Rising competition from upstart makers of processors for machine learning failed to slow Nvidia Corp. as the maker of graphics chips reported better-than-expected fiscal second-quarter earnings.

The news, and investors’ anticipation of a positive earnings report for the quarter that ended on July 31, sent Nvidia shares up 2 percent today, to a record $59.70 a share, before they rose 2.5 percent more in after-hours trading. “Nvidia business is solid and strong,” Global Equities Research analyst Trip Chowdhry said in a note to clients today. “Fundamental strength may last for another 18 months or so.”

The company, whose chips are still mainly used in gaming personal computers, reported a net profit of $253 million, or 40 cents a share. Earnings before certain costs such as stock compensation rose 24 percent to 53 cents a share — way above Wall Street expectations of 37 cents — on revenues of $1.43 billion.

Despite a traditional focus on gaming, which is 55 percent of sales, that market grew only 18 percent from a year ago. Growing far faster were chips aimed at data centers, up 110 percent, and automotive systems, up 68 percent.

Nvidia Chief Executive Jen-Hsun Huang (pictured above) has made artificial intelligence, in particular the branch of machine learning called deep learning neural networks, a particular focus of development and marketing efforts in the past few years. That’s because Nvidia’s graphics processing units (GPUs) have proved to be especially adept at the parallel processing needed for deep learning. In May, the company introduced new graphics processors that Huang said were 10 times faster than the previous generation.

Apple, Facebook, Microsoft and other companies have been building and buying GPU-based systems for their data centers, driving much of the upside in Nvidia’s results. “The vast majority of growth comes from deep learning by far,” Huang said in comments on the earnings conference call today. In particular, about half the data center chip business is for deep learning, with another 35 percent for high-performance computing. “We find ourselves at the epicenter of this great dynamic,” he added.

Other chipmakers have also targeted deep learning recently, with a number of them offering different kinds of chips and computers that purport to do deep learning better. Google Inc. has designed its own deep learning chip, called a Tensor Processing Unit, for its own use. And Intel Corp. recently said it would buy Nervana Systems Inc., a maker of deep learning chips and services.

Huang expressed few worries about the competition. “Our sense is our lead is quite substantial,” he said, especially given the ecosystems of partners and software companies that can use the same chip architecture for applications ranging from cars to PCs to high-performance computers. “But we’re not sitting on our laurels.”

Nvidia also issued a forecast of $1.68 billion in revenues in the current quarter. Analysts on average had expected $1.45 billion, according to Thomson Reuters I/B/E/S. “I still see room for growth,” said Patrick Moorhead, president and principal analyst at Moor Insights & Strategy. “More and more will use neural nets and this is where the growth could come from.”

Photo by Robert Hof