now, I believe AMD's combined (desktop and notebook) market share of the PC CPU business is at ~16%. I have my estimate at 20% combined share by 2023, as AMD begins to ramp into mobile, and expand into the higher margin high-end of the desktop market. There are already estimates that AMD will hit 20% in mobile market share in Q1 of this year. As AMD continues to ramp a strong product roadmap over the next several years, continued market share expansion is likely. My target is combined market share of 20%. By 2023, I would anticipate that this billion TAM expands another 10%, to billion. On 20% share, AMD should see PC related revenue of .6 billion.

AMD's graphics business, again a combination of desktop and notebook, should be stagnant over the next several years. Nvidia continues to dominate the high-end of the discrete GPU market, while Intel continues to dominate in mobile, with their integrated graphics technology. While AMD should remain competitive in the low to mid range of the graphics market, and should continue to gain business from the newest PlayStation and Xbox releases, it is hard for me to see AMD really breaking into the graphics market and dominating. As of now, AMD is having a tough time competing with Nvidia's HEDT (high-end desktop) offerings like the RTX 2080 Ti even with their newest releases. Also throw in the fact that Nvidia could launch their own 7nm high-end processors (code name Ampere) some time in Q1, gaining market share will be difficult for AMD. AMD pegged their TAM estimate for 2020 at billion, but this was before we saw the correction in the GPU market caused by the cryptocurrency bubble. I estimate that by 2023, the overall GPU market (excluding any ML/DL GPU market) will be valued at billion. Assuming 15% share (reasonable considering AMD's technological lag), AMD gets to GPU revenues of .8 billion.

So, Ryzen and Radeon should combine for overall revenues of .4 billion.

The juicy part is when we get to AMD EPYC. AMD's highest recorded market share in the server business was ~26% in 2006. This was based on lower-end pricing, with lower-end technology. However, AMD is clearly demonstrating that they are a leader in server technology, with multiple performance improvements versus Intel's Xeon server processors. While I do not see an expansion from the ~5% (estimated) Q4 market share to 26% in three years, I believe it is reasonable to anticipate an expansion to 20% by 2023. First of all, management is guiding for "double-digit market share" in data center, a 10%+ number. Assuming AMD can keep up this pace over the next few years, with 7nm Rome, 7nm+ after that, and 5nm, keeping a solid lead on Intel, the company should easily achieve 20% share. Assuming the market grows from billion in 2020, to billion by 2023 (as cloud capex begins to reaccelerate), AMD EPYC sees billion in revenue. This is one of the two riskiest assumptions in the model, assuming that AMD will find a way to scale their server business so far so fast, and that Intel will not get their act together in the next couple of years. I still believe it is a reasonable assumption, but AMD really needs to execute to see that 20%+ market share come to fruition.

Now that we have a revenue estimate of .4 billion, the attention has to turn to margins. My projection is that AMD gets to 60% gross margins, comparable to Nvidia and Intel's ~63% gross margins. AMD gets to these margins with continued moves into higher performance/higher price segments of the market. This will allows them to leverage higher ASPs, with little unit cost growth. This combination, especially with a push into the enterprise cloud market should spur margin expansion. This, other than my datacenter market share, is likely the biggest stretch in my valuation model. Keep in mind, as of Q3'19, AMD's gross margins are at ~43.1%, meaning this expectation is for ~17% margin expansion in just three years. Even for a company that has executed as well as AMD has, that is a lot to ask for. So, 60% may be an ambitious goal. With regards to operating expenses, in calendar 2018, Nvidia and Intel both saw opex as a % of revenue at 28.7% and 28.8% respectively. In 2018, AMD's opex level was a little more inflated, at 30.7%. I believe AMD will invest heavily in R&D over the coming years to maintain their leadership in CPU, retain their console partnerships, and attack the high-end of the GPU market. So, I am anticipating 30% of revenue will be devoted to operating expenses. This yields operating income of .32 billion. Assuming a 20% tax rate, AMD should see a net income of .456 billion.

I believe an earnings multiple of 20X is justified, considering AMD's likely technological lead in CPUs (server and PC) beyond 2023. This values the business at .12 billion. Assuming an annual discount rate back to 2020 of 10%/year, the business is valued at .388 billion. On 1.136 billion shares outstanding, the stock should be valued at ~.35/share.

Conclusion

AMD shares, at least on a fundamental basis are overvalued. While I believe fundamentals continue to improve in a drastic manner, it is priced into the stock. Overall, I am optimistic on AMD's ability to take share from Intel in the PC and datacenter markets. The stock does not have enough downside to warrant a sell rating, so I'm sticking to a hold for now. Overall, I believe the business is firing on all cylinders, but the stock is overvalued.

TIPRANKS: HOLD

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: This is not financial advice. Please do not interpret this as financial advice. I am not a financial advisor. This is just my personal opinion.

0
0
0
0
0
0
0 0 80
Submit comment
    No comments yet

Oct 19 2020 - AMD: A Gambler's Bet

Apr 20 2020 - AMD: Still Going Strong

Jan 19 2022- Nvidia Vs. AMD: Which Is The Better Growth Stock?

Oct 10 2021 - AMD: Prepare For A Selloff

Apr 20 2020 - Updating My 'Short' Thoughts On AMD

AMD Closes At New Record High Is This The Most Overvalued Chip Stock On The Market?

My Top Tech Pick Of 2020 - AMD Has Beaten The S&P By 15x Since I First Recommended It

Mar 5 2020 - AMD Shares Jump as it Discloses Financial Targets and Much More

Submit media
Enter your nickname

Show

Show

Enter your email address and we will send you an email explaining how to change your password or activate your account.

Back to login form

Close