The US is restricting NVIDIA from selling its advanced AI hardware to China and it will heavily impact the Santa Clara-headquartered company’s revenue since China is one of their most important markets. 

China is a key market for Nvidia, so much so that the revenue from China and Hong Kong accounted for 22% of the company’s revenue in 2022, according to its financial statements.

“Over the long term, restrictions prohibiting the sale of our data centre GPUs to China, if implemented, would result in a permanent loss of opportunities for the US industry to compete and lead in one of the world’s largest markets and impact on our future business and financial results,” said Colette Kress, NVIDIA’s chief financial officer during an investment conference in June this year.

Furthermore, US officials plan to tighten these export curbs. The new move is aimed at Nvidia’s A800 chip, which it created to bypass the regulations and in order to continue to sell to China.

“We are aware of reports that the US Department of Commerce is considering further controls that may restrict exports of our A800 and H800 products to China,” Kress said recently at an investment conference.

Exploring New Frontiers

While the business with China has become tough because of geopolitical reasons, Jensen is searching for new markets to re-route the supply. And India is the next big market after China. 

In 2020, India’s AI market was worth $3 billion, constituting about 1% of the global market, according to IDC. However, AI spending in India is projected to grow significantly, reaching $11.78 billion by 2025 and contributing $1 trillion to the country’s economy by 2035.

A joint study by Microsoft and IAMAI found that India’s AI market is set to grow by 20% in the next five years, second only to China globally. 

Last week, NVIDIA CEO Jensen Huang visited India and made some big announcements. The company partnered with major Indian companies such as Tata, Reliance and Infosys. Under this partnership, NVIDIA will deliver AI computing infrastructure and platforms to these companies for developing AI solutions.

The partnerships also underlined the demand and the willingness on these companies’ part to integrate AI and expand on their ability to build and deploy AI models. 

India, though a small AI market, contributes 16% of global AI talent, ranking among the top three talent producers worldwide, surpassing its own demand. With a legacy of exporting tech talent globally, India produces more talent than it consumes.

“I would like to accelerate the building of AI infrastructure here in India. You have your own data, lots and lots of data. You also have very diverse languages. You have the great talent of computer scientists,” said Huang during the press meet. 

What the Partnerships Entail

The partnership ranges from creating LLMs in local Indian languages to building GPU cloud infrastructures and upskilling employees on a large scale for the AI revolution that’s on the horizon.

NVIDIA, in partnership with Reliance Industries, aims to develop India’s foundational LLM trained on the nation’s diverse languages, tailored for generative AI applications. The project will create AI infrastructure surpassing India’s fastest supercomputer, with access to NVIDIA’s GH200 Grace Hopper Superchip and DGX Cloud for high performance. This initiative will enable Reliance Jio Infocomm to develop AI applications for its 450 million customers and provide energy-efficient AI resources for Indian scientists and startups, expanding to 2,000 MW data centres.

“As India advances from a country of data proliferation to creating technology infrastructure for widespread and accelerated growth, computing and technology super centres like the one we envisage with NVIDIA will provide the catalytic growth just like Jio did to our nation’s digital march,” Reliance chairman, Mukesh Ambani said. 

Meanwhile, Infosys is poised to collaborate with NVIDIA, leveraging its infrastructure and expertise to develop AI models and applications. NVIDIA’s CEO Jensen Huang discussed plans to reskill over 300,000 employees in AI, aligning with NVIDIA’s recent partnerships with Reliance and Tata to build significant AI infrastructure in India. These endeavours, utilising NVIDIA’s GH200 Grace Hopper Superchip and DGX™ Cloud, will empower TCS to innovate and enhance the skills of its extensive workforce of over 600,000 employees.

In the rapidly evolving global AI landscape, these partnerships signify India’s potential to become a major player. 

OpenAI’s five-year journey is a testament to AI’s growth, and India, with plans to acquire tens of thousands of GPUs and achieve AI supercomputers 50 to 100 times faster by the next year, is poised to achieve similar success in a shorter timeframe.

“We are going to bring out the fastest computers in the world. These computers are not even in production [so far]. India will be one of the first countries in the world [to get them],” Jensen said. He added that by the end of 2024, India will have AI supercomputers that are an order of magnitude faster (i.e. 50 to 100 times faster), lowering the cost of foundational model training.