For the last 18 months, cryptocurrency growth has been partially attributed to the revenue growth for Advanced Micro Devices (NASDAQ: AMD). As CEO Lisa Su has put it, “On the Computing and Graphic segment, …, If you look at blockchain in particular, our estimates are that it was about a third of that growth,” (Q4 2017 ER). On the other hand, it is this same correlation which led the street to downgrade AMD in light of the volatility and the potential competition ahead in crypto. Depending on whom you ask, AMD bulls and bears all have different argument how crypto impacts AMD. In this post, I used the shareholders’ perspective to present my argument that how and if cryptocurrency risk affected AMD share values.
AMD Stock Prices Moved with Bitcoin Prices
An obvious to start is to examine the historical relationship between crypto demand, i.e., Bitcoin prices, and AMD stock prices, since Ryzen and Radeon are widely used in crypto mining. At the first level, there is a strong positive correlation between the two price levels (Figure 1). The economic intuition behind this relationship seems straightforward. Higher Bitcoin prices, reflecting higher crypto demand, will contribute to AMD’s Computing and Graphics segment revenue. In other words, a high Bitcoin price may lead to a high AMD stock price.
It should not come as a surprise to anyone that Bitcoin prices are volatile and crypto demand is not sustainable. There is a worldwide shortage of memory cards (GDDR5 VRAM memory) and all high-end GPU card along with mid-tier cards sold out during December 2017. It is not a secret that there is a tug of war between gamers and crypto miners. Even if AMD’s revenue may have been helped by the short-term crypto demand, it is unrealistic to expect that the crypto-based revenue will replace gaming-based revenue. I would like to think that AMD investors may have incorporated the expectation in stock prices over time as Bitcoin exhibits more and more volatility.
AMD Is Less and Less Affected by Bitcoin.
To verify this assertion, I measured the price sensitivity between AND prices and Bitcoin prices over the last thirteen 6-month periods. The price sensitivity or Bitcoin beta, measured in a similar way like stock beta, is computed between AMD stock prices and Bitcoin prices over each of the thirteen 6-month periods. As expected, AMD stock prices became less responsive to Bitcoin prices over time (Figure 3). The Bitcoin beta has been dropped off from 25 to 0.9 in last 5 years, suggesting that AMD stock prices have been less affected by Bitcoin prices.
AMD Stock Returns Do Not Move with Bitcoin Returns
However, the appearance of a positive correlation may be “spurious” if both prices may have been driven by a third common factor while Bitcoin and AMD are not related to each other. In other words, any two random time series that has increased over 2010 and 2017 would exhibit a similar relationship, yet providing meaningless economic intuition. As investors should only be interested in price changes (returns) instead of price level, the more relevant comparison should be on the relationship between AMD monthly stock returns with Bitcoin monthly returns (Figure 2). In this case, it is clear there is no obvious relationship between AMD stock returns and Bitcoin returns. In other words, rising Bitcoin prices do not lead to rising AMD stock prices.
Crypto Impact Already Factored in AMD Prices
The difference between stock price and expected return on investment may be subtle but important. As stock price levels should have reflected the expected new information, stock price should not change unless there are changes in expectation. Since 2016 when Bitcoin prices rose sharply, AMD and Nvidia were widely reported as the main beneficiary of the recent Bitcoin craze. While there's a general reluctance to acknowledge the impact on their traditional gaming customers, it is estimated that crypto mining contributes 5%-6% to the company's top line.
In his last report, Susquehanna Financial Group’s Christopher Rolland estimated that AMD and Nvidia split more than $500 million in Ethereum-related sales of their graphics processing units in the fourth quarter of 2017. But recognizing the complex tradeoff, Rolland also considers crypto-related sales as “false revenue” because of future potential ASIC replacements or crypto prices crashing. During the last xx months, crypto has been declined by xx%. In his second report, it is the same Roland that cited impending competition from Bitmain, a cryptocurrency mining company. Furthermore, Morris Chang's commentary on Q4 stated, ‘ASICs are and will be taking some of this market,’ but there are new currencies being introduced all the time that will continue to drive GPU demand.’”
In short, I believe that crypto risk has been well expected by the market. If analysts have reflected the cryptal risk into their revenue estimates, stock prices should have not reacted to crypto risk anymore after being adjusted to the street estimates. It is under this context, I re-estimated the Bitcoin beta after first adjusted stock prices with analysts’ forecasts of revenue. In Figure 4, it is clear that most Bitcoin betas are virtually zero, suggesting that the AMD stock reaction of crypto risk has been overstated.
Takeaways for AMD Longs
For a long time now, AMD shareholders have been bothered by the fact that AMD share prices seemingly failed to respond to company’s positive earnings releases. The finding in this post may provide some relief on the doubt of market efficiency. It is highly likely that AMD stock price has not reacted positively to the revenue beats in Q3 & Q4 2017 because investors have already priced in the unsustainability of revenue increases from crypto mining.
Therefore, this is not the time to overreact to it.
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.