There's a lot of chatter out there about a potentially perilous long-term fate for NVIDIA (NVDA +1.19%). In one corner, you've got those who claim that the discrete GPU is essentially dead, because integrated solutions are getting more powerful. In the other, you have those claiming that NVIDIA's investment in its Tegra mobile chips is largely futile, since this space will probably be gobbled up by bigger, more powerful players in the semiconductor space.
While there is some legitimacy to these concerns, the bottom line is that no matter how you slice it, NVIDIA will be just fine.
The death of the discrete GPU is overstated
There's no doubt that integrated graphics solutions from AMD (AMD +2.23%) and Intel (INTC +2.17%) in the PC space are becoming faster with each successive generation. That's what Moore's Law buys you. While this will pressure the lowest-end discrete graphics solutions, particularly in notebooks, it's tough to be too worried here, since the fears are predicated on the basic fallacy of "good enough" in graphics.
The thing about computer graphics is that the demand for computing power is insatiable. Until PC games are indistinguishable from reality and can run flawlessly on integrated graphics -- today's PC games barely run on integrated graphics -- there will be a continued need for higher-performance discrete graphics chips. Further, in the professional/workstation space, where time equals money and better performance saves time, there will almost always be a total cost of ownership case for ultra-high-end professional GPUs.
On top of the "good enough" theory being a fallacy, it's also tough to ignore that while the low end of the PC market is struggling, the high end of the PC market, at least according to Intel, is actually growing. This is where NVIDIA's highest-margin GPUs are sold, which means that even if the low end of the PC market erodes nearly entirely and/or goes to integrated graphics, the vast majority of the margin dollars will remain intact courtesy of the high end.
Tegra is a win-win
There's no doubt about it -- the mobile processor business is currently a very difficult one for almost every player. Intel, NVIDIA, Broadcom, and many others are seeing an incredibly fierce pricing environment. Most, if not all, of these players are losing money. The hope is that with volumes come sufficient economies of scale to make these ventures profitable in the long run.
But the unfortunate truth is that this market, like all other lucrative tech markets in their infancy, will eventually see consolidation. Some companies will simply exit the business entirely to focus on other things. Others will acquire potent competitors with key technological edges, and still others could even be driven out of business -- although this is unlikely, as there are very few pure-play mobile SoC companies without other core businesses. Will NVIDIA emerge as one of the survivors? Will it be acquired? Or, will it simply exit the business? Nobody knows.
However, it's really a win-win. If NVIDIA does make it as one of the last few SoC vendors in this space, then the gross margin for the Tegra business will be quite high, and the company will be vastly more profitable than it is today. If NVIDIA doesn't make it, there's a good chance that it's either acquired -- its graphics patents/technology are extremely valuable -- or that it simply shifts its efforts toward other businesses.
What could it focus on instead?
NVIDIA's GPU business is incredibly robust. Even without Tegra, the company is set to grow along with PC gaming, high-performance computing, and heavy-duty workstation graphics and visualization needs. This is all on the back of the same fundamental hardware and software investments for its GPUs. Further, if NVIDIA simply jettisoned Tegra, it would see an immediate improvement in its bottom line at the expense of potential growth.
But NVIDIA's management hasn't given any indications that it wants to be less aggressive. It wants to use its GPU cash cows in order to turn NVIDIA into a robust growth story. If Tegra doesn't work, NVIDIA could very well double down on its position as a player in high-performance computing. It's no coincidence that NVIDIA joined the POWER consortium with IBM and Mellanox. An idea could be that NVIDIA ends up licensing the POWER 8 designs, merging with Mellanox, and positioning itself as the one-stop-HPC-shop, fighting Intel's data center group.
Foolish bottom line
There are many ways to grow, and with NVIDIA's assets and partnerships, success in consumer mobile system-on-chip products isn't the only way to meaningfully grow the top and bottom lines. If Tegra works out, then wonderful. If not, then the company still has plenty of options -- and plenty of cash to make them happen.
