Back in May, I wrote that Nvidia (NVDA) should reward shareholders with a big share buyback. At the time, I pointed out that Nvidia's cash stockpile had been growing rapidly over the past decade, while the share count has increased by roughly 30% since 2004 as a result of stock options. The company initiated its first share buyback in 2005, and repurchased $1.46 billion of stock between then and the end of 2008. However, since that time Nvidia has almost entirely stopped repurchasing shares; the company still has $1 billion remaining in its total repurchase authority of $2.7 billion.
When asked by an analyst on the Q2 (August) earnings call about Nvidia's plans for its cash, CEO Jen-Hsun Huang stated, "The only plans that we have ready to announce right now is to make a lot more cash." Needless to say, as a shareholder I was not particularly happy with this answer. If the company needed to maintain over $3 billion in cash for working capital purposes or to make major investments, this might be a reasonable stance. However, the vast majority of Nvidia's cash is "excess"; it sits in bank accounts, money market funds, etc. Given the prevailing low-interest rate climate, maintaining a high cash balance is not really helping Nvidia to make more cash.
Today, cash and short term investments represent a large portion of Nvidia's valuation. As of Monday's closing price at $11.92, Nvidia had a market cap of $7.4 billion. However, the company also held $3.4 billion in cash and investments, representing roughly 46% of the company's value. Excluding cash, Nvidia's shares are worth approximately $6.40, which seems extremely low at only 8X TTM earnings. Whereas Nvidia's most direct competitor, AMD (AMD), is struggling mightily, and Intel (INTC) is expected to post declining revenue and profit for FY12, Nvidia showed year over year growth in revenue and profit for the just-ended Q3 and is expected to repeat that achievement in Q4. At face value, it would seem that Nvidia should be trading at a premium to those competitors.
I believe that a major reason why Nvidia is not being rewarded for its strong operating performance is that investors are discounting the value of the cash on its books. This is entirely reasonable, as $3 billion that might not be returned to shareholders for ten or even twenty years has a present value well below $3 billion. Therefore I was very encouraged to see Nvidia take a step in the right direction last week. The company announced that it will initiate a quarterly dividend of 7.5 cents per share, and will extend the share repurchase authorization to the end of 2014. While this decision does not obligate Nvidia to repurchase any shares, it was very good to hear management affirm its commitment to returning cash. Given that the company had never given an inkling that it was considering a dividend, I think there is a reasonable likelihood that management will begin repurchasing shares again next year. Given the very low enterprise valuation (i.e. stock multiple excluding cash), this should give the stock a significant boost in 2013.
Ultimately, I would like to see management increase the share buyback with the aim of pushing the share count back down to 500 million or less. While Nvidia should maintain a solid cash position to cover the cost of any acquisitions and to permit continued investment in the event of a downturn, $1.5 billion would be more than sufficient for these purposes. Nvidia's most recent major acquisition was Icera, a company which specializes in baseband processors (wireless modems). This filled the key gap in Nvidia's mobile portfolio, as Ashraf Eassa recently pointed out, by allowing Nvidia to create integrated modem/application processors. Yet this acquisition only cost $367 million. It is unlikely that Nvidia would need to spend more than that amount for any future acquisitions in the mobile space. Moreover, the company has a steady source of income for the next few years from its patent cross-licensing agreement with Intel, worth over $200 million annually. This significantly reduces the likelihood that Nvidia would need larger cash reserves to offset negative cash flows.
As I have discussed elsewhere, I think that Nvidia's fundamentals are very strong. The discrete GPU business continues to take significant share from AMD due to the superiority of the Kepler architecture for most graphics tasks. Moreover, the company is likely to release an integrated application/baseband processor (codenamed "Grey") within the next year. This will vastly increase the TAM (total addressable market) for Nvidia's Tegra line of SoCs because phone-makers increasingly favor integrated SoCs of this sort.
To date, Qualcomm (QCOM) has been a huge beneficiary of this trend, as essentially the only company able to provide such a solution. Many bearish commentators have noted that Nvidia has been frozen out of the two top smartphone/tablet vendors, Apple (AAPL) and Samsung (SSNLF.PK), as these two companies design their own SoCs. Yet most Samsung smartphones actually use Qualcomm's Snapdragon integrated SoC. Samsung clearly prefers having the integrated solution over using its own standalone SoC and a standalone baseband processor from Qualcomm or another vendor. I believe that Grey will be competitive with Qualcomm's integrated solutions and will thus give Nvidia an opportunity to get some of the design wins at Samsung (as well as at other, smaller vendors). As long as Grey is reasonably competitive with Qualcomm's Krait, Nvidia is likely to take some share from Qualcomm, as handset makers will not want to become overly dependent on a single supplier.
In summary, I think Nvidia's fundamentals are strong, and the company's low valuation is at least partially the result of investors discounting the cash on Nvidia's balance sheet. By initiating a dividend and extending its share repurchase program, Nvidia is indicating that it may be ready to pursue a more shareholder-friendly capital allocation policy. I would therefore advise buying on weakness. If Nvidia does resume share repurchases, I think the stock could reach $18, which is my current estimate of its intrinsic value.