Small Business and Technology Focus

Return to Article

Nvidia Banks on Split-Personality Chip

Alexei Oreskovic

Hundreds of video gamers eagerly convened in San Jose, Calif., last November to celebrate the arrival of Nvidia's G80 graphics processor.

Since then, it's been Nvidia hitting the road, striving to convince a different group of computer users of the merits of a chip that the company hopes will bring a big boost to its business.

In June, Nvidia will launch a new brand dedicated to selling the G80 -- not as a graphics accelerator for PCs and workstations, but as a chip intended to take on data-crunching computing chores currently handled by microprocessors.

Nvidia calls the concept "GPU computing" and contends that the 128 individual "stream" processors packed into the G80 chip make it ideally suited for such computational heavy lifting.

According to CEO Jen-Hsun Huang, the G80 boasts 10 times the floating point computational muscle of today's top-of-the-line PC microprocessor.

"We believe GPU computing will usher in an era of the personal supercomputer, and will dramatically accelerate the adoption of new methods from computational chemistry to computational finance to computational genomics," Huang said in a February conference call with financial analysts.

The effort is being spearheaded by Andy Keane, who joined Nvidia last year and has worked at microprocessor outfits like Intel and MIPS Technologies.

Last year, speculation grew that Nvidia was secretly developing a PC microprocessor based on the x86 instruction set to better compete with the recently merged Advanced Micro Devices and ATI Tech, which fields both graphics chips and microprocessors, as well as Intel, which already has both capabilities.

For now, though, Nvidia appears to be approaching the computing market from another angle.

Nvidia says the GPU computing business will focus on high-performance computing, a market in which multimillion-dollar supercomputers and banks of interconnected servers plow through massive amounts of data to simulate nuclear blasts, uncover oil and gas deposits and run complex financial models.

All those applications require parallel computing, something which graphics processors are inherently good at doing.

Compared to current hardware requirements, the potential to use mass-produced Nvidia graphics chips to do these jobs offers customers a lot of bang for the buck.

"Because the graphics guys are turning out these chips that have a very high inherent floating point capability, and doing it in much higher volumes than any supercomputer could ever think about ... the economics are very compelling," says Nathan Brookwood, a microprocessor analyst with research firm Insight 64.

"If you have a problem that lends itself to lots of computation in parallel, then this is a very cost-effective approach," Brookwood says.

But Wall Street is taking a wait-and-see attitude to the graphics processor's alter-ego.

"It's an intriguing avenue of development," says Stifel Nicolaus analyst Blake Fischer. "The issue is, what is its applicability into the real world? From what we're seeing at this point, we view it as more of a niche product."

Mainstream PC users don't really need parallel processing capabilities, nor is the software that runs on standard PCs optimized to take advantage of parallel processing.

That leaves Nvidia's GPU computing group fishing in a small, albeit lucrative pond, say analysts.

The Cell processor, jointly developed by Sony, IBM and Toshiba, also boasts parallel processing prowess. But besides being featured in Sony's own PlayStation 3 game console and certain medical imaging products from Mercury Computer, the chip has yet to gain much mass appeal.

However, Nvidia isn't dissuaded by fears of a limited addressable market, and believes the new GPU computing business will help its professional group grow to a $1 billion business in the next couple of years, says Michael Hara, vice president of investor relations.

Of course, the company's professional group, which currently provides traditional graphics accelerators for workstations, is already a $500 million businesses growing at 30% a year. So, the $1 billion mark would be attainable in a few years anyway at current growth rates.

However, GPU computing adds another engine to Nvidia's professional group, giving investors confidence that that business will be able to keep growing, says Stifel Nicolaus' Fischer.

"But I don't think many people have started to bake that into their estimates," Fischer says. Stifel Nicolaus makes a market in Nvidia shares.

That said, any contribution from the new GPU computing effort can only help the bottom line. Nvidia's current crop of professional products carry profit margins in the 60% range, well above the corporate average of 44%. And the forthcoming GPU computing products promise similar margin advantages, since they leverage the investments Nvidia has already made in its mainstream graphics processors.

The expectation on the Street is that Nvidia will raise its corporate gross margin target from its current level of 45% sometime in the next few quarters.

If the GPU computing business lends a hand in hoisting the margin, it will already have earned its keep.