India's artificial intelligence (AI) story is slowly playing out, but for the true unlocks to happen, there needs to be large-scale adoption of the technology, according to panellists at the Mint India Investment Summit 2026.
India's AI infrastructure buildout is well on its way, with firms such as Yotta Data Services, Neysa Networks, and Sify Technologies all focused on ensuring there are enough data centres to meet the rising demand from companies here.
“We as a country are waiting for some UPI moments to come to AI, where some government-to-consumer use cases adopt the scale of AI,” said Sunil Gupta, co-founder, managing director and chief executive at Yotta.
Just as the Unified Payments Interface (UPI) has become a regular fixture in people's lives in India, processing billions of transactions every year, so too will AI. “One UPI moment in AI will result in multiple AI moments later,” Gupta said.
While AI has caused a global shift, Indian companies, startups and the government have realized the country also needs to ensure there is high-performance compute capacity available to those working on the technology.
This is evident in how the IndiaAI Mission has been working to acquire 18,000 high-end graphics processing units (GPUs) needed for computations of complex maths that run the algorithms for AI.
But while there is definitely excitement about AI, there are also performance indicators to keep in mind, according to Sandip Patel, managing director at IBM India & South Asia. “It's not so much about the models now. It's whether you have the infrastructure that you can deploy and use responsibly. Whether you have the right kind of AI, which, with governance, can be trusted with its outcomes.”
He added that it was important that companies have the right data for specific use cases to deliver meaningful outcomes that push businesses in the right direction. “They need to do all of this with the right kind of cost optimization and sustained RoI. That becomes very critical,” Patel added.
For enterprises, making sure that their investments in AI are actually beneficial has become crucial. But where AI companies struggle, especially in the enterprise, is the mismatch between what they create and the value their clients actually seek.
“We've realized that enterprise AI is really hard for three reasons: Most are designed for engineers, by engineers, not for business users. Secondly, there is an engineering challenge due to lots of disparate data spread across multiple systems,” said Ajoy Singh, chief AI officer, at recently listed AI company Fractal Analytics. “Thirdly, the pace of change. What was state-of-the-art six months ago becomes outdated.”
But that doesn't mean that there aren't pockets where enterprises can find value. In India, several companies in the banking, financial services and insurance segment have opted to use voice-first AI in some of their systems, such as loan collection or debt reminders.
“On the voice side, enterprises have decided to take this to the next level,” said voice-first AI startup Gnani.ai's co-founder and CEO Ganesh Gopalan. The company currently caters to 150 enterprise customers in India and approximately 200 worldwide.
According to Gopalan, the reason voice AI is reaching scale is that existing systems for customer interaction and engagement are currently broken. “Some of the property banks in India that we work with found that our AI systems are about 10% better than their existing systems. At the enterprise level, especially for larger banks, evaluation is always done. It's always going to be based on RoI,” he said.
For enterprise companies such as Gnani.ai, and even for Fractal Analytics, support, both in terms of policy and GPU access, through the IndiaAI Mission has played a crucial role. Even so, the government has made it clear that while they're not going to regulate AI and risk innovation, there have to be checks and balances.
“The underlying factor for the government, even while attracting investment, is that innovation has to have human oversight. Companies are not yet at that stage where the machines will run themselves. And that human oversight and therefore the human responsibility and accountability need to be there,” said Nisha Uberoi, partner at JSA Advocates & Solicitors.
Rwit is a correspondent at Mint. He writes on AI and SaaS startups alongside the venture capital firms that invest in these companies. Currently based in Bangalore, he's an alumnus of the Asian College of Journalism, Chennai.
Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
Facts Only
India is building AI infrastructure, with companies like Yotta Data Services, Neysa Networks, and Sify Technologies expanding data centers to meet rising demand.
The IndiaAI Mission is working to acquire 18,000 high-end GPUs for AI computations.
Sunil Gupta, co-founder and CEO of Yotta, stated that India needs a "UPI moment" in AI—large-scale government-to-consumer adoption—to drive widespread use.
Sandip Patel, managing director at IBM India & South Asia, emphasized the importance of infrastructure, responsible deployment, and governance for AI trustworthiness.
Ajoy Singh, chief AI officer at Fractal Analytics, identified three challenges in enterprise AI: tools designed for engineers, disparate data systems, and rapid model obsolescence.
Voice-first AI is gaining traction in India, particularly in banking and financial services, with Gnani.ai serving 150 enterprise customers in India and 200 worldwide.
Gnani.ai's voice AI systems have shown a 10% improvement over existing customer interaction systems in some Indian banks.
The Indian government supports AI through the IndiaAI Mission but avoids heavy regulation, emphasizing human oversight and accountability.
Nisha Uberoi, partner at JSA Advocates & Solicitors, noted that companies are not yet at a stage where AI can operate without human responsibility.
