TAIPEI -- Nvidia does not feel threatened by a growing number of potential rivals entering the artificial intelligence field, but rather the graphics chipmaker sees the surge of interest in the emerging technology as welcome and a promising growth opportunity.
"Any time, the more companies work on AI, the faster it will grow. For five years it was very, very silent when we worked on AI. The more people started to work on AI, the faster we started to grow; the more people work on AI, the more people talk about AI, the more people use AI, the faster we will grow," Nvidia chief executive Jensen Huang told a small number of reporters on the sidelines of the annual Computex technology show in Taipei.
"Today the AI industry is approximately zero, in the future it will be the largest industry in the world ... So long as it's as fast as possible, it's good," Huang said, adding that he was confident Nvidia would continue to deliver the best AI platform.
Nvidia's graphics processors are used in handling large amounts of image and video data for AI and other emerging technologies. The company is also working with Audi, Tesla and Toyota on self-driving technology.
Huang was commenting on Apple's reported efforts to build its own AI chips, first revealed by Bloomberg on May 27, and Google's ongoing efforts to create its own Tensor processing unit, or TPU, to train machines to learn without being explicitly programmed to do so.
Nvidia noted in its latest annual report that it is a supplier for the Google Cloud service.
At a time that Google is showing signs that it does not want to continue to rely on external chip suppliers, with the release of the second-generation of the TPU in mid-May, Huang took a light jab at his customer.
"Our [graphics processors are] much more robust and much more flexible because we have to do many things. As a result, almost every network that you can think of, any artificial intelligence network, runs perfectly on our GPU. TPU has limitations, because precision is limited, design is simpler, as a result it has limitations. But again they are just two chips," he said.
According to research firm IDC, worldwide revenues for cognitive and AI systems will grow 59.3% from 2016 to reach $12.5 billion this year, and achieve a compound annual growth rate of 54.4% through 2020, when revenues will top $46 billion.
Robust first quarter
The company delivered a robust first quarter, with both revenue and earnings beating estimates. GAAP earnings per share surged 126% year-over-year to $0.79, on revenue of $1.94 billion, up 48% from the year-ago period.
It also offered strong guidance for the current quarter, forecasting revenue of $1.95 billion, plus or minus 2%.
The company's shares, which have gained more than 200% over the past 12 months, closed at $144.87 on Tuesday.
Nvidia unveiled new gaming products including graphic cards for notebooks and monitors during Computex.
It also announced it will work with four major Taiwanese contract electronics makers -- Hon Hai Precision Industry, Quanta Computer, Inventec, and Wistron to build new data center servers with Nvidia's AI design.
Data centers are now a part of essential infrastructure for major internet companies and require massive computing power to gather and analyze lots of data.
Foxconn is the largest server maker in the world, while Quanta is now known for supplying servers directly to major U.S. tech companies, including Facebook and Google.
While demand for data center servers is strengthening with the rise of AI and other next-generation technologies, Huang said that he still does not see servers replacing gaming as Nvidia's main source of income.
Nvidia is betting big on AI, but Huang is aware of potential social and security issues that the new technology may entail.
He is promoting open dialogue on new technologies and democratizing AI, or making the technology accessible for everyone, everywhere, to prevent possible abuse.
"The best way to protect society from any technological threat is to democratize it. If everyone has the access to the same capability, if every country has access to the same capability, if every company has the access to the same capability, then it will be very difficult for it to be used improperly without everyone else using the technology to stop it," Huang said.
With the worry that rogue states will also invest in AI, "we must invest in AI even faster and even greater, all of us," he said.
"I think because people are generally good, therefore if we give generally good people access to the best possible technology, I believe they will be generally used in a good way."