India's AI Moment

 AT A TIME WHEN many city dwellers in India, especially those in New Delhi, are reeling under dangerously poor air quality, what could be more heartening than coming across an academic paper in the journal Scientific Reports exploring how artificial intelligence (AI) can be employed to combat air pollution. The study by KS Rautela and MK Goyal concludes that precise forecasting of the air quality index can help devise region-specific measures to reduce air pollution and save lives by sparing people the grim prospect of breathing air equivalent to smoking several cigarettes, a condition linked to reduced life expectancy and various cancers, particularly of the lungs.

But let us first look at where India stands in terms of AI adoption, or to paraphrase it, where does the India AI story go from here? Is the trajectory of growth towards achieving AI independence? What to make of IT behemoths such as OpenAI, Perplexity, Google Gemini, and so on investing more in India, tying up with local institutions and aligning with the government’s IndiaAI Mission to develop national AI capabilities?

Just a recap: AI company Perplexity’s AI browser Comet is available in India to Perplexity Pro plan subscribers. It had partnered earlier this year with telecom operator Bharti Airtel to offer a free 12-month Perplexity Pro subscription (a premium AI-powered chatbot that otherwise costs around ` 17,000) to all 360 million Airtel subscribers. For its part, OpenAI, the tech giant backed by Microsoft, which is also busy hiring local talent, has launched a ₹399 monthly plan to make ChatGPT more affordable for locals. Anthropic, the maker of the ‘Claude’ AI chatbot, is also opening its first India of­fice in Bengaluru in early 2026. Google has committed an investment of $10 billion in India over the next five-seven years with a focus on digitalising the economy and building India-first products. The bottomline is this: several top-notch AI players are looking to tap the Indian market for local talent and to set up data centres as part of their growth strategy.

At the same time, India is also looking to make indigenous AI products as envis­aged in its IndiaAI Mission, the federal programme with a ₹10,300 crore budget to develop indigenous AI capabilities, improve computing infrastructure and foster AI talent and startups. That includes developing a sovereign AI model by Feb­ruary 2026. Speaking recently at the India Mobile Congress 2025, Electronics and IT Secretary S Krishnan said that though In­dia was a latecomer in AI, it has scaled up its computing infrastructure rapidly with the deployment of 38,000 GPUs (graphics processing units) and building up its own foundational models besides making AI-driven solutions exclusively for women and youth. Homegrown products include AI-power translation platform Bhashini, agentic AI Krutrim, and language model Sarvam-1, among others.

SEVERAL INDUSTRY EXPERTS and CEOs of AI companies Open spoke to are of the view that while the AI growth story in India is for real, it is much more as a technology consumer than a technology producer. Simply put, the country has a big user base, often the biggest after the US, for US-based AI giants, while AI independence remains a work-in-progress at an early stage.

The reasons are not hard to compre­hend. Primarily it is our history and then it is the mindset.

Tanish Shewani, founder of Bharat Carbon, a climate-tech company that is into carbon accounting, net-zero strategy and so on, tells Open that while India, as a country with largescale mobile phone penetration, offers enormous scope for AI, IT companies back home often suffer from a legacy of being a services provider. He points out that historically Indian companies prefer stability and return on investment. For someone who upholds sustainable development, Shewani notes that one of the biggest challenges for India would be in upskilling its engineers to make them job-ready for new frontiers of AI.

Rajeev MA, who is part of the Artificial Intelligence (AI) Program at TCS Digital Enterprise, broadly agrees with these ar­guments. Rajeev, whose responsibilities include deep learning, offers his views on why India is more of a tech consumer than a maker. “Many factors influence this. Re­search and development spending, risk capital, education systems that primarily give more importance to memorisation as opposed to applying concepts learned on new problems, the stigma around failures as opposed to seeing it as a stepping stone to success, risk aversion, etc are some of them,” says this alumnus of IIT-Madras. He goes on, “Leaders are also moulded in the same system and very few are able to come out of it. I think it is changing with many young startups that need to be en­couraged. Service-oriented companies in my opinion have very little appetite for risk-taking because they are answerable to the stock market quarter-on-quarter. Companies like Zoho don’t suffer from this issue because they are either self-funded or through private equity.” The TCS veteran also emphasises, “You need to be leaders in three areas to be successful (in AI): data, algorithms, and infrastruc­ture or compute. India might have some unique data that are region-specific. We are weak in algorithms and infrastruc­ture. Take a look at data centres in the US versus India. We are far behind. The gov­ernment is purchasing accelerators for deep-learning training. But we are (still) in the early stages.”

