Late last year, chip giant Intel (NASDAQ:INTC) announced to the world that it would be entering the market for high-performance stand-alone graphics processors. Such processors are used in a broad set of applications, such as 3D games, artificial intelligence training, professional rendering, and scientific computing.
Although Intel has built graphics processor technology for years, those efforts have been fairly low-performance products integrated inside the company’s personal computer processors. Those solutions are competent for basic computer use (e.g. video playback, web surfing, and even some casual games), but they simply don’t cut it for high-performance use cases. Stand-alone graphics processors — also referred to as discrete graphics processing units (GPUs) — are required in those situations.
Image source: Intel.
Those usage models are becoming increasingly popular. The leader in discrete GPUs, NVIDIA, has seen robust growth in sales of both its gaming and data center GPUs. According to analyst estimates, NVIDIA is on track to rake in $13.05 billion during its fiscal year 2019, with that figure expected to grow to $14.71 billion in fiscal 2020 .
There seems to be a lot of money to be made from selling stand-alone graphics processors, and while my fellow Foolish colleague Harsh Chauhan called Intel’s push into this market a “waste of time” — I disagree, and here’s why.
A discrete push benefits integrated graphics
Back in 2011, Intel introduced a processor family known as Sandy Bridge (the marketing name was second-generation Core). One of the reasons that Sandy Bridge was so successful for Intel was that it incorporated a graphics processor on the same silicon die as the CPU. That integrated graphics processor wasn’t a gaming powerhouse, but it was good enough for the average consumer who wanted to play videos and surf the web.
That integration paid off for Intel in the form of higher processor average selling prices. Then-CEO Paul Otellini said during the company’s Q2 2011 conference call that “what we’re seeing with Sandy Bridge is the trend to sell up a bit for every SKU because of the demand for the performance and the graphics capability.”
In other words, the integration of graphics led to a bump in average selling prices and, ultimately, revenue.
In subsequent years, Intel tried to move upmarket by building versions of its processors with even higher-end graphics in a bid to capture additional dollar content, but those efforts didn’t pan out as Intel had hoped thanks to a combination of underwhelming product performance as well as, in recent years, stagnation in the company’s graphics processor designs.
Intel’s focus on trying to build discrete graphics products should force the company to increase its overall hardware and software efforts with respect to graphics technology. Since it’s likely that Intel’s future stand-alone graphics processors will be based on the very same technological DNA as its future integrated parts, Intel could see a dramatic improvement in the competitiveness and value proposition of its integrated graphics processors as result of its discrete GPU ambitions.
New revenue opportunities
Even if Intel’s discrete graphics efforts don’t generate a huge financial return in and of themselves, the side effect of stronger integrated graphics technologies would still be a worthy consolation prize.
However, if Intel’s efforts succeed, then they could bring in a healthy amount of incremental revenue. When Intel announced that it was hiring industry veteran Raja Koduri to run its Core and Visual Computing Group, the press release said that he would “expand Intel’s leading position in integrated graphics for the PC market with high-end discrete graphics solutions for a broad range of computing segments.”
It seems to me that, at a minimum, Intel will leverage this technology in areas like gaming, professional visualization, data center, and automotive — all areas that NVIDIA targets with its own GPU portfolio. It’s too early to tell what kind of market segment share Intel can achieve across these segments (and any share gains are likely to take time), but with a longer term outlook (assuming, of course, strong execution on Intel’s part), I wouldn’t be surprised to see discrete GPUs add more than $1 billion to the company’s top line.
Ultimately, Intel’s discrete GPU push makes sense, because it’s a very good risk-to-reward proposition for the chip giant. In the worst case scenario, Intel should at least see a dramatic improvement in its graphics technology as a whole, which would benefit its core personal computer processor business. In more optimistic scenarios, Intel is able to capture a non-trivial amount of the fairly large total addressable market for discrete GPUs, bringing revenue and gross margin dollars to the company that it simply didn’t have before.