Advanced Micro Devices (AMD) has a long history of being number 2 in two semiconductor domains: versus Intel (INTC) with CPUs and versus Nvidia (NVDA) with GPUs (graphics processing units). Recently AMD has conquered a sub-domain, custom CPU/GPU combo chips for game consoles (Xbox One, Playstation and Wii). At the same time traditional desktop and laptop PC sales have gone into a global decline, which affects all three companies. Other opportunities have arisen in smartphones and tablets, but so far those segments have been dominated by players like Apple (AAPL), Samsung (GM:SSNLF) and Qualcomm (QCOM).
Investors have treated Intel, Nvidia, and AMD as losers in 2013, while much of the rest of the stock market has roared. This article will focus on the AMD vs. Nvidia rivalry and what it means for investors. Intel will be brought in because of its importance in the CPU/GPU market, but will not be a focus.
A few key numbers will be referenced to in this article. I'll use the third quarters of 2013 and 2012, with the caveat that Nvidia is on a fiscal year, with the quarter ending in October rather than September. Given that consumer chips ship in seasonally high volumes in October and November, this skews the Nvidia numbers somewhat.
AMD had revenue of $1.46 billion, up 15% y/y from $1.27 billion. Nvidia had revenue of $1.05 billion, down 12.5% from $1.20 billion year-earlier. On the face of it AMD is now doing much better than Nvidia in dealing with growing revenue against a backdrop of flat to decreasing PC markets.
For GAAP net income, AMD had $48 million, which was a welcome improvement on the loss of $74 million in Q3 2012. Nvidia GAAP net income was $118.7 million, which was down 43% from $209 million in Q3 2012. So Nvidia is making considerably better profits on its revenue, but the one-year trend favors AMD, with net income going in the right direction.
As to non-GAAP net income, keep in mind that AMD and Nvidia don't necessarily use the same criteria for exclusions from GAAP. AMD non-GAAP net income was just $31 million, but was up from negative $150 million in the year-earlier quarter. Again, Nvidia did better in the quarter, but was trending the wrong way, with non-GAAP net income of $153.8 million, down 37% from $245 million in Q3 2012.
Finally, the ability to carry on a head-to-head competition depends partly on cash reserves. AMD ended the quarter with $1.2 billion in cash, but long term debt of $2 billion. Nvidia ended with cash and equivalents of $3 billion and no debt.
One way to sum up these numbers is to recognize that Nvidia has done better in the past. AMD's past performance has been hammered by its head-to-head strategy against Intel in CPUs.
In technology yesterday's winners and losers can switch roles in almost any future time frame. Entire businesses, like mini-computers, can collapse, while others arise, like social cloud companies. The semiconductor segment has been somewhat more stable, but the competition is fierce. The future is hard to see (though the winners usually claim they saw it all along).
At this moment it seems that desktop computers are losing some of their mass appeal, but will remain around for power users, including high-end gamers. The computational power of an AMD A10 processor or an Intel processor combined with a graphics card is far beyond what can be got from a chip powering a smartphone, but all that power is not needed just to download some communications from the cloud.
AMD is playing in a broader spectrum of end-markets than Nvidia, which has no x86 CPU instruction set presence. Nvidia's main products are GPU chips for graphics cards and ARM + graphics chips for smartphones and tablets. The GPU chips go to several end-markets: consumer (including mainstream and high-end gaming); professional workstations (design, video production, science); and server/cloud/HPC (high performance computing). Except for its own demonstration Shield handheld gaming device, Nvidia has been muscled out of the gaming console business.
AMD lags Nvidia by quite a bit in workstation and server/cloud graphics market share. AMD has been most competitive in consumer graphics. AMD is trying to leverage its new-found advantage with consumer consoles into PC gaming. This is mostly software driven. If you looked at almost any head-to-head comparison between gaming cards (at Anandtech, for instance) that are in the same price range, in the past some games would run faster on an Nvidia card, while others did better on an AMD card. In general, but not always, Nvidia had some kind of advantage overall. But that advantage was not always from the design of the GPU chips; it was often because Nvidia had more market share, so more companies chose to optimize to the Nvidia chips.
With most games available both on consoles and PCs, gaming companies have to decide how to approach the optimization question, given the changes taking place now. Ideally, there are switches in the software so a game will run optimally on a console , on an Nvidia card, or on an AMD card. But optimization takes time and money, so many games will play sub-optimally on one of these devices. A lot of consumers won't care, but the serious game players, the kind who buy $500 gaming cards, will. The incentive system has shifted: now it makes sense to optimize for AMD, which covers both consoles and AMD cards.
Nvidia management claims they are not worried about AMD's console victory and its halo effects. Nvidia does have options. The simplest is to discount the consumer GPU card prices. They announced major price cuts recently to bring their price/performance ratios more in line with AMD's latest GPU offerings. They could go even lower, effectively giving up profits for market share. Given AMD's cash position, AMD would be hurt, but could last long enough that the greater concern for Nvidia management would be a shareholder rebellion. Plus AMD has a monopoly on console chips. I don't see the latest Nvidia price cuts as anything other than the normal cycle of discounting aging lines, then eventually bringing in higher-priced offerings again. They won't hurt Nvidia margins in the long run.
