During the past few years, Advanced Micro Devices' (NASDAQ:AMD) strategy has not made a lot of sense. The company has been trying to compete with Intel in both the PC and server chip markets, while at the same time facing off against NVIDIA in the GPU market, only to find itself losing market share across the board. On top of that, AMD has launched a variety of ambitious initiatives, despite the fact that its research and development spending has been slashed in each of the past three years.

Earlier this month, AMD held its 2015 Financial Analyst Day, and it seems that the company is finally taking the proper steps to narrow its focus and turn the business around. There are still no guarantees, and returning to profitability will be a difficult task, but it looks like AMD is finally doing something right.

Picking its spots
During its financial analyst day presentation, AMD listed four recent challenges: a weak PC market, channel inventory, market share losses, and investments not being sufficiently focused. It is the last point that has caused most of AMD's problems. The company has been throwing its limited resources behind all sorts of initiatives, and this lack of focus has allowed the competition to continually win market share.

About a year ago, AMD announced Project SkyBridge, an initiative to create a single platform for both x86 and ARM CPUs. Project SkyBridge is now dead, and the company's custom K12 ARM core has been delayed to 2017. These decisions are likely for the best: Project SkyBridge was too ambitious for AMD, and with the ARM server market still in its nascent stages, rushing a custom ARM core to market makes little sense.

Along with scaling back its ARM ambitions, AMD is exiting its SeaMicro dense server business. AMD acquired SeaMicro in 2012 for $334 million in an attempt to take advantage of the growth in micro-servers. However, as CEO Lisa Su stated during the latest conference call, the market has not developed fast enough, and AMD's core competency is processors, not systems. That is $334 million wasted and a costly lesson learned.

The most important part of its new strategy isn't where the company is focusing but where the company isn't focusing. AMD is not going after the smartphone market or the low-end tablet market, and it is reducing its exposure to the low-end of the PC market. And while the Internet of Things is expected to be a huge opportunity in the coming years, the company will not be putting its chips into IoT endpoints such as smartwatches and smart appliances.

These decisions are the right ones. There is simply no way that AMD can compete in these markets against Intel and companies making ARM chips. Intel barely has a presence in smartphones today, and it had to heavily subsidize its tablet chips to gain market share. That is despite Intel spending billions maintaining its state-of-the-art fabrication facilities and outspending AMD on research and development by more than a factor of 10.

The plan to win market share
The key to the new CPU strategy is Zen, the new high-performance core expected to launch in 2016. AMD is now focusing on the high-end of the PC market. With Intel currently facing essentially no competition there, it is likely the best opportunity to win back some PC market share. Zen is also the key to AMD's server strategy, and the company hopes to gain server market share starting in 2017.

With Zen, AMD has switched to using simultaneous multithreading -- what Intel calls Hyper-Threading. In its previous CPU architecture, AMD used a different form of multithreading, and this contributed to its lackluster performance compared to Intel chips. Zen should help close the performance gap.

In the GPU market, AMD has been losing market share to NVIDIA for the past few quarters, but the company plans to launch new GPUs in the next few months. These will be the first to use high-bandwidth memory, which should give the new cards an advantage when it comes to gaming at 4K resolutions. NVIDIA will likely launch some new GPUs in response, so there is again no guarantee that AMD will be able to regain ground in the market.

AMD is still going to have a tough time winning market share in the CPU and GPU markets, but for the first time in years, its strategy make sense. The company no longer appears to be flailing, pushing all sorts of initiatives and trying to compete on multiple fronts. Instead, AMD is starting to focus on the areas where it has the best chance at being competitive, and that should be a welcome development for investors.