For Advanced Micro Devices (NASDAQ:AMD), 2016 will prove to be a pivotal year. The company continues to struggle, posting a massive $660 million net loss in 2015, which followed a $403 million net loss in 2014. AMD has seen its share of both the CPU and GPU markets disappear, and extremely weak demand for PCs is only adding to the pressure.
A few major developments will play out this year in all of AMD's core businesses, and it's not hyperbole to say that the future of the company is at stake. With that in mind, here are three numbers that will be extremely important for AMD in 2016.
The only real success for AMD over the past couple of years has been its semi-custom business. AMD supplies semi-custom SoCs for all of the major game consoles, based on the company's CPU and GPU technology. This revenue and profit have prevented the company's finances from deteriorating even further.
But even with strong sales of game consoles, AMD receives less per chip as time goes on. During the fourth quarter, for example, semi-custom revenue slipped 15% year-over-year, while operating profit tumbled 46%. Game consoles will not be a source of growth until the next generation of devices is announced, which is likely years away, and that assumes that AMD remains the only supplier.
The good news is that AMD has announced a few additional semi-custom design wins, and the company expects to begin realizing revenue from these wins during the second half of this year. The first two wins were announced in October of 2014, and AMD expects to generate a total of $1 billion of revenue over the course of three years. A third design win was announced last year, but details are scarce.
This new revenue should help offset declines in console SoC sales, and depending on how quickly console revenue falls off, AMD may actually be able to grow the semi-custom business. With the company's PC business still flailing, the semi-custom segment provides a much-needed source of revenue and profits for the company.
During the second quarter of 2014, AMD shipped about 40% of discrete graphics cards, with rival NVIDIA (NASDAQ:NVDA) claiming the rest. Since then, AMD has seen its market share collapse, driven by both disruptive products from NVIDIA and a lack thereof from AMD. During the third quarter of 2015, AMD's market share had slumped to just 18.8%.
The company has taken some steps to improve the performance of its graphics business. Late last year, AMD formed the Radeon Technologies Group, which now oversees graphics development for all AMD products, including discrete GPUs, APUs, and semi-custom chips. This separation could make it easier for AMD to spin-off the graphics business at a later date, although there's been no indication that the company plans to take that step.
This year, AMD will have a chance at regaining some of its lost market share. The company plans to launch a brand new graphics architecture, Polaris, toward the middle of this year, and the company promises major boosts in performance and efficiency. According to AMD, Polaris is substantially more energy efficient than NVIDIA's current Maxwell architecture, opening the door for AMD's graphics chips to power laptops and other small form factor devices.
That said, NVIDIA will be launching new GPUs of its own this year, based on its next-gen Pascal graphics architecture. Pascal promises major performance and efficiency gains of its own, and at this point it's impossible to say whether Polaris or Pascal will prove superior. AMD doesn't need to blow NVIDIA away in order to win back some market share, but the company's execution needs to be better than it was in 2015.
Toward the end of 2016, AMD plans to launch the first CPUs based on its new Zen microarchitecture. Intel (NASDAQ:INTC) is currently as dominant as it has ever been in both the PC and server CPU markets, thanks in part to AMD's previous Bulldozer architecture being a major disappointment. In terms of single-threaded performance, Bulldozer fell short of comparable Intel products by a wide margin.
Zen is expected to fix this problem. AMD has stated that Zen will bring a 40% improvement in instructions per cycle, a measure of single-threaded performance, compared to Excavator, the latest revision of the Bulldozer architecture. This is a big leap, and if AMD can both hit this target and launch Zen-based chips on time, the company may have its first truly competitive high-end CPUs in years.
If an AMD turnaround is going to take hold, 2016 is the year it will happen. With new semi-custom revenue and a chance at winning back some lost GPU and CPU market share, the string of big revenue declines may finally come to an end. AMD expects to return to non-GAAP operating profitability during the second half of 2016, which would be quite the reversal from the same period last year when the company posted a non-GAAP operating loss of $136 million.
AMD will still post a net loss based on this outlook, given the $40 million in quarterly interest payments the company makes, but if AMD meets its goals, it would be the first true sign that the turnaround is making progress.