A better strategy for Nvidia, and one that is being pursued by both companies, is making it easier for gaming companies to optimize to Nvidia. This can even involve Nvidia staff helping out specific gaming companies. That is less costly than a price war and encourages loyalty. By extension, payments can be made to encourage staying with the Nvidia systems.
A third strategy is more partnering with Intel. Intel's on-chip graphics already use some Nvidia technology, resulting in $66 million in license revenue for Nvidia in the quarter. As SoC chips that combine the traditional CPU with a GPU become dominant (if they are not already), the GPU card is likely to decline in sales except for high-end gaming, professional workstations, and cloud systems. Nvidia is not likely to license x86 CPU technology from Intel, but could try to compete with ARM CPUs on SoCs with high-end graphics capabilities; Tegra on steroids could run a PC. In the meantime, Nvidia could try jumping down to Intel's 14 nm process like Altera (ALTR) to achieve a clear competitive advantage over AMD for pure GPUs.
Aside from their head-to-head battle in consumer GPUs, and its extension into workstation and cloud computing (where AMD's market share is so small it can't really go down further), Nvidia's fate probably depends on how well it does with Tegra 4, its smartphone/tablet chip, and future extensions of the Tegra line. AMD has been trying to compete in tablets, but not smartphones, with little traction, but could score with a lower-power consumption chip due out in the first half of 2014.
Tegra 4 is a good chip, and has seen some adoption, but the competition is fierce and Nvidia does not have a clear competitive advantage. Graphics design knowledge is not as arcane as it was a decade ago. ARM licenses both CPU designs and GPU designs; other companies license GPU designs that will work with ARM. Many competitors license ARM designs, and some, notably Marvell (MRVL), are capable of squeezing their own design improvements out of them. Everyone is constrained by the same process technologies (currently 28 nm going on 20 nm) except for Intel, which does not use ARM. Even AMD is preparing to use ARM, combined with its own technologies, for server chips. A number of Chinese companies are manufacturing smartphone chips using ARM, and Chinese phones may appear in the U.S. sometime in 2014.
On the positive side, a lot of the new Tegra 4 chips will be going into Xiaomi smartphones in China. It is not so much that Tegra 4 chips are not competitive, as that it is very difficult to predict which companies will survive when the industry consolidates. Samsung and Qualcomm yes; I don't see any other company as a guaranteed long-term survivor at this point.
In the short run, both AMD and Nvidia are positioned to survive. Both have initiatives underway that have the potential to lead to profit expansion in 2014 and beyond. Nvidia could be a smartphone winner, and AMD could win by extending their "semi-custom" chip expertise beyond gaming consoles. Either company might have initiatives in the R&D stage, or be eyeing acquisitions, that could also be game changers.
As an analyst and investor, I weigh risks (probabilities of future scenarios) against current market capitalization. At Friday's closing $3.50 per share, AMD had a market capitalization of $2.5 billion. Nvidia closed at $16.17 per share, for a capitalization of $9.4 billion.
Using GAAP net income, for Q3 the ratio of Nvidia to AMD is [119/48] about 2.5 to one. Market capitalization, Nvidia to AMD, is about 3.75 to one. So at current prices you can buy more profit at AMD than at Nvidia.
Sadly, we are not just buying Q3 net profit. AMD's cash situation, while not unhealthy, is considerably worse than Nvidia's. AMD has a much worse history of disappointing investors than Nvidia, though that is largely because of its competition with Intel, which AMD is now trying to minimize. Nvidia has also been using cash to buy back stock, and to pay a dividend [$0.085/quarter], which also helps keep its stock price up.
It is conceivable that both AMD and Nvidia could turn into losers if other companies win the race to move the ARM architecture up to PCs and servers, and graphics become commoditized. That would take several years, but it has to be seen as a possibility.
Both could also turn out to be long-term winners. For AMD, that would most likely come from taking graphics market share from Nvidia or taking server SoC server share from Intel. For Nvidia it would most likely mean outperforming the competition to become one of the top 3, or even top 2, smartphone chip makers.
Rather than choosing one company over the other, I would say AMD is more suited to risk-hungry investors and Nvidia is more suitable to conservative investors. At worst with Nvidia you have the dividends, the stock buy-backs, and probably at least 4 years of cash flow from the GPU business. At best you have market share gains from AMD and the pile of future smartphone profits.
At worst from AMD you have a failure in the traditional PC segment and market share losses to Nvidia. That would result in a lower stock price than at present, but my assessment is AMD's combined IP and gaming console contracts are worth the $2.50 per share. At best you have several opportunities that could build cash flows and restore investor confidence. They include AMD APUs gaining market share against Intel, GPUs gaining share from Nvidia, SoCs in new fields like industrial and commercial displays, and the new cloud server model chips.
I see no reason to set a target price for either company, given the great changes that will take place in 2014. Instead I will be looking for results quarter to quarter, hoping to spot trends.