Gireendra Kasmalkar, managing partner at Pentathlon Ventures LLP, argues that the future of artificial intelligence (AI) in India does not depend on metros, but looks deeper and beyond. In an interview, Kasmarker said he believes innovation thrives on depth, not spectacle, and that the founders who will define the future of AI in India are not the loudest voices about GenAI. He feels that India needs to foster this depth if it is to build a meaningful AI future. We have to support these founders. In that sense, this Pune-based VC's perspective is not just a critique of the hype, but an invitation to build with intention.Edited excerpt:
You often describe your professional life as unfolding inning by inning. How does that framework shape the way we understand innovation today?
Living your career innings-by-inning gives you a kind of longitudinal perspective that most people tend to miss. My early years, my first innings, were driven by necessity. I built it because I had to, not because I had the luxury of knowledge or sophistication. It was an era defined by raw execution and survival. The second time, the development expanded further. I founded a global software testing company, learned how scale actually worked, experienced the intricacies of operational details, and eventually moved to a publicly traded company in Germany. This acquisition taught me the discipline of enterprise technology, global thinking, and the harsh realities of running a B2B company at scale.
It's the third inning now, and I feel completely different. It's no longer about building just for yourself, it's about enabling others. Through Ideas to Impacts and Pentathlon Ventures in Pune, I had the privilege of watching founders build their first innings. This advantage fundamentally shapes how I view innovation today. It makes me realize that innovation is not just a bolt of lightning. It's a long continuum where each inning builds on the last. Whether it's QA automation years ago or AI today, the tools change, but the basic rhythm remains the same. That means understanding real problems, being close to your customers, and honing your skills over years, not months.
You've said many times that innovation doesn't just belong to subways. Why do you believe the future is distributed?
Because I have seen it firsthand, not as a theory, but as a living reality. When I was running my previous company, a surprising pattern emerged. Almost half of my employees were from towns and small cities outside Pune. These are places that no one thought of as startup hubs, but the talent that has emerged from them has been extraordinary. They were sharp, down to earth, and deeply aware of the issues around them. What they lacked was not ability, but exposure, access, confidence, and the ecosystem that Metro takes for granted. Once we built the Idea to Impacts ecosystem, amazing things started happening. We've seen entrepreneurship surface in places invisible to the mainstream venture capital narrative. Cities like Surat, Nashik and Kolhapur, which rarely feature on the startup map, have started producing founders with great clarity. One of our recent portfolio companies is in Surat, but very few have venture capital connections. But the founders' understanding of their space was sharper than half the pitches seen from traditional hubs. As talent is distributed, so is the future. What is not distributed is opportunity, access, and capital. And that is slowly starting to change. As this landscape evolves, we will see that the next wave of meaningful B2B innovation in India will not just come from Bengaluru or Gurgaon. It will come from the quiet corridors of India, where there is a much deeper understanding of real issues.
Are there differences in how metro and non-metro founders approach problem solving? Has this shaped the way Pentathlon evaluates founders?
There is no single archetype, but there are recognizable patterns. Founders in big cities tend to be more sophisticated. They understand the vocabulary of fundraising, know how to package proposals, and are often good at communicating in investor-friendly language. On the other hand, founders from smaller cities often bring very different strengths to the table. You can get a sense of raw transparency. They tend to be close to the problem, the customer's world, and the underlying economic and operational issues.
Pentathlon is more about substance than style. We look for founders who truly understand the space in which they operate. A founder who can explain not just what the problem is, but why the problem persists, what the customer is feeling, and what the real means of improvement are. This depth cannot be faked, nor can AI manufacture it. It was born out of many years of struggle with reality in the field. If a founder has that depth, we can help them with presentations, fundraising, scaling, recruiting, and all the other aspects. But there are no shortcuts to depth.
You've been vocal about the dangers of building AI for AI's sake. Why is this so dangerous in today's environment?
Because AI has made it dangerously easy to create the illusion of progress. Create prototypes using AI. You can generate pitch decks using AI. AI tools can also help you write code faster. But this doesn't get you closer to your actual customers. None of these lead to understanding the domain. None of these validate whether the problem needs to be resolved.
