“We joke internally that founders seem to want to start only two kinds of companies, AI (artificial intelligence) and quick commerce in India,” remarked Ritesh Banglani, Partner at Stellaris Venture Partners, at the recent ET Soonicorns Summit 2025 in Bengaluru. But beneath this entrepreneurial enthusiasm lies a counterintuitive reality that’s reshaping how investors think about the next wave of startups.
Contrary to widespread belief that artificial intelligence will make company building cheaper and faster, leading venture capitalists (VCs) are discovering the opposite: AI startups are proving more capital-intensive than their traditional counterparts.
Banglani’s revelation emerged from a panel discussion on ‘AI as Catalyst: Early Bets & Growth Levers,’ where he was joined by Harshjit Sethi, Managing Director at Peak XV Partners, who sits on the board of Sarvam AI, a company focused on Indic language models.
Speaking at the panel moderated by Pratik Bhakta of ETtech.com, Sethi stated: “I don’t necessarily believe it’ll be cheaper,” addressing the common assumption that AI will reduce startup costs. Featuring a day-long lineup of sessions navigating AI and the Indian startup ecosystem, the ET Soonicorns Summit 2025, India’s largest congregation of soonicorns, returned for its fourth edition to Bengaluru on August 22.
The capital reality check
The enthusiasm for AI investing in India is undeniable, but context matters. Sethi offered a sobering perspective: “Indian startup funding is probably two orders of magnitude lower than what we’re seeing in the valley (Silicon Valley). Some of that, of course, is driven by the large lab that exists in the valley.”
Despite these higher capital requirements, investor conviction remains strong. Banglani's portfolio demonstrates this commitment: “Over the last two years, we have invested in 12 AI companies. And just to give you a sense, that is roughly, easily, 60% of all our investments in the last two years have been in AI companies.” This portfolio spans five AI software-as-a-service (SaaS) companies, two consumer applications, two development software firms, and three AI services companies—all AI-native from inception.
Sethi concurred with the sector’s potential, noting that “as we move beyond the model layer to the application layer,” India’s opportunity becomes more pronounced. “The hope is that we will start to see more companies play here, and we'll see sort of people solving real problems, population, scale, applications, if you will, that will emerge from India.”
The economics of building smarter: Value over cost reduction
A persistent narrative suggests AI will make company building cheaper and faster. Both investors challenged this assumption, particularly regarding capital requirements. “If you believe software is a commodity, which a lot of coding agents will lead you to believe that at least it’s becoming significantly easier, then the question is, you’re still competing for the same share of wallet.”
Banglani reinforced this point with ground-level evidence: “I don’t know if you have tracked the cost of tokens on LLMs. I think a lot of the expenses of AI companies today are going in that direction. So as of today, at least, I don’t think AI native companies consume any less capital.”
More intriguingly, Banglani shared insights from enterprise conversations that reveal a fundamental shift in AI adoption drivers: “Just yesterday, I was speaking to a bank CEO, and he said that their intent in adopting AI has to do a lot more with improved business outcomes, has to do with increased business velocity, and they are not able to underwrite this on the basis of cost reductions. They are not even looking at cost reductions.”
The disruption advantage: Speed over scale
Where AI’s true power emerges is in enabling new entrants to fundamentally reimagine established business models. Banglani illustrated this with a compelling example from the travel sector: a startup using AI not for trip planning—“the easiest thing in the world”—but for selling holiday packages through voice bots that qualify leads and determine customer readiness.
“Now, can MakeMyTrip do it? Absolutely. Will MakeMyTrip do it tomorrow? I highly doubt it. They have to change all their processes. They will have to—what will they do with the 800 people who are selling holidays?” Banglani asked, highlighting the coordination costs that plague incumbents.
Sethi echoed this sentiment: “The one good thing that we’ve seen in some of the AI-native companies is that, in general, the teams tend to be a lot smaller. So coordination costs, the cost to actually sort of get feedback from a user, understand what to build, and launch quickly. All of that, we see that cycle being significantly faster.”
The fintech sector exemplifies this transformation potential. Banglani outlined how AI is revolutionising the entire stack: “voice bots for distribution, for lead qualification, funnel management, AI for funnel management, and onboarding. KYC processes are being completely transformed by AI, backend processes... collections are getting much more efficient with the use of AI.”
The talent and capital conundrum
Despite the optimism, both investors acknowledged significant structural challenges. The talent shortage at foundational layers presents a particular