As far as I can tell, Nvidia (NVDA), a leader in graphic computing chips, is in a death struggle with its old competitors AMD and Intel (INTC), as well as a slew of smartphone and tablet chip competitors. Yet the main takeaway I got from CEO Jen-Hsun Wong in the May 9 analyst call was that NVDA does not need all of its cash to compete. (See Nvidia Fiscal Q1 2014 Earnings Call Transcript.)
In the short run returning cash to shareholders should prop up the stock price. In the long run, if the cash should have been used for R&D, the move could hurt shareholders. There are several variables that need to be looked at here: cash on hand; current margins; cash flow; the current valuation of the stock; and the need for cash to counter competitive threats.
First some background. Nvidia, along with former rival ATI (now part of AMD), made graphics chips a separate business in the PC market. AMD absorbed ATI because it was clear the trend towards SoCs (System on Chips) meant the sequential CPU and parallel graphics component, or GPU, could be on the same piece of silicon. Nvidia in theory could have tried to compete by licensing x86 CPU technology from Intel and AMD, but would have had no clear competitive advantage in the space. Instead Nvidia decided to combine its graphics IP with an easier to license CPU technology, ARM, that addressed the low-power portable device market.
Currently the bulk of Nvidia revenue still comes from discrete graphics chips that work with Intel or AMD CPUs. In the ARM space Nvidia has had a number of victories in tablet computers with their Tegra line, notably Google's (GOOG) Nexus 7. It is also about to compete with smartphone chips, variations on Tegra.
Looking at first quarter numbers (quarter ending April 28, 2013), I see no evidence that revenue should make Nvidia confident. Q1 is typically seasonally down across sectors for Nvidia, as its end market is largely consumer-oriented. Revenues were $954.7 million, down 14% sequentially from $1,106.9 million. The y/y growth is uninspiring: up 3% from $924.9.
When looked at against the backdrop of a declining PC market, 3% looks better. AMD's Q1 revenue was down 31% y/y, while Intel saw a decline of 7%.
Clearly Nvidia benefited from entering the tablet/ARM space. It appears to have gained market share over AMD in the discrete GPU space, with gaming revenue up 24% y/y.
The Tegra mobile processor line revenue was down 50% sequentially and 22% y/y, but that is because Nvidia moved up the transition to Tegra 4 and Tegra 4i, the smartphone version. OEMs have ramped down production of soon-to-be obsolete Tegra 3 devices. A meaningful ramp of Tegra 4 revenue is expected in Q3.
Profit margins were good in the quarter. GAAP net income was $78 million, cash from operations was $176 million, and free cash flow was $110 million. Cash and equivalents ended at $3.7 billion, with no debt. $100 million was used for stock repurchases in the quarter, and $46 million was paid out in dividends.
The dividend rate is currently near 2% ($0.075 per quarter). The stock went ex-dividend May 21. Market capitalization is $9.2 billion.
The announced plan is to return over $1 billion to shareholders this fiscal year. Unless the dividend is changed, that means under $200 million in dividends and over $800 million in share-buy backs. That would likely leave some $3 billion in cash.
If Nvidia were in steady-state, I don't think the buybacks would be harmful, given the amount of cash that would be left on hand. I don't think buybacks are a particularly great investment; it is not like P/E is in the single digits (it is 16.0 right now). Acquisitions could be more attractive, depending on specifics.
If management's confidence is justified, then the buybacks do make sense. If Tegra 4 is going to greatly accelerate profits, and outside investors don't see the value, then even current long-term investors can get a great deal of benefit from doing the buybacks now, before the new information is incorporated in a higher stock price.
For me the critical factor is R&D spend. In the latest quarter Nvidia spent $327 million on R&D. Intel spent $2.5 billion, but has many more products to support. AMD spent $312 million. [All GAAP numbers]
In the short run AMD is the main competitive threat to Nvidia due to direct competition in GPUs for gamers and other high-end users. Already, however, in the x86 space both AMD and Intel are selling processors with increasingly sophisticated graphics integrated on the chip. This means less people over time should need any kind of discrete GPU card or chip in their computer at all. There will always be a high end, especially since GPUs are increasingly used for software outside of the graphics domain, but it looks like we should plan for a declining GPU attach rate to mid-range and lower end PCs. This trend will probably play out in the next 4 to 6 years, at which point even high-end PCs should not need discrete graphics cards.
In answer to an analyst question on the conference call Jen-Hsun Wong said that tablets are disrupting the low end of the PC market, not the high end. To do serious work efficiently you need a large screen and fast graphics processing. That is the end of the market Nvidia is already serving. Wong also dismissed AMD's win of the next generation of gaming consoles, saying they constitute a small market and are not liked in China. I think Wong is wrong on that point; it is a serious threat to Nvidia (See AMD Game Console Triple Crown).
But just as we saw PCs replace minicomputers in the 1980s, and GPU cards replace specialized graphics computers in the 1990s, the process of miniaturization will mean that Nvidia's only workable long-term solution will be more sophisticated ARM CPUs on chips combining with its high-end GPU technologies. Then it will be an ARM vs. 86x battle.
Nvidia has a lot of cash (but not as much as Intel) and a foot in the door in the mobile & ARM space. The problem is that anyone can license ARM, and reasonably powerful GPU designs are also available. The same is true for the wireless part of the smartphone SoCs that will be coming out. The competition is intense: Samsung (SSNLF.PK), Apple (AAPL), Qualcomm (QCOM), TI, Marvell (MRVL), and a handful of up-and-coming Chinese competitors, among others.
I don't buy the thesis that ARM is automatically going to win out in the long run over the 86x architecture. I certainly could not run my business PC on an ARM chip today. Innovation cuts both ways. ARM was a low-power, and low-computation, architecture that worked for the first round of smartphones and tablets. It can become more computationally powerful, but not without consuming more power. In the last few years some variations of x86 chips have become more power efficient, so that we have begun to see tablets running on them. At 20 nm process technology they should be competitive for tablets and possibly smartphones. That means in 2014, not far away.
So I see the cashing out as hedging a bet. Tegra 4 might be the magic device that brings Nvidia stockholders untold wealth, but it could be just a holding action against the other ARM chip makers. If 2014 is the year that discrete GPUs sales start a substantial decline and Tegra loses market share, then cash use becomes a much more critical issue.
I'm not calling it either way. I have been very impressed by Nvidia management and engineering over the years. But Nvidia is going up against other well-managed, well-funded companies with great engineering teams.
One final data point. Qualcomm's R&D budget in the latest quarter was $1.2 billion. Over 3 times what Nvidia spent.
On the conference call CEO Wong's response to a question about smartphone chip competition was that the market is still doubling each year. So despite its dominance by Apple, Samsung, and Qualcomm (and Marvell in China), there is still room for niche players. That may be, but in my experience after the growth slows down the industry consolidates to just 3 or even 2 profitable players. If I were running Nvidia I would worry more about the competition than the stock price. I'd leave the dividend or maybe up it a bit and put the cash to use in R&D, marketing, and strategic acquisitions.
In writing this I convinced myself not to buy NVDA at this time at this price (I have owned it in the past and did well with it). I would reconsider at a lower price point, or at this price if Tegra 4 takes a larger market share than I expect it will, or if AMD (which I own) falters with its APU and GPU introductions.