Advanced Micro Devices (AMD) Valuation

Our first couple posts focused on the front end of the semiconductor supply chain with Applied Materials (AMAT) and LAM Research (LRCX). This week, we are jumping to the back end of the semiconductor supply chain with Advanced Micro Devices (AMD). AMD’s stock price has been hit harder than most, 53.1% decline, over the past year within the rising interest rate environment along with a slowdown in some consumer segments of the market. Same as we did over the last couple posts; we will go through both a relative and DCF valuation analysis to estimate a target share price. Next week, I will take a look at one of AMD’s biggest competitors in Nvidia.

Advanced Micro Devices (AMD) Overview

AMD was founded in 1969 in Silicon Valley and is headquartered in Santa Clara, California. AMD is what is known as a fabless semiconductor company which means they design and sell semiconductor chips, but they outsource the actual fabrication of the chips. They use semiconductor contract producers known as foundries to fabricate chips to their designs. AMD then turns around and sells those chips to their customers with include equipment manufacturers, cloud service providers, independent distributors, and retailers. AMD’s customers span the globe, but the bulk of their sales are concentrated in the United States and China. Here’s a breakdown of their revenue by geographic region from 2017 through 2021 (AMD has not released a geographic revenue breakdown for 2022 yet).

Until this past year, AMD broke their business into 2 reporting segments including Computing and Graphics, and Enterprise, Embedded, and Semi-Custom. The Computing and Graphics business segment included desktop and laptop microprocessors, accelerated processing units, and graphics processing units (GPUs).  Whereas Enterprise, Embedded, and Semi-Custom included embedded processors, System-on-Chip products, and products for game consoles. Here’s a summary of AMD’s segment revenue breakdown from 2017 through 2021.

In the second quarter of 2022, AMD decided to change their reporting segments to align more closely with the end markets they serve. These new reporting segments include Data Center, Client, Gaming, and Embedded. It appears that the original Computing and Graphics business unit contributes mostly to the new Data Center and Client business units. Whereas, the previous Enterprise, Embedded, and Semi-Custom business units mainly contribute to what is now the Gaming and Embedded business units.

In February 2022, AMD acquired a company called Xilinx which now has their revenue rolled into AMD’s reporting. Most of Xilinx’s revenue is reported in AMD’s Embedded unit with a small portion contributing to Data Center. Below is a breakdown of the segment revenue breakdown reported in AMD’s latest quarterly filing.

Relative Valuation Analysis

I started with a relative valuation analysis before diving into the discounted cash flow (DCF) valuation. For the comparable company’s portion of the analysis, I used the following companies:

• Nvidia (NVDA)
• Broadcom (AVGO)
• Texas Instruments (TXN)
• Analog Devices (ADI)
• Microchip Technology (MCHP)

As of 1/8/2022, AMD is trading at an 18.5x EV/EBITDA multiple and P/E ratio of 39.7x. Both values for AMD fall above both the average and median values calculated for the comparable companies. In fact, AMD’s P/E ratio is higher than all 5 comparable companies and only Nvidia has a higher EV/EBITDA ratio. 

With the comps table complete, we can proceed to the historical P/E portion of the analysis. I decided to use AMD’s P/E ratio data back to the end of 2017. Prior to 2017, AMD consistently showed negative earnings and therefore did not have an applicable P/E ratio. After all the data is in Excel from macrotrends.com, we can identify our upside, middle of road, and downside price scenarios. Below is the full relative valuation table that allows me to identify any overly conservative or optimistic scenarios that should be excluded from consideration.  In the summary section below, I calculate a potential upside reward of +$43.40 (+67.8%) and a downside risk of -$20.52 (-32.1%) relative to AMD’s price per share of $63.96 as of 1/8/22.

DCF Model Assumptions

Now it’s time to move onto the Discounted Cash Flow (DCF) valuation portion of the analysis. Just as we did for previously, I used the past 5-years of financial statements to analyze historical trends and averages. AMD’s top line revenue has grown drastically over the last 5 years from $5,253 million in 2017 to an estimated $19,036 million in 2022 without the addition of Xilinx (total estimated revenue for 2022 is $23,502 million with the inclusion of the Xilinx acquisition). This growth rate equates to a 5-year compound average growth rate (CAGR) of 29.4%. AMD’s cost of goods has consistently decreased over the past 5-years from about 62% in 2018 to just above 50% in 2022. Here you can see a breakdown of the 5-year averages for most of the factors used in the DCF evaluation.

