Indian market has a cost advantage; need to solve India-specific problems, says panel

New Delhi: A panel of top artificial intelligence (AI) leaders in India said making the technology in the most populous country can aid in creating it for the world as such models are relatively easy to scale up.

Manish Gupta, senior director, Google DeepMind, indicated that developing AI models in India would make the models robust, as they would work for a large and diverse population. “Google started its AI research lab in India about five and a half years ago. So, we consciously picked this theme of ‘inclusive AI’…How do we develop AI in a manner that benefits everyone. And we saw huge gaps in capabilities of our own models. For instance, in languages like English versus so many languages spoken in India. I look at all the challenges in India as a source of competitive advantage,” said Gupta.

Echoing the sentiment, Geetha Manjunath, chief executive officer and founder, Niramai Health Analytix said the Indian “jugaad” method has a considerable cost advantage over developing AI in other jurisdictions. “We know there are a lot of problems in India, and that is good. Because necessity is the mother of innovation. What you do in a jugaad way to save costs and develop high-quality products with limited resources will definitely be useful to the rest of the world. Because we want to save costs everywhere, be it the UK’s healthcare costs for the NHS, or be it mediclaim-medicare costs in the US, reducing costs is always good. So we can develop for India,” said Manjunath.

Also read | Five artificial intelligence trends to track in 2025

“Jugaad” refers to an innovative way of improvising to complete any task.

The global AI market, in the limelight thanks to Sam Altman’s OpenAI and ChatGPT, was disrupted recently by the Chinese large language model (LLM) DeepSeek. When DeepSeek produced results matching ChatGPT at a fraction of the cost, the AI discussion moved towards efficiency, not in terms of getting results with a prompt, but in terms of reducing costs for computing and development.

AI’s development has garnered a massive market, driving chipmakers such as Jensen Huang’s Nvidia to become the most valuable company in the world with a market cap briefly touching $3.53 trillion, as reported by Reuters on 26 October 2024.

The US, led by the newly elected Donald Trump, decided to pour in investment worth $500 billion in the Stargate AI project, a partnership between OpenAI, Oracle, and Softbank, on 21 January this year. The investment would bolster the AI ecosystem by building data centres and power stations, and reduce regulatory hurdles in developing the technology.

India’s minister of information technology and electronics Ashwini Vaishnav announced that India will also develop its own foundational AI model and host DeepSeek on Indian servers soon, Mint reported on 31 January.

After the tech industry, AI will disrupt the financial services industry next, said Dilip Khandelwal, managing director and CEO, Deutsche India | Global CIO corporate function at Deutsche Bank AG. “If there is one industry which is getting heavily disrupted (by AI), and is closest to tech, it is the banking and finance industry. And what is happening there is not so surprising. Gen AI has created a lot of noise and buzz over the last two years, but if you look at what was happening over the last 5-6 years, backend processes were being taken over by machine learning tools,” he said.

Also read | OpenAI discussing localization of ChatGPT India data

AI has entered the world of financial services and made executives question deep-rooted legacy systems, Khandelwal added. “So, people have started looking at what is the future of these systems, or do we need to start looking at the software in a completely new way. That’s why financial industries are trying to leverage cloud infrastructure, Gen AI, and AI on top, to create those products which are completely different from the products they have seen in the past,” he said.

Making new AI products is also the core theme at Samsung, according to Mohan Rao Goli, corporate VP and MD, Samsung Research and Development, Bangalore (SRI-B). “AI is going to touch everybody. We are going through a transition from smartphones to AI phones. That is the theme at Samsung. For instance, the vision intelligence, we have to use computer vision algorithms blended with the artificial intelligence models, so that we can see the low-light scenes and then create that imaging experience,” he said.

A key challenge for India’s AI developers is skilling. “Skilling…the fundamental skilling of each engineer is important. We need to understand how AI, especially generative AI, can influence skills such as computer vision, communications, and connected intelligences—all things that we have learned over the years. We should come out of the comfort zone and learn about AI, starting from linear algebra to reinforcement learning, and then utilise these nuances as tools for these domains,” said Goli.

Lack of ambition and data privacy and protection are also critical challenges in the sector, according to the panel.

“To me, the biggest challenge is ambition. Or sometimes the lack of it. The lack of guts. The biggest thing I learned from my time in the US was not the technical skills, but daring to dream. Google has been a true pioneer in that. What I have seen us lacking is that boldness,” said Gupta.

Also read | Company boards set for rejig, ESG, AI skills in demand, Deloitte’s Goradia says

Manjunath highlighted concerns over the vast amounts of data that AI worked on, and complying with data privacy and protection laws in various jurisdictions. Legal disputes are likely to arise with the advent of AI, as illustrated by the flurry of lawsuits against OpenAI. 

While costs and skills are challenges, the solution may be creation of demand, according to Venkat Sitaram, senior director and country head- infrastructure solutions group, Dell Technologies India. “This has to happen with a partnership between the industry and the government,” he said. “Innovation and ambition is a part India’s AI industry, and businesses will start creating that demand for AI. We will start looking at infrastructure, we will start looking at models we want to develop. We will look at new ways of trying and testing new things, and not be completely risk-averse.” 

Doing this would generate economies of scale for the nascent AI industry in India, said Sitaram. “Costs will start driving down by itself. We play an important role in bringing costs down. ‘Make in India’ is one way of seeing how we can bring costs down by making it easy and accessible. We have invested heavily in our supply chains significantly on bringing the costs down, on how our customers can order a PC and get it delivered in 24 hours, that too the configuration that you want to buy, not the configuration we have in our stock,” he said.

AI will help create a model of consumption patterns and other relevant information that will allow the supplier to procure such goods accurately, Sitaram said.

And read | 2025 to be India’s year in semiconductors and data protection: Minister Ashwini Vaishnaw