Shares of Advanced Micro Devices (AMD 2.37%), the longtime rival to Intel (INTC -9.20%) in the PC processor market, plunged following the company's earnings report. The actual second-quarter numbers weren't terrible relative to expectations -- coming in line on revenues and missing by a penny on the earnings per share line -- but the guidance fell significantly short of expectations.
Intel's Bay Trail appears to be taking low-end share
AMD's computing solutions group saw a 20% year-over-year decline in sales. AMD noted that it sold a richer mix of products (which implies higher average selling prices), implying that units were down in excess of 20% year over year.
This is share loss to Intel's low-cost Bay Trail-M and Bay Trail-D notebook and desktop products, respectively. While Bay Trail-T (the tablet variant of Bay Trail) drove a bill of materials too rich for very cheap tablets (driving a need for contra-revenue offsets), it faced no such issues for low-cost PCs. Further, the cost structure appears good enough that even with these processors driving 60% of Intel's low-end PC chip mix (and 20% of total notebook mix), Intel is reporting eye-popping gross margins (64.5% in the most recent quarter; 66% guided for the current quarter).
AMD's Lisa Su even addressed this on the company's conference call:
So if you take a look at U.S. specifically about Bay Trail and we certainly see Bay Trail no question about it, we see them at, very, very low entry price points going up into the mainstream of the notebook and desktop market. And there are places where we chose not to compete because it's just not profitable business.
This appears to be an admission that AMD doesn't have the cost structure to defend itself against Intel's 22-nanometer Bay Trail parts in the low-cost PC market. Given that Intel is projecting that costs for its next-generation 14-nanometer Celeron/Pentium products due in 2015 will be lower than this year's Bay Trail-M/D, and given that the cost per transistor at the 20-nanometer foundry nodes (which AMD will be beholden to) go up, it's difficult to see AMD improving its competitive positioning here.
Source: Intel.
Graphics and console business not driving as much upside as thought
Investors seemed to be expecting a large uptick in game console chips during the second half of the year. However, AMD noted that these chip sales will be spread more evenly across the year (which helped Q1/Q2, but will mute a potential ramp into Q3/Q4). Further, AMD is calling for game console chips to peak in Q3 in contrast to last year, when AMD saw a sequential increase in game console chip sales in Q4. This was likely a surprise to investors/analysts, and now models are being updated to reflect this Q3 peak.
The company also saw discrete graphics card sales (particularly for the enthusiast/add-in-board market) slump this quarter as a result of a flood of used cards hitting the market as demand from crypto-currency miners softened (likely as a result of dedicated chips for mining hitting the market).
Is there hope?
At this point, we know what went wrong, but is there any reason to believe in the company's turnaround? There's hope that eventually the semi-custom and embedded businesses can offset the PC decline, but that remains to be seen given that the PC business is still a huge part of AMD's revenues (and it's likely that Intel will continue to gain share).
There's also the possibility that AMD could gain share on NVIDIA in discrete graphics, but it would be unwise to underestimate NVIDIA's competitive positioning. Finally, there's a chance that ARM-based servers could gain traction and that AMD would be well-positioned to benefit, but that -- again -- requires going up against behemoth Intel.
In short, there's a whole lot of uncertainty surrounding the AMD story, and the post-earnings share price drop reflects that.
