Nvidia’s GTC event will shine a spotlight on 200 AI startups

Nvidia's GPU Technology Conference (GTC) kicks off Monday with a keynote speech by Nvidia CEO Jensen Huang, whose company has become one of the most powerful in the tech world with a market value of $322 billion. And the event will run alongside Arm DevSummit, which is a coincidence as Nvidia has agreed to buy Arm for $40 billion. Normally held as a physical event, the virtual event will highlight an ecosystem that has millions of developers.

More than 200 new AI startups will appear at GTC in some form, said Jeff Herbst, vice president of business development and head of the Inception virtual AI program at Nvidia, in an interview with VentureBeat.

“We love startups,” said Herbst. “We're like a mouthpiece and a helper for the companies. We help them learn from the collective knowledge of the industry. Our ecosystem is growing like a weed. We have passed two million developers on our platform.

If Nvidia gets approval to buy Arm, the developer count will cross 15 million. That means that Nvidia, which evangelizes developers to use its graphics chip and AI technology, has come a long way.

Nvidia's GPU Technology Conference (GTC) started in 2009 in San Jose, California. I went to the first event and remembered how it was nowhere near filling the sprawling San Jose Convention Center. Over the years, I went to many events where Herbst drew attention to startups using Nvidia's tech.

Herbst's division invests in startups that use Nvidia's technology, but more often than not Nvidia now gives away its technology to the companies that are getting started with AI or graphics technology. Over the years, Herbst has formed alliances with venture capitalists to help them narrow their focus on the best startups. Nvidia has invested alongside the VCs when it finds the best companies.

For many years, Herbst ran a contest for the best GPU (graphics processing unit) startups, but the startups grew so plentiful that it was hard to narrow down the pack to a single winner, Herbst said.

“We have over 6,000 companies in the Inception program, and it's still growing,” Herbst said. “These developers train on our platform for years. It's becoming more popular, like a flywheel. We have found other ways to help them than to invest money into them.”

GTC has more than 30,000 mostly paid registrants, and it promises to be one of the largest, most complex virtual tech conferences, with five days of programming, nearly 24 hours a day, across seven time zones.

More than 1,000 live and pre-recorded sessions will be available in 40 topic areas, in one of five languages: English, Mandarin, Japanese, Korean and Hebrew.

Huang's keynote will take place at 6 a.m. to 8 a.m. Pacific time on Monday. His talk will cover advances in chips, AI, enterprise and edge computing, robotics, and remote collaboration. He will speak again with Simon Segars, CEO of Arm, at the Arm DevSummit event on Tuesday morning at 8:20 am. Pacific.

Among participating organizations are: Amazon Web Services, American Express, Audi, BMW, Caltech, Carnegie Mellon University, Facebook AI Research, Google Cloud, IBM, Johns Hopkins University, Lockheed Martin, Microsoft, MIT, Oak Ridge National Labs, OpenAI, Oracle, Sony, VMware, and Volvo.

On Tuesday at 12 p.m. Pacific, Huang will join Herbst at the Inception virtual startup incubator to talk about AI startups. Nvidia has more than 6,000 startups in its Inception program, where the company provides the startups with software and hardware, without taking an equity stake.

“We're going to talk about the combination of Nvidia and Mellanox and how we're approaching things at a system level for compute and storage,” Herbst said. “If AI is about programming using data, the system approach is about facilitating the movement, cleaning, and storage of data. Jensen's view is that the datacenter is the new compute unit.”