AMD is a major company in the semiconductor industry which is considered a growth industry, there are predictions that it will continue its strong growth trajectory through 2030. Which is why I decided to use the same 8-year forecast period with 2030 as the terminal year. To start our forward-looking forecast assumptions, I identified line items that were consistent throughout the 5-year historical period which I believe will remain consistent throughout the forecast period.

• R&D – 20% of Total Revenue
• SG&A – 10% of Total Revenue
• Interest Rate – 5.5% of Debt
• Tax Rate – 15% of EBT
• Accounts Receivable – 70 days
• Inventory Days – 90 days
• Prepaid Expenses – 40 days
• Accrued Liabilities – 55 days
• Accrued Expenses – 27% of COGS
• Capital Expenditures – 2.5% of Total Revenue
• Share-based Compensation – 2.5% of Total Revenue   

For the remaining line items below, we needed to take some different approaches throughout the forecast period.

• Cost of Goods Sold – AMD’s gross margin has improved over the past 5-years but I expect their pricing power to fall slightly over time as the competitive landscape changes. COGS in the first forecast year is consistent with 2022 at 51% of revenue. I decided to gradually increase COGS each year until I reached the historical average of 55% of revenue in the final 4 years of the forecast period.
• Accounts Payable – Accounts payable days spiked up in 2022 compared to the previous couple years. Therefore, I started the forecast period at the historical average of 75 days in 2023 then tapered down each year until I got to 60 days for 2026-2030.
• Total Revenue – I estimated revenue to increase by a modest 7% in 2023 before ramping back up close to the average historical growth rate over the middle of the forecast period. I then estimated the revenue growth rate to pull back over the last few years mainly due to the uncertainty in predicting a high growth industry that far into the future.

To determine AMD’s top line revenue growth forecast, I parsed together information I found through research and then made some assumptions myself on the extended years of the forecast. I stumbled upon a website called Koyfin.com recently that I found very useful when analyzing financial performance data along with professional analyst forecast assumptions. AMD has cited recent weakness in the Client business segment and unfortunately that is the largest segment in terms of revenue for AMD. Therefore, I believe AMD will have a slow couple years before they ramp back up once we get to the other side of the predicted upcoming economic recession. Due to the rising interest rates to combat inflation, it is likely that consumers and enterprises will hold off on some larger technology purchases until the economic future is clearer. Here’s my table summarizing the annual revenue growth rate for each of the 8 years in my forecast. 

DCF Output

Now it’s time for us to put everything together in the DCF valuation. First, I performed a Weighted Average Cost of Capital (WACC) analysis to determine a free cash flow discount rate of 12.6%. I settled on a terminal EV/EBITDA multiple of 24x based on an average EV/EBITDA multiple across the relative valuation scenarios above. The 24x multiple appears to be a conservative estimate as it falls between AMD’s 1st quartile and average historical PE ratios.

The resulting DCF analysis calculated a diluted price per share of $78.06 (+22.0% potential upside compared to AMD’s share price of $63.96 on 1/8/23). Here you can see the full breakdown of my DCF valuation.

Just like previous valuation assessments, I performed a DCF share price sensitivity analysis to see how the price fluctuates under different assumption conditions. Below is the data table showing AMD’s DCF share prices under different WACC and EV/EBITDA multiples assumptions.

AMD Price Target

To determine an estimated price target, I put together a “football field” chart as shown below that incorporates relative and DCF valuation ranges along with AMD’s 52-week high/low and analyst price targets found on Yahoo Finance.  I tend to give more weight to the DCF and relative valuation components of the analysis.  Therefore, my current price target for AMD as of 1/8/22 is $80 (+25.1% potential upside compared to the current price of $63.96).  However, this estimated price target is based on an 8-year forecast and my current forecast estimates 2023 earnings to be flat compared 2022. Therefore, any further headwinds to AMD’s business not currently factored into the market price could very easily drive the short-term price for AMD lower. Also, AMD’s current P/E ratio is higher than their sector competitors and slower growth could lead to a decreased P/E ratio that will bring it more in line with their peers. Any short-term pullbacks in AMD’s share price would make a potential investment more attractive as their long term growth potential is still very high.

Please join me next week as I look into Nvidia (NVDA), one of AMD’s main competitors. We will go through the same valuation process and then provide some thoughts on how Nvidia and AMD stack up to one another as investment options.  

Disclaimer: These valuations are based on my personal assumptions, calculations, and assessment.  I am not a financial advisor, nor should my assessment be considered an investment recommendation.