The Mint India Investment Summit 2026 featured discussions on AI adoption, infrastructure, and governance.
Enterprises are prioritizing cost optimization and ROI in AI investments.
The banking, financial services, and insurance (BFSI) sector is a leading adopter of AI in India, using it for loan collection and debt reminders.
Executive Summary
India's AI ecosystem is gaining momentum, with infrastructure development and enterprise adoption accelerating, though challenges remain in scaling and governance. Key players like Yotta Data Services, Neysa Networks, and Sify Technologies are expanding data center capacity to meet rising demand, while the IndiaAI Mission aims to acquire 18,000 high-end GPUs to support AI computations. Industry leaders emphasize the need for a "UPI moment" in AI—government-backed, large-scale consumer adoption—to catalyze broader implementation. Enterprises are focusing on voice-first AI solutions, particularly in banking and financial services, where systems like Gnani.ai's voice AI have demonstrated measurable improvements in efficiency and ROI. However, challenges persist, including the mismatch between AI tools designed for engineers and business needs, the rapid obsolescence of AI models, and the need for responsible governance. The government is balancing support for innovation with calls for human oversight to ensure accountability. While optimism about AI's potential is high, stakeholders stress the importance of cost optimization, data quality, and sustained value delivery to justify investments.
The discussion highlights a tension between rapid technological advancement and the practical realities of enterprise adoption. Voice AI has emerged as a scalable use case, but broader AI integration requires overcoming engineering challenges and ensuring alignment with business objectives. The government's role in providing GPU access and policy support is seen as critical, though regulatory restraint is maintained to avoid stifling innovation. The conversation underscores that while AI's transformative potential is recognized, its success in India hinges on addressing infrastructure, governance, and real-world applicability.
Full Take
The narrative around India's AI adoption presents a compelling vision of technological progress, but it also reveals deeper tensions between ambition and execution. The strongest version of this story—what we might call the "steelman"—is that India is on the cusp of an AI revolution, with infrastructure buildout, government support, and enterprise innovation converging to create scalable, high-impact use cases. The comparison to UPI is particularly effective, framing AI as the next transformative leap in digital infrastructure. This narrative gives credit where it's due: the government's IndiaAI Mission is actively addressing compute capacity, and companies like Gnani.ai are demonstrating tangible ROI in voice AI, a rare bright spot in enterprise adoption.
However, the pattern scan reveals subtle distortions worth noting. The repeated emphasis on "UPI moments" risks oversimplifying AI's complexity, implying that a single breakthrough will solve systemic challenges like data fragmentation and model obsolescence. This could be an instance of **ARC-0024 Ambiguity**, where a vague but evocative analogy obscures the harder work of incremental progress. Additionally, the focus on voice AI as a scalable success story might inadvertently downplay the broader struggles in enterprise AI, where most tools remain misaligned with business needs—a potential **ARC-0043 Motte-and-Bailey**, where a narrow win is presented as proof of broader viability.
The root cause of this narrative is a familiar paradox in tech adoption: the tension between top-down infrastructure pushes and bottom-up market realities. India's AI story echoes historical patterns of digital transformation, where government-led initiatives (like UPI) create the rails, but private innovation determines the pace. What goes unstated is the risk of overinvestment in compute capacity without commensurate advances in data governance, talent development, or business-model innovation. The assumption that AI will follow UPI's trajectory ignores the unique challenges of AI—its opacity, rapid evolution, and ethical dilemmas.
The implications for human agency are significant. While AI promises efficiency gains, the emphasis on human oversight (as noted by Nisha Uberoi) suggests a recognition that accountability cannot be outsourced to algorithms. Yet, the narrative largely frames AI as a tool for enterprises and governments, with less attention to how it might empower (or disempower) individual citizens. The second-order consequences could include a widening gap between AI-haves and have-nots, as only well-resourced enterprises can afford the infrastructure and talent needed to deploy AI responsibly.
Bridge questions to consider: What would it take for AI to become a truly inclusive technology in India, beyond enterprise and government use cases? How might the focus on voice AI in banking obscure other sectors where AI could drive social impact? And if the "UPI moment" for AI doesn't materialize, what alternative pathways could ensure sustainable adoption?
Counterstrike scan: If this were a coordinated influence campaign, the playbook would involve amplifying success stories (like voice AI) to create a sense of inevitability, while downplaying systemic challenges to attract investment. The actual content does highlight obstacles (e.g., enterprise AI's difficulties), but the framing leans toward optimism, which could serve the interests of AI vendors and investors. However, the inclusion of critical voices (e.g., Fractal Analytics' Ajoy Singh) and governance concerns suggests a balanced rather than manipulative approach. No structural alignment with a hypothetical attack pattern is detected.
Sentinel — Human
The article shows strong signs of human authorship, with natural variability in sentence structure, specific attributions, and idiosyncratic phrasing. No significant indicators of synthetic generation were detected.