India to chip away at the AI market

Last week, while the spotlight was on global political leaders gathering in New Delhi for the G20 summit, away from the limelight India’s tech scene took a significant step onto the world stage.

American company Nvidia, the tech giant known for its graphics processing units (GPUs), announced that they are teaming up with Indian powerhouses Tata and Reliance to turbocharge the growth of artificial intelligence (AI) applications and services across India.

This is a big deal as these partnerships are about to open up a world of possibilities as these new joint ventures can develop AI tools tailor-made for the unique needs of India and local companies.

For instance, healthcare could see AI stepping up the game in diagnostics and patient care even in the remotest corner of the country. And in education, we’re looking at personalised learning experiences for students who are unable to access higher learning centres that are mostly located in urban areas.

Until now, companies in the Indian domestic market have primarily depended on open-source AI platforms like OpenAI, which were predominantly trained on data from developed countries. That brings in inherent biases and other limitations, requiring significant customisation for local companies integrating AI into their products.

Back in January 2020, Niti Aayog, a think tank established by the Indian government, proposed the creation of a special AI infrastructure called AIRAWAT (AI Research, Analytics, and Knowledge Assimilation platform). It had the potential to be a game-changer, but unfortunately, it never gained much traction.

In contrast, countries like South Korea and Japan have allocated substantial funds to develop their own Large Language Models (LLMs) and have already made progress in establishing their AI platforms to support local companies, researchers, and students.

As for China, the capabilities and supercomputing infrastructure it has are almost light-years ahead of India. In the list of the top 500 supercomputers (those with more than one petaflop of capabilities), China has 228 such facilities, followed by the USA (117 systems) and Japan (29 systems), said the Niti Aayog report.

Chinese companies have already released several LLM models and their aim is to eventually end the US dominance in the sector. Meanwhile in India, even heavyweight tech companies depend on cloud-based AI services like Amazon's AWS and Microsoft Azure.

Now, the Indian private sector is stepping in to bridge the gap. Reliance and Tata are gearing up to integrate AI into their business operations and have forged partnerships with Nvidia to bring state-of-the-art AI capabilities to Indian companies.

One of the most exciting projects in this mix is India’s very own Large Language Model (LLM). This LLM will undergo training on the rich linguistic diversity of India, encompassing its 1.4 billion people who speak a multitude of languages and dialects.

The potential here is vast. This could lead to the development of AI voice assistants and chatbots, bringing technology to millions who remain illiterate and offering assistance to the older generation left behind by the emerging technology divide as digitalization becomes a part of daily life for citizens.

Another significant aspect of these deals is tapping into India’s immense talent pool. India boasts one of the world’s largest AI talent pools, with 1.6 million digital wizards. Surprisingly, the number of AI-focused startups doesn't quite reflect this talent, and some experts attribute it to a gap in AI skills within the workforce.

Nvidia’s CEO, Jensen Huang, hit the nail on the head during a media event in Bangalore when he highlighted that Indian AI experts often have to travel abroad to access supercomputers.

“You have your own data, lots and lots of data. You produce more computer scientists than any country on the planet.”

 “Produce AI here. You have everything you need. Everything you need to build the AI and use the AI here in India, but you need to have infrastructure.”

Both Jio Platforms of Reliance and Tata Communications have announced plans to develop cloud-based AI infrastructure that will harness some of Nvidia’s most powerful tools, including the Grace Hopper Superchip. This initiative aims to provide high-end computing access to companies, startups, scientists, and students across India.

However, it’s not the Indian development alone that forms the core of Nvidia’s expansion plan. These partnerships the company has unveiled serve a dual purpose: they allow Nvidia to broaden its reach while strategically navigating the US-China chip conflict.

When the US imposed a ban on exporting certain high-end chips to China, Nvidia had to tread carefully to avoid being entangled in the Washington-Beijing crossfire.