Even so, there is great excitement around overseas AI titans setting sights on India as part of their next stage of growth. Apart from India being the most populous country in the world, it is also a nation where 92 per cent of knowledge workers use generative AI tools compared with the global average of 75 per cent, ac­cording to a study jointly done by Micro­soft and LinkedIn last year. As evident from India’s credentials as a massive user base for social-media entities and messag­ing platforms, generative AI has clicked well in the country too, especially thanks to mass penetration of smartphones and mobile internet. As many as 85.5 per cent of households possess at least one smart­phone according to the latest National Sample Survey. In addition, around 86.3 per cent households in India have access to the internet within home premises. In­ternet access among the youth is much higher, according to this survey. In the age group of 15-29 years in the countryside, around 95.5 per cent own a smartphone in rural areas. In urban areas, around 97.6 per cent own a smartphone in the same age group. Many IT giants who offer online payment facilities through UPI stand to gain immensely thanks to the surge in the numbers of those who use internet tele­phony to make payments. Further, India is home to more than 825 million mobile broadband users.

The sectors that attract huge potential for these AI companies include health­care where AI-powered diagnostics and telemedicine can help address the short­age of doctors, especially in rural areas. Agriculture is another area that needs AI for weather and crop prediction as well as sensors and drones for precision farm­ing, among other requirements. As with education, players like OpenAI and oth­ers have already begun collaborations to offer products and services. A section of analysts states that OpenAI wants to do what Microsoft once did with Windows: quietly colonise young psyches (‘Chat­GPT Comes to the Classroom’, Open, Sep­tember 16, 2025). Governance is also one segment that calls for AI integration. No­tably, Indian MNC Zoho, which hosts the emails of lakhs of government employ­ees, has its own language model, Zia, and has bundled new agentic AI features into existing subscription plans. Zoho, found­ed by Sridhar Vembu, is being pitched by the government as a ‘Swadeshi’ MNC and a rival to Google and Microsoft in India.

BUT CERTAIN CHALLENGES, as raised by TCS’ Rajeev MA and other experts persist, and there are no signs yet of India tackling such hurdles very soon. A scarcity of high-quality da­tasets, especially in Indian languages, is a significant obstacle for training accurate and unbiased AI models. Notwithstanding an overall progress, many rural areas still lack robust and high-speed internet, cloud computing and data storage infrastructure. Most importantly, skills gap and high com­pute costs are major causes for worry.

Bengaluru-based author and AI expert Sreejith Sreedharan points out a few remarkable features of India’s AI story. “India has quickly emerged as an AI implementation powerhouse. Sectors like energy already report 84 per cent pro­ductivity gains. Investment tells the same story. Private AI funding has jumped from $3.14 billion (2013-20) to over $8 billion in just four years, putting India among the world’s top AI destinations,” he says, emphasising that beneath the visible mo­mentum lies a worrying fragility.

“India invests just 0.64 per cent of GDP in R&D (GERD), compared with 3.45 per cent in the US, 2.68 per cent in China, and nearly 2 per cent in Singapore,” Sreedharan says, adding that outside of space technology, India’s deep-tech base remains thin for a country aspiring to global leadership. “The consequence is predictable on the AI frontier too: depen­dence on foreign frameworks for foun­dational AI, semiconductors, and major computing infrastructure. That depen­dence will be tested as India scales data centres and their vast energy needs,” he warns. For an energy-starved nation, this could be a double-edged sword: “Rapid deployment of generative AI also risks unsettling millions of jobs with entry-level roles many young Indians rely on most. In the short term, adoption may deepen the rural–urban gaps already visible in incomes, digital access, and opportunity.”

