Shares of graphics-chip company NVIDIA (NASDAQ:NVDA) soared in 2015, driven by market-share gains and positive financial results that diverged from the slumping PC market. The company's core gaming business is performing well, while growth opportunities in the automotive and enterprise markets provide the potential for higher revenue and profits going forward. Here's a look at three numbers that will important for NVIDIA in the coming years.

NVIDIA has been the leader in the discrete graphics card market for years, but its unit market share hovered around 60% from 2010 through the middle of 2014, leaving rival Advanced Micro Devices (NASDAQ:AMD) with a respectable 40% share of the market. This changed toward the end of 2014, when NVIDIA launched the disruptively priced GTX 970 and GTX 980, leaving AMD with no real competitor at the high-end of the market. During the third quarter of 2015, NVIDIA's market share reached 81.1%.

Data from Job Peddie Research.

AMD was able to stop the bleeding by launching new high-end graphics cards in the middle of 2015, but the launch was lackluster at best, leaving NVIDIA with the lion's share of the GPU market. These market share gains helped NVIDIA report positive results last year despite weakness in the PC market, leading to the stock rising 64% in 2015.

This year, both NVIDIA and AMD plan to launch new GPUs, and both companies are promising substantial gains in performance and efficiency. After being stuck using a 28nm manufacturing process for years, both NVIDIA and AMD are moving to 14/16nm FinFET manufacturing processes, and both will be incorporating the second generation of high-bandwidth-memory.

NVIDIA's massive lead over AMD, as well as AMD's string of quarterly losses that are straining the company's finances, gives NVIDIA a major advantage. While big market share gains are unlikely for NVIDIA in 2016, the company has a good chance at holding on to the bulk of its market share gains from 2015.

233 million
One of the major opportunities for both NVIDIA and AMD in the coming years will be virtual reality. The Oculus Rift is set to be released this year, and while the cost of headset will be too high for most people, this first-generation device could be just a taste of what's to come.

The Oculus Rift will cost $599, on top of requiring a powerful PC with a minimum price tag of about $1,000. This PC requires a powerful graphics card, an NVIDIA GTX 970 or an AMD Radeon 290 at minimum, in order to process the 233 million pixels per second that the device churns through. That's nearly twice as many pixels per second as a 1080p screen running at 60Hz requires, so it's no wonder that a low-end graphics cards won't be enough.

Source: Oculus VR.

The high price of the Rift makes the device unlikely to drive increased sales of graphics cards, but it could cause some gamers to shift from lower-end cards to higher-end cards, potentially improving the mix and profit margins for both NVIDIA and AMD. If VR takes off and becomes more affordable, graphics card sales could get a boost in the long run, although at this point it's still too early to tell whether VR will truly go mainstream. If it does, both NVIDIA and AMD stand to benefit.

$836 billion
The National Highway Traffic Safety Administration estimates that the total amount of economic loss and societal harm caused by automobile accidents each year is $836 billion. What does this have to do with NVIDIA? The company's Drive PX platform aims to bring advanced driver assistance features to cars, with the eventual goal of providing the processing power behind the self-driving car of the future. This number could be reduced dramatically in the coming years, driven by technology like that of NVIDIA.

Source: NVIDIA.

NVIDIA is far from the only company scrambling for its place in the self-driving car of the future, but its graphics technology is well suited for the task. Drive PX is a machine learning system, able to detect and track large numbers of objects from camera and sensor inputs by first training on an enormous quantity of data. Once objects are detected, a plan needs to be developed, and the 8 TFLOPS of computing power contained within NVIDIA's second-generation Drive PX system should be up to the task.

I think that a mainstream self-driving car is still years away – it's a really difficult problem to solve. However, advanced driver assistance features enabled by the kind of processing power and software that NVIDIA's Drive PX provides are a step in the right direction. ADAS systems have yet to widely adopted, creating a major growth opportunity for NVIDIA.