Already, they’ve struck deals in countries like Saudi Arabia to expand their business presence. In India, they likely see a market with potential rivalling that of China, provided the planned projects progress as intended.

Nvidia, accounting for roughly 60 to 70 percent of the global supply of AI server chips, has witnessed an unprecedented surge in demand for its products, propelling them into the trillion-dollar company club. This surge in demand shows no signs of slowing down in the foreseeable future, as both companies and governments are eager to build their own AI platforms.

The company has been present in India since 2004 and has established four engineering development centers across the country – in Gurugram, Hyderabad, Pune, and Bangalore – employing approximately 3,800 people. Huang, who was born in Taiwan, mentioned that his company collaborates with over 15,000 AI-based startups worldwide, including a significant number in India.

“I wouldn't be surprised if around 150 or so of these startups are right here in India.”

In India, the demand for AI-skilled professionals is expected to surge in the coming years, with both domestic and international companies seeking to recruit AI-savvy talent. A recent report from LinkedIn revealed that 57 percent of company executives surveyed in India are actively searching for AI talent this year.

As the demand for AI-trained individuals and AI-powered services continues to soar, India’s recent steps in this direction couldn’t have come at a more opportune time. However, there are several challenges to address, including concerns about privacy protection and potential job displacement due to AI.

Despite these challenges, Nvidia’s CEO remains optimistic about the future. “You [India] will export more AI expertise than any country in the world. It potentially could become your largest export,” said Huang.

Huawei’s new phone sets off alarm bells

Usually, we leave news about the release of new phones to others to report. However, Huawei’s new 5G Mate 60 Pro is something we thought we should discuss, as it is opening another front in the US-China confrontation. The release of the phone itself was unusual because the company didn't hold any glitzy events. The 7-nanometer processor in the Mate 60 Pro has set off alarm bells in Washington, as Huawei is one of the companies blacklisted by the US, and chip makers are barred from selling high-end products to them.

While the US has ordered a probe to determine if some company violated its sanctions and supplied chips to Huawei, the news has sparked patriotic chest-thumping on Chinese social media. Many comments praise the technological breakthroughs achieved by the Chinese despite the US sanctions. A teardown of the phone revealed some major advances made by China in developing domestic tech capabilities, reported  Bloomberg.

Student space project takes off

The students of Barton Hill Engineering College in Trivandrum have taken a significant step towards realising their dream project of becoming the first educational institution to launch their own satellite. Their goal is to develop and launch a cube satellite that will test long-range, low-power (LoRa) communication from outerspace. Using a balloon, the students recently floated a device into the atmosphere and successfully received signals from heights as high as 12 kilometres away. The next step for the students is to analyse the data they have gathered and plan for a higher-altitude launch. While this presents various technological challenges, the students express confidence in overcoming them. However, they are not sure about another problem they face: a lack of funds.

Sardine data can help food tech firms

Kerala is famous for its fish curry and oil sardines, or mathi as it is known locally, is one of the top favourites of the local people. The Central Marine Fisheries Research Institute in Kochi says it has decoded the entire genome of this species and the information can lead to the understanding about its evolution and the development of a sustainable strategy to ensure its abundance in our waters. The genetic data was 1.077 Gb in size and contained 46,316 protein-coding genes. The information can provide details about the Omega-3 oil the fish has and it can lead to food tech opportunities, reports Malayala Manorama.

This is out of the world, baby

Marriages are made in heaven, we know. But now babies too? Zip up and read on before you ask how to sign up. A Netherlands-based startup says it has developed a miniaturised in-vitro fertilisation (IVF) and embryo incubator that it hopes to launch into space soon. The firm says it “researches and enables different stages of human reproduction in space, enabling independent human settlements beyond Earth”. Initially, the firm will start experiments here on Earth with mouse cells, and even that could take five years, reports Interesting Engineering. Moreover, whether such space projects will get an official nod is also doubtful, it says. Thank your lucky stars for that.