Given the news that Intel (Nasdaq: INTC) appears to be leaving the discrete graphics card business (the segment where Advanced Micro Devices' (NYSE: AMD) ATI division and NVIDIA (Nasdaq: NVDA) derive most their sales), you'd expect some high fives, chest bumps, and maybe even a fist pump or two at AMD and NVIDIA headquarters. Yes, on the surface it's great news for both companies, but I also think Intel made the right call here and is now better positioned to compete with more valuable parts of the graphics card industry.
Let's blow this town ... or not
The "news" of Intel's departure from discrete graphics cards came in this blog post on their corporate site and was anything but definitive. The most relevant snippet from the blog post: "We will not bring a discrete graphics product to market, at least in the short-term," takes pains to leave the door open for future Larrabee products targeting the non-integrated consumer graphics market. Intel might be saying, "Yeah, we won't be coming back ... in the short run. Nudge nudge, wink wink." Despite the cliffhanger ending, don't bet on Intel coming back here -- they've wizened up.
What Intel has realized, like NVIDIA before them, is that the real action is in so-called "high-performance computing." Not that AMD and NVIDIA can't mint some cash marketing graphics cards to consumers, but the market is tiny relative to Intel's other ambitions. Also, marketing graphics cards increasingly targets a small niche of enthusiasts, not exactly the forte of a large goliath like Intel that's focused on huge sales to large enterprises.
Finally, while there might have been concerns back in 2007 that graphics cards couldn't handle the increasing loads on a variety of computing functions, Intel's new integrated graphics releases and other solutions like Broadcom's (Nasdaq: BRCM) Crystal HD Accelerator have proven to be "good enough" solutions for handling most the multimedia that more robust graphics processors threatened to steal from the central processor.
High-performance computing, high-performance profits
Focusing on "high-performance computing" (which scales from powerful workstations up to supercomputers) plays much better into Intel's hand. These systems are a more significant investment and generally purchased by companies where Intel has existing relationships.
Also, right now the workstation market is controlled by NVIDIA, which has around a 90% market share in workstations. It's a brand new market for the company with much higher gross margins -- between 60%-70% for NVIDIA) -- and an area where Intel can add to its profits while defending against technology that threatens to overtake its own processors.
I do wonder, though, how far Intel will push into the market. Competing against NVIDIA's Quadro line makes a lot of sense. However, scaling Larrabee up to take on more robust products like NVIDIA's Tesla might be cannibalizing existing sales Intel already controls. From that standpoint, an assault on the workstation market looks to be priority No. 1.
Bottom line
So, at the end of the day, AMD has the most to celebrate. Its recent product launches have been strong compared to NVIDIA, it has a cost advantage on most graphics cards, and now Intel has left its strongest market. NVIDIA gets the bonus of knowing its only consumer graphics competition is AMD, but Intel's strategic realignment is targeted directly at a segment where they see their best margins, and I see the most future potential.
You might say Intel's not retreating, they're advancing in another direction. As far as whether or not Intel will have better success in high-performance computing than they had in consumer graphics cards, let's just say skepticism is the best view until proven wrong.
