AMD reported Q1 2020 on April 27 and it was spectacular. Revenue came in at $3.45B and guidance is $3.6B for Q2. AMD further increased whole year guidance to 50% (originally 37%) increase over 2020.
However, AMD share price had languished since, and is now merely $77, even after announcing a significant stock buyback program of up to $4B.
Even if AMD goes on to merely meet this latest guidance, 2021 revenue will come in at $14.64B. Using a gross margin of 47% and CAPEX of 26%, and 15% tax rate (all provided by AMD), net income will come in at $2.61B. Divided by total shares of 1.24B, net income will be $2.11 per share.
In other word, Mr. Market is assigning a forward PE ratio of 36 to AMD.
My personal guess is that AMD will beat the above numbers, given that Lisa Su has a history of giving very conservative guidance that she can beat. I am leaning toward 2021 revenue of $15B, and since AMD chips are selling so well, it’s possible for gross margin to come in at 48%. So net income will be $2.8B and earnings per share will be $2.26. Give it a PE ratio of 40, AMD share price should be $90.
So the question is, what PE ratio does AMD deserve? How does 2022 and beyond looks like? Let’s examine.
In 2018, AMD’s total revenue was $6.48B, earning per share was $0.46.
In 2019, total revenue was $6.73B, earning per share was $0.64.
In 2020, total revenue was $9.76B, earning per share was $1.29. Remember this is the pandemic/work from home year.
Now in 2021, when there is global chip shortage, earning will be $15B, with earning per share of $2.26.
I think normally, given the rate of such astonishing increase of both revenue and net income, PE ratio of 50 is well deserved.
However, the current market is discounting AMD’s future. So what are the negative factors?
First, we have inflation and 10 year treasury yield going up. Sure, the US printing USD is a huge problem (national debt up $6T in one year, Federal Reserve balance sheet increased from $4T to $8T in one year), but AMD’s business will continue to flourish regardless, short of a global crash. In that case, no asset will be safe and companies like Nvidia shouldn’t be at $600+ either.
Then there is the Xilinx acquisition, which has acted as a dead weight around AMD’s neck. I find this very strange. Xilinx will bring many benefits to AMD’s future. FPGA, SmartNIC, AI, Automotive, 5G, embedded industrial, software expertise are just the obvious benefits. Using Infinity Fabric, AMD can further integrate Xilinx IP on their SOC. AMD already said that this acquisition is immediately accretive. Xilinx brings lots of cash, higher gross margin and high quality revenue and earning. Additionally, no objections have been raised by any government, companies or organizations. Yet the market is treating this deal like shit. Meanwhile, Nvidia’s purchase of ARM is treated as a great benefit, even though I have significant doubt that it will go through. There are many objections by other companies, and also Great Britain, and in the end, likely China. So, once AMD’s deal goes through, and Nvidia’s buy does not, perhaps share prices will act differently.
Sector rotation. Money is flowing back into energy, commodities, economy re-opening plays etc.
I believe that the cloud and data center will continue to expand. Computing and gaming will continue to grow. Use of chips in autos will continue to grow. Use of 5G, IoT etc will expand. I don’t buy this simple theory of PC sale will slow once pandemic is over, therefore AMD’s growth will slow.
Money is flowing into crypto. I do believe that a significant amount of retail money is flowing into crypto, and probably even wall street money. I can’t comment on the future of crypto, but I think gradually an equilibrium will be achieved, where this flow ceases to be a huge speculative WallStreetBet type of meme.
Chip shortage. The market seems to think that since there is chip shortage, that AMD is unable to grow next year and beyond. For the first four months of this year, I was unable to buy the Zen 3 Ryzen CPUs on Amazon or anywhere else, not the 5800X, not 5900X and certainly not the 5950X. Then suddenly around April, the Ryzen 5800X became in stock on Amazon. Now you can buy Ryzen 5900X and 5950X on Amazon, albeit at an inflated price. I was able to buy a Ryzen 5950X at a local MicroCenter store a few days ago, at MSRP. The store got a shipment of twenty-five 5950X in the morning and by the time I got there around noon there was five left. Same thing goes for the 5900X. Incredibly, plenty of motherboards are now available, whereas a few months ago they were very rare. So the point is, as Lisa Su had stated, AMD’s wafer supply from TSMC will improve significantly in H2 2021. Next to improve will be the GPUs. By 2021, AMD will continue to grow by taking market shares from Intel and Nvidia.
Apple’s M1. Apple was using Intel chips. M1 performs very well, using 5 nm process and super software integration. This doesn’t mean that X86 CPUs are suddenly obsolete. AMD’s mobile CPUs will also be very competitive, in terms of power use and performance.
Amazon Graviton CPU. Alibaba RISC-V custom processors. Big companies can design their own custom chips for their own use. General data center instances will still be X86 based. AMD has a full ARM license and can design ARM or RISC-V based processors. Lisa Su basically said that there is no money to be made in that direction right now. However, AMD can go that direction in the future if that’s where the TAM is going. I trust Lisa Su’s vision in this regard.
CAGR of 20%. I believe this projection by AMD is holding AMD back. AMD has not revised this number yet. Given AMD had grown by 45% in 2020, and now even 50% for 2021, is AMD expecting to grow much slower in 2022 and beyond? Shouldn’t AMD be revising the CAGR up to 30% or more?
I look forward to AMD holding another Financial Analyst Day like the one they last hosted in 2020.
Since Xilinx has not been growing at such a rapid pace (especially due to US sanction on Huawei in 2020), AMD needs to reassure investors that even after folding Xilinx into AMD, AMD will continue to grow at a fast pace.
So what are the positive drivers?
Lisa Su will give the keynote speech at Computex in early June. She will likely talk about the Frontier exascale supercomputer. Hopefully she will give us some surprises.
Samsung will use Radeon GPU in the upcoming Exynos 2200 chip, which will be in the Galaxy S22 Ultra and maybe even in new Samsung laptops/tablets. This brings AMD into the mobile space, which will be huge, going forward. We will likely know about this later in the year, since the Galaxy S phones are usually launched in January of each year.
News on a successful Xilinx acquisition and perhaps news of Nvidia failing to buy ARM will boost AMD.
PS5 and XBox will sell every unit they put out, for the next four five years.
Epyc Milan will finally reach that inflection point Lisa Su always talked about. Epyc had finally breached 10% server volume. Once it becomes clear that Epyc will reach 20% + then AMD will get additional boost.
Ryzen, Radeon will continue to gain market shares. We need to hear about Instinct gaining traction. This is AMD’s data center GPU. Nvidia reign supreme in this space. Instinct needs to make more of a splash. Once it is deployed in the Frontier and many more supercomputers, ROCM will gain more respect.
Toward the end of 2021, I hope to hear more about how AMD intends to use Xilinx to grow more business/TAM. Right now it is vague and the market is discounting Xilinx entirely.
Looking ahead to 2022, if AMD can grow another 30%, revenue will be $19.5B. If you add Xilinx’s $3.5B, total revenue will be $23B. Net income could be $4.5B! Let’s see how the market value AMD by that time.