When founders start with a technology they want to build with AI, they often end up looking for problems after the fact. This is the worst direction. The tool becomes the hero, not the problem, and the startup becomes the hammer looking for the nail. If you remove AI from a startup and the idea dies, then the idea had no reason to exist in the first place. That's the core of what I call AI cleaning. Although impressive from the outside, there is no real value inside.
Real innovation starts on the other side. Start with the customer, the problem, the pain point. If AI is the best way to solve that, that's great. If not, that's fine as well. AI is just a tool. It's a powerful tool, to be sure, but it's still just a tool.
Distinguish between AI-first and AI-native startups. Could you please explain this difference in detail?
This distinction is important but often misunderstood. AI-first startups use AI to reimagine existing workflows and use cases. The problem already existed. Workflow already existed. But with AI, you can rethink processes, automate parts of them, and turn traditional manual tasks into intelligent ones. The opportunity lies in redesigning what people are already doing. AI-native startups are fundamentally different.
These are products and experiences that could not exist without AI. Think of something like ChatGPT. Without a language model, this category would not exist. AI-native startups tend to be more horizontal, more open-ended, and often B2C in nature. Most Indian founders do not need to pursue AI-native ideas or feel pressured to do so. The real opportunity in India lies in AI-first innovation that takes deep-rooted B2B pain points and uses AI to solve them 10x better than before. This is where India's strengths truly align.
From your perspective as an investor and mentor, what makes a truly strong AI startup?
Powerful AI startups are not defined by the sophistication of their models or the cleverness of their technical architecture. It is defined by its relationship to the problem it seeks to solve. The first and most important factor is deep domain expertise. Without this, AI becomes a decoration. AI systems are completely context-dependent, and context comes from a deep understanding of the domain.
The second factor is access to relevant, high-quality company data. This is the true moat of AI. LLM is available to everyone, but not everyone has access to the unique datasets that drive domain-specific intelligence. The third factor is the existence of real, validated use cases rather than hypothetical scenarios. AI-based startups must be solving problems that exist today, not ones imagined on paper.
Finally, startups must be anchored in a niche with depth. Trying to be everything to everyone rarely works. The most effective AI companies focus on well-defined verticals and work at unusually deep levels. This depth makes them competitive and defensive.
Where do you think India's real AI opportunities lie: Hardware? LLM? Applications?
India's opportunity is not to compete directly with global giants in hardware or LLM development. These are capital-intensive fields that require billions of dollars and huge talent densities. Rather, India's strength lies in the sheer size and diversity of its corporate environment. Indian companies operate in a complex and fragmented environment. They generate proprietary datasets, accommodate multilingual interfaces, and struggle with compliance, logistics, manufacturing, and operations in ways that global companies don't fully understand. This is where Indian AI startups can build something truly defensible. They can train models on corporate data that the rest of the world doesn't have access to. They can solve problems that even global players are not aware of, because those problems are specific to our market. India's AI differentiation comes from the complex middle ground: application layer, domain layer, and enterprise workflows.
You said India is not a DIY market. How will that shape AI adoption?
India is an intermediary economy. In many fields, people do not directly adopt products. They rely on trusted intermediaries, ASHA officials in the health sector, bank correspondents in the financial sector, VLEs in the local governance sector, teachers in the education sector, and shopkeepers in the commercial sector. These individuals act as a bridge between the system and the public. AI in India will scale only if it augments these intermediaries, rather than bypassing them. Imagine a bank correspondent equipped with an AI assistant helping rural customers understand loan terms in their local language. Also, imagine teachers using AI-powered tools to provide personalized learning paths. These intermediaries bring trust, and trust is India's most valuable currency. AI should empower them, not replace them.
As AI becomes central to the economy, what should be India's policy stance?
India should not try to build the next OpenAI. It's a distraction. Instead, India should take inspiration from what worked with India Stack. We need to create an enabling ecosystem, a public infrastructure, on which private companies can build. That means making more datasets accessible in a secure and anonymized format, accelerating the development of Indian language models, expanding access to computing resources, and creating a regulatory sandbox where startups can experiment without fear. Policies should enable innovation, not dictate its direction. Our strength lies in our ability to build large-scale digital infrastructures that significantly reduce the cost of innovation. The same philosophy should guide our AI strategy.
