Advanced Micro Devices (NASDAQ:AMD) posted a significant revenue gain of 9% y/y for Q2 to $1.027 billion. AMD posted a slight operating loss of $8 million, which was a considerable improvement over the $137 million operating loss of a year ago. This is being hailed as a sign of AMD's imminent turnaround. I remain unconvinced.
Let me first start out with a disclaimer concerning my financial interest in AMD. This is something I do occasionally when it's apparent that the standard disclosure at the bottom of the article is being frequently ignored in comments.
I have no interest in AMD, and I am not shorting the stock. The last time I had any interest in AMD was in 2013 when I had a long position. I have often seen it alleged in comments that I am shorting the stock, apparently in order to imply that is why I take a negative view. I have yet to short the stock.
My negative view of AMD is largely driven by the overall competitive landscape in which the company functions and the market dynamics related to personal computers. As I made clear in my recent article on Intel (NASDAQ:INTC), I see the traditional PC market as being in decline. This has been driven by the growth of mobile computing devices, specifically the smartphone, which have taken over many of the functions of the traditional PC.
As I state in the Intel article, INTC's failure to penetrate the mobile market I believe to be due to the fundamental cost advantage that ARM architecture processors have over x86 architecture. Eventually, I believe that ARM architecture devices will take over all of personal computing as their capabilities become increasingly suitable for traditional desktop and notebook applications. I expect ARM to take over the data center as well. This process of ARM architecture taking over traditional x86 markets could take many years.
My opinions about the x86 computer market have necessarily shaped my views of AMD. Following the lack of market success for its ARM Seattle server processor, AMD's management decided to focus on x86 for server and PC applications. If I don't believe there's a future for Intel with x86, why would I think there's a future for AMD? AMD's goals of increasing market share in PCs and servers amount to fighting for a larger piece of an ever shrinking pie.
Intel appears to be committed to x86 for the foreseeable future, and I felt that was probably more bad news for AMD, given that INTC is the stronger competitor. Likewise, Nvidia (NASDAQ:NVDA) is the larger and more profitable competitor in graphics processors. I have compared AMD's competition with its stronger rivals as equivalent to fighting a "two-front war."
A Closer Look at Q2
That's the broad-brush context with which I view AMD. Now, let's look at some of the specifics in its earnings report. There are some insights to be gained from the segment information. Computing and Graphics segment revenue grew y/y by 15% to $435 million, but actually declined sequentially from the March quarter revenue of $460 million. Losses deepened sequentially as well to $81 million from the March loss of $70 million.
I see this a confirmation of my basic take on the impacts of the RX 480 and Bristol Ridge launches. Following the Computex launch, I thought Bristol Ridge was dead on arrival. Bristol Ridge, fabricated on an antiquated 28 nm process, couldn't hope to be competitive with Intel. It's still dead, regardless of what AMD's management may want to represent.
The botched release of the RX 480 may not be reflected in the Q2 financial results, but certainly the results appear to confirm my take that the pricing of the RX 480 meant that it was not going to contribute meaningful operating income.
Why did Computing and Graphics revenue decline sequentially in Q2 despite the new product introductions? My view is that this was primarily due to the advent of Nvidia's GTX 1080 and 1070 Pascal architecture graphics cards. Numerous reviews showed them to be superior to AMD's high-end offerings such as the Fury X as well as the lower-performing RX 480. Once Nvidia's Pascal cards hit the market, AMD lost the high end. That loss of revenue was not compensated by the RX 480 release.
AMD also appears to have lost GPU market share in the process. Unlike the prior quarter, AMD did not claim to have gained GPU market share during the Q2 earnings conference call. I think it's likely that AMD lost GPU share overall, although it claimed to gain share in professional GPU applications.
Last week, Wells Fargo claimed that AMD offered "renewed competition" to Nvidia following the RX 480 release. This is rather typical of the kind of ignorance technology company analysts display about technology.
AMD is in a worse competitive position relative to Nvidia than before the companies released their new Pascal and Polaris architecture devices. Prior to those releases, both companies were fabricating their devices on equivalent process nodes (28 nm), and both companies had reasonably competitive graphics cards in all price segments.
That is completely changed now. AMD has nothing competitive with Nvidia except for the RX 480 at the ~$200 price point. Furthermore, AMD is unlikely to have anything competitive with Nvidia above the $200 price point until Vega is released, which we were informed at the conference call will not be until next year.
Likewise, investors were informed at the conference call that volume availability of Zen microprocessors would not begin until 2017. Confident predictions that Zen would be available in late 2016 have proven incorrect.
That leaves AMD's Semi-Custom segment to carry the ball for the rest of the year. AMD did see some growth in this segment in Q2 of 5% y/y to $592 million as it got an early start on the new slimmer Microsoft (NASDAQ:MSFT) Xbox One S. Operating income grew a very substantial 211% to $84 million.
Combined with the normal seasonal uptick in consoles for Q3, it appears that Semi-Custom segment will grow y/y in Q3 as well. AMD is expecting another non-console semi-custom win in Q3, and then there are the new consoles that will come out in 2017, Scorpio and Sony's (NYSE:SNE) Neo. These probably don't affect revenue this year, however.
It was the Semi-Custom segment income of $84 million combined with the $150 million realized from the ATMP joint venture that put AMD in the black this quarter, with net income of $69 million. Without the ATMP income, AMD would have had a pre-tax loss of $49 million. So, it's really not clear that AMD's Q2 profitability is sustainable.
Despite the seemingly inevitable jump in stock price post earnings, my view of the intrinsic worth of AMD is unchanged. In graphics, its competitive position has gotten weaker, not stronger, and it has likely lost market share as a result. It's likely to lose more to the better, faster and lower power GTX 1060 graphics card, which is priced comparably to the RX 480. I expect to see y/y declines in revenue and increasing losses in the Computing and Graphics segment through the end of the year.
Enterprise, Embedded and Semi-Custom segment is more uncertain. It's expected to rise seasonally in Q3, but AMD wasn't providing much more guidance than that. AMD's guidance was for an 18% sequential revenue increase on the strength of semi-custom and graphics products. I believe this is probably over optimistic to the extent it's predicated on the graphics business, but I'm sure that AMD can count on the semi-custom revenue its expects.
Although semi-custom is a bright spot, I still consider the prospects for the Computing and Graphics segment to be poor. Given the unrealistic expectations for growth in the Computing and Graphics segment among AMD supporters, I still rate it a short opportunity, but that opportunity probably doesn't arrive until after Q3. Current short interest is very low, understandably.
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.