Sreedharan feels that the solution is straightforward in outline though hard in practice: “India must move from being the fastest adopter to becoming the coun­try that builds the foundations others rely on. That requires pairing our adaptive strengths with sustained, long-horizon deep-technology research, backed by ro­bust government-led R&D, clear public-private partnerships, and targeted invest­ment in chips, power, and skills. Only by choosing that path can India shape, rather than be shaped by, the future of AI.”

Being the back office of the global soft­ware industry, it is easier said than done. However, there are optimists who argue that local talent will flourish if they are given a suitable ecosystem to grow and make a name for themselves globally. “India’s AI story is still unfolding, and it is too early to say our mindsets won’t change fast enough. There are bright rays of hope, especially because of growing col­laborations between public and private sectors,” says a government official based in Chennai who is close to the matter. He adds that entrepreneurs and senior hands from academia are thrilled about the future of Indian AI.

Building AI data centres and infra­structure to enable smooth computing remains a huge task. Similar to semi-con­ductors that need scarce ultra-pure water, the country could require an additional 45-50 million square feet of real estate space and 40-45 Terawatt Hours (TWH) incremental power by 2030 to meet the growing demand for AI, according to a re­port by Deloitte. The question that comes up: Where will that power and water re­source come from?

We don’t know yet. “Overcoming re­source constraints, such as the shortage of skilled talent and the significant financial investment required, will be critical to the long-term viability of the AI Stack,” notes a recent report titled ‘Digital Sovereignty and AI: Developing India’s National AI Stack for Strategic Autonomy’ by Pankaj Pandey. An AI stack is a collection of technologies, frameworks and infrastructure compo­nents that facilitate using AI systems.

Technologists like Tanish Shewani— who calls himself a “future synthesist” because he prioritises each decision by thinking long-term, meaning how a prod­uct will impact the future—warn against injudicious use of natural resources. Then there is this question of bang for the buck.

Zoho’s Vembu, in a recent interview, said that AI is entering bubble territory. “For all the investments already made, we must be seeing dramatic gains in at least knowledge work. We are seeing some, but that’s not enough to [be] commensurate with the amount of investment going in,” he said. Again, a country like India faces challenges in training language models because of linguistic diversity—primarily because a massive amount of high-quality web data is available in English, while In­dian languages collectively make up less than 1 per cent of online content.

It is estimated that AI could add $1.7 trillion to India’s economy by 2035. A KPMG report released this year notes that 57 per cent of Indian CEOs plan to allocate 10-20 per cent of their budgets to AI over the next year, signalling growing excite­ment around the technology. Yet myriad problems continue to surface even as Indian policymakers make ambitious statements about next-generation data centres powered by renewable energy and other futuristic goals. Among these challenges are the caste, religious, and racial biases embedded in large language models (LLMs), which often reinforce stereotypes specific to India. Recently, despite the euphoria among some analysts about the Indian AI story, investment research firm Bernstein warned in late September that a repeat of the internet-era competition cannot be ruled out: “US players, with deep pockets and stronger infrastructure, are poised to win all over again—by using low prices to bury any home-grown alternative, just like last time.” In fact, the disruption caused by China's DeepSeek LLM had significantly inspired India’s AI efforts. Interestingly, a comparison of the most widely used AI products in India and China shows that while US companies hold a near-monopoly in the Indian market, Chinese firms dominate theirs.

Now, a report by the Germany-based Friedrich Naumann Foundation for Freedom has warned of an emerging AI divide, akin to the digital divide, that could trigger a brain drain from low- and middle-income countries (LMICs). “With­out targeted efforts to expand AI-related employment, offer competitive salaries, and nurture local innovation ecosystems, LMICs risk perpetuating the cycle of tal­ent outflow, further widening the global AI divide,” the report states.

It looks like when it comes to quickly catching up with AI development, India will have to rely on its history of jugaad innovation combined with bolder action from key stakeholders.