Nvidia Inc
Huang spent this week on a country tour where he met with tech luminaries, business moguls and the Indian Prime Minister Narendra Modi.
"You have the data, you have the talent," Huang told Modi at a news conference in Bangalore. "This is going to be one of the largest AI markets in the world."
In keeping with Nvidia's secretive corporate culture, Huang was not specific about his company's Indian ambitions. Still, it's possible to piece together from partnership announcements earlier this month and during Huang's trip what Nvidia's doing there.
There are three things that make India really vital to Nvidia's AI push, Bloomberg TV reported. The first is that there is currently a restriction on trade between China and the US over the provision of graphic processors which are key to AI development. That means Nvidia has to find somewhere else to make them. Second, India is a huge talent hub for AI engineers. Third, India is fertile ground for manufacturing, which is also core to AI product expansion.
In order to better harness India's potential in these three areas, Nvidia announced early in September that it is building the country's - and one of the world's - most advanced AI supercomputers with best-in-class GPU and cloud.
In terms of core data and infrastructure, Nvidia is working with Reliance Industries Limited
To do this, Nvidia and Reliance will run AI processors that are a thousand times more powerful than the fastest supercomputer in operation today.
The new AI will be built on Reliance's Jio telecom infrastructure. Nvidia and Reliance aim to create customized AI service applications for Jio's 450 million users across a spectrum of languages using 5G, broadband and other networks.
While India lacks some of the GPU chip set expertise that Taiwan has, what it has in size is data. This data is a core component to building and training any AI, and Reliance's Jio platform is the country's largest of them all.
For GPU processing capability, Nvidia has said it will bring in Tata Group. Leveraging Nvidia's GH200 Grace Hopper Superchip, Tata will build the cloud that this high-powered supercomputer is going to run on and where all the data will be stored.
"Data centers worldwide are shifting to GPU computing to build energy-efficient infrastructure to support the exponential demand for generative AI," Huang said at the time.
Finally, Huang let on this week where he intended to get all the engineers to do the actual AI programming and the global management expertise Nvidia will need to expand these applications into foreign markets: Infosys Limited
The two companies jointly revealed this week that in particular they would focus on building advanced video software, speech and translation AI and AI generative software. Furthermore, Nvidia plans to train 50,000 Infosys engineers specifically on its generative AI programming.
"Generative AI will drive the next wave of enterprise productivity gains," said Huang.
Generative AI is Nvidia's core business and is a term used to describe technology that creates real-life text, image and video. Analysts concur that it will become a core component of enterprise and commercial applications over the next decade, excitement over which has seen the company join the elite ranks of $1 trillion plus market cap tech giants this year.
A New Virtual World
From what we can tell so far, the purpose of Huang's trip was to cement corporate and government deals so that Nvidia can get to work on developing its forthcoming next-generation suite of virtual reality software. A huge chunk of this development, it appears, will be done in India.
Infosys hinted that this new suite of applications it was making with Nvidia would entangle concepts such as world simulation and world building, 3D, "digital twinning" and hyper-advanced super-real interactive human-to-computer simulations. The bulk of this would be for use by major corporations around the world, Infosys' chairman Nandan Nilekani told reporters.
"Our clients are ... looking at complex AI use cases that can drive significant business value across their entire value chain," said Nilekani.
Perhaps telling in how Nvidia aims to proceed with developing its ambitious new product suite is the next destination of Huang's next overseas visit, which is another country that starts with the letter "I"" Israel. Israel is home to some of the most advanced mobile, cloud and data encryption technology.
NVIDIA's AI Summit in Tel Aviv will be fertile ground for application developers then, with over 2,500 participants expected to attend mid-month.
"The global generative AI race is in full steam," Haang said.
