In November 2017, chip giant Intel (INTC 5.00%) announced that it would be entering the market for high-performance, stand-alone graphics processors.
From a strategic point of view, this makes perfect sense as graphics processors become increasingly critical in fast-growing areas of the computing market, such as gaming-specific computers. Building a successful stand-alone graphics processor business would ultimately accelerate Intel's revenue and profit growth rate.
Image source: Intel.
Moreover, since Intel already invests heavily in graphics technology (nearly every personal computer processor that Intel ships has an Intel-designed graphics processor integrated inside), entering the market for high-performance stand-alone graphics processors seems like a good way to leverage work that's already being done.
However, there is one risk to Intel's efforts here that I think is worth keeping in mind.
The manufacturing millstone
Graphics processors are made up of many relatively small computing cores that, in aggregate, deliver quite a lot of performance. Moreover, the performance of a properly designed graphics processor tends to scale with the number of cores that it has, all else being equal.
Additionally, graphics processors tend to be power constrained, so the more efficient each of those graphics cores is, the more of them will fit in a given power envelope, which ultimately means better performance for the intended use.
Image source: Intel.
With each new generation of chip manufacturing technology that chipmakers like Intel put out, transistors -- the building blocks of computer chips -- become substantially smaller (meaning more of them can fit in a given area) and significantly more efficient.
It's not surprising, then, that the biggest leaps in graphics processor performance tend to be tied to generational advances in chip manufacturing technology.
In the past, Intel was known for its ability to migrate to denser, higher-performing chip manufacturing technologies faster than its industry peers. However, in recent years, Intel has faced significant difficulties in bringing new manufacturing technologies to market.
Those difficulties have been so severe that they have led to product cancellations and the insertion of low-risk products built using slightly tweaked versions of its current manufacturing technologies into its product road maps. Ultimately, Intel's competitive positioning in the market has weakened because of its difficulties in bringing new manufacturing technologies into production.
By contrast, Intel's main manufacturing competitors -- Taiwan Semiconductor Manufacturing Company and Samsung -- have migrated to new manufacturing technologies at a brisk pace. This has allowed these two companies to beat Intel's current products in terms of transistor density.
The key player in the stand-alone graphics processor market, NVIDIA (NVDA 3.17%), relies on both TSMC and, to a lesser extent, Samsung, for the manufacture of its graphics processors.
So, if Intel continues to face difficulties with its manufacturing technologies while its chipmaking competitors successfully migrate to new technologies, then NVIDIA could have access to more advanced chip manufacturing technologies for its graphics processors than Intel will for its own.
NVIDIA is already substantially ahead of Intel when it comes to the quality of its graphics processor architectures and its relationship with major game developers, so Intel will undoubtedly have to fight hard to win stand-alone graphics business away from NVIDIA.
Being at a manufacturing technology disadvantage relative to NVIDIA on top of other disadvantages would make it that much harder for Intel's efforts in stand-alone graphics to succeed.
