At first light of the rumored Tesla-AMD deal (“the Deal”), Advanced Micro Devices’ shares (NASDAQ: AMD) were up by 8.54%. The supposed rival Nvidia’s shares (NASDAQ: NVDA) were down by -5.20%. The 13%-14% stock return spread reflects market expectations that there will be some future Tesla or automobile segment revenue transfer from Nvidia to AMD. In this post, I will quantify the negligible revenue impact on Nvidia, yet a probable, significant positive impact for AMD, if AMD were to venture into self-driving technology. With this potential “cannibalization” effect, I also update the relative valuations for AMD, Nvidia and Intel.
A Non-Event for Nvidia?
Despite the fact that the Deal was quickly refuted by many parties, the market’s first response in terms of stock price reactions often tells a pretty good story regarding the potential revenue impact. Given NVDA’s P/S ratio of 13, a 5.2% stock return loss translates into a $5.64 billion market capitalization loss, or $33.4 million in annual revenue loss (Bloomberg). This would be equivalent to -0.40% loss on Nvidia’s annual revenue (Table 1), if the Deal were to go through.
Let’s put this -0.4% potential loss in perspective. If the market has had a significant reaction to the Deal, we should use a similar standard to examine Nvidia’s long list of other automobile partnerships. In addition to Tesla as Nvidia’s major partnership in autonomous car technology, recently, Volvo and Autoliv selected DRIVE PX for self-driving cars targeted to hit the market by 2021. ZF and HELLA, two leading automotive suppliers, announced a system based on DRIVE PX to deliver the highest NCAP safety ratings for cars. Baidu announced that its Project Apollo open-source self-driving platform for the China market will use DRIVE PX.
But, most notably is the addition of Toyota to the list. This partnership holds extreme potential, and could potentially allow up to 10 million cars per year to include Nvidia Drive series chips. Tesla has been incorporating Nvidia’s old version of automotive focused chips in their cars since 2014, and they were buying them from Nvidia for prices ranging between $1,000 and $1,500. If Toyota were to include the new series of Nvidia chips in just 10% of their cars per year, at the same price that Tesla was charged, this could increase Nvidia’s annual revenues by $1 to $1.5 billion just for the Toyota partnership alone. That would be equivalent to a 13% to 20% annual revenue increase for Nvidia. In short, this is the basis for the statement, “The Alleged Tesla-AMD Deal Should Be a Non-Event for Nvidia.”
A Potentially Positive Impact on AMD’s Revenue.
On the other hand, if the Deal were to happen, AMD stands to realize a significant push in automobile segment revenue. From Table 1, given AMD’s P/S ratio of 2.75, a +8.54% stock return translates into a $1.08 billion market capitalization increase, or a $143.5 million annual revenue increase. This is a significant 3.1% growth in revenue. This estimate comes close to Paulo Santos’s $100 million revenue increase to AMD.
Cannibalization and Zero-Sum Gain in Valuations
This is a unique time for AMD, NVDA, and Intel’s shares (NASDAQ: INTC). Each of them has posed a significant threat to the others in various segments of the chip market, as many new products have been rolled out. In the meantime, the market is in the midst of a significant tech rally, and every player is getting a different piece of a growing pie. While the term “market share” implies a zero-sum gain, losing market share does not automatically equate to a loss in revenue growth or a loss in share values.
That being said, relative stock performances will be affected by the relative market share changes. Therefore, in addition to standalone valuations on individual stocks, I incorporate the information of market share changes, which will enhance the accuracy of the stock valuations. Since Intel’s loss on CPU market share will be mainly AMD’s gain, the near zero-sum gain nature suggests that Intel CPU revenue growth will be negatively correlated to AMD’s. Consequently, their stocks should be valued in a way consistent with the tradeoff of their revenue growth.
Using the same logic, NVDA’s gain on GPU market will be AMD’s loss. NVDA should be priced fairly to AMD, relative to their respective GPU growth rates. In all likelihood, as the relationship between any two stocks should be determined by the relative growth rates, three relative valuation relationships should be held up simultaneously.
Revenue Estimate Revisions
Since we are not the only ones following the news, it stands to reason that professional analysts also revised their estimates accordingly. For example, after last earnings call, NVDA’s 2017 annual revenue estimate has been raised from $8.09 billion to $8.28 billion (Bloomberg). Over the same time period, AMD’s revenue estimate has been revised upward from $4.73 billion to $4.82 billion, and Intel’s revenue estimate has been raised from $59.9 billion to $60.2 billion.
One interesting observation is that the similar timing of analyst revisions on three stocks may be a result of the recent tech sector move. It is also interesting to notice that upward revisions were often accompanied by simultaneous stock price increases for NVDA and AMD, suggesting that current stock prices may have factored in the expected revenue increases.
How to Value Stocks under Zero-Sum Gain?
The short answer is that AMD will not have positive profit in the near future, so the valuation cannot be on profitability. The long answer is that shareholders of these three players have been more focused on the new product market share of revenue; stock prices have not reacted to the earnings outlook. For that reason, I need a valuation model that relates expectation of revenue growth to stock prices. The standard Sales Franchise Value Model (SFV) is used because of its emphasis on revenue growth.
On the relative valuation of these three stocks, I need to stress that the three paired relations should be simultaneously held in order to ensure the zero-sum nature of the gain. For details, please see here. In Table 2, the current estimates along with the same tests performed two months ago are compared:
- If there is a three-way valuation relationship among INTC, AMD, and NVDA, that relationship will require that INTC’s fair value is $37, AMD’s fair value is $15, and NVDA should be traded at $195. This suggests that both AMD and NVDA are undervalued at the current market levels, and that INTC is fairly valued. Note that AMD’s fair value at $15 has not changed after the Deal.
- If stocks were trading at these fair values, NVDA’s $195 has priced in a 14% annual revenue growth rate. AMD’s $15 has factored in 13.5% revenue growth rate, and INTC’s $37 reflects a 3.6% revenue growth rate.
- NVDA’s 14% implicit growth rate is far below analysts’ estimate of 19.8%. This suggests NVDA has a significant upside beyond the $195 price target, if the actual revenue goes close to street estimates.
- AMD’s 13.5% growth rate is in line with analysts’ consensus estimate of 14%. There is less of an upside over their $15 fair value, compared with that of NVDA.
- As of today, AMD is 12% undervalued. Nvidia is 8% undervalued.
Finally, valuing three stocks simultaneously is superior to valuing three stocks independently, since the nature of zero-sum gain among AMD, Intel and Nvidia provides additional information. The fact that the revenue growth rates estimated from the market prices come remarkably close to the independent analysts’ consensus estimates further validates the price targets.
Disclosure: I am/we are long NVDA.
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.