PALO ALTO, U.S. -- Perhaps little known outside Silicon Valley, Nvidia's sharp growth in graphics processing units (GPUs) -- supported by rising demand for artificial intelligence -- has catapulted the company into a serious challenger to the chip industry leader Intel.
Defying industry stall
At Nvidia's May 10 earnings announcement, CEO Jensen Huang challenged the longstanding Moore's Law predicting the accelerating innovation of semiconductors, i.e. that the number of transistors in a dense integrated circuit doubles roughly every two years. Intel co-founder Gordon Moore put the idea forward in 1965.
"Microprocessors no longer scale at the level of performance they used to -- the end of what you would call Moore's Law," Huang said. The black-leather-jacketed executive's provocative jab at the long-dominant chip giant reflected his confidence that GPUs will take center stage in the age of AI.
AI development requires high-speed processing of massive data, including images. GPUs may not be suitable for precision processes handled by central processing units (CPUs), but they have an edge in processing speed.
Since its 1993 founding, Nvidia has specialized in GPUs. The company has built its success on GPUs for video game consoles, and has survived competition by constantly honing its technologies, while rivals disappeared in buyouts by other chipmakers.
Nvidia stock has been thriving as well, surpassing Intel in share price for the first time in February 2016. Intel stock has since moved little, but Nvidia shares have kept rising. As a result, Nvidia stock is nearly five times more valuable than the blue chip today.
The broadening presence of Nvidia can be seen in its collaborative agreements as well. In March, it announced a tie-up with Microsoft. And last October, the company unveiled plans to work with Japanese industrial robot giant Fanuc on "connected factories."
A recent highlight is a series of alliances with automakers. In May, Nvidia announced a collaboration with Toyota Motor that will see the leading Japanese automaker develop autonomous vehicles using the U.S. partner's GPUs. Other automaker partners include Audi, Tesla and Ford Motor.
Developing an autonomous driving system requires vast amounts of programming to factor in various scenarios to control brake timing and other functions, with test driving over many months being an essential part of the process.
Nvidia's deep learning AI dramatically speeds up the process, thanks to its ability to predict scenarios on its own. Using the AI, Audi was said to have finished the amount of programming that would have taken two years previously in just four hours.
Danny Shapiro, the senior director of Nvidia's automotive-related operations, says carmakers can simply add their own specifications on top of the brain of autonomous system Nvidia provides to deliver a certain driving feel.
Nvidia is developing a license-plate-size AI system and calling for automakers around the world to use it in their self-driving vehicles.
Nvidia expects GPU demand to rise for virtual currency applications as well. Huang said at a telephone conference after the earnings event May 10 that GPUs are suitable for mining virtual currencies. "What we've done, our strategy is to stay very, very close to the market. We understand its dynamics really well. And we offer the coin miners a special coin-mining SKU. And this product configuration -- this GPU configuration is optimized for mining," he said.
Huang was highly optimistic about the future of GPUs at the earnings event, but his prediction will come true only if the GPUs become the industry standard for autonomous driving. Nvidia's battle to replace Intel as a semiconductor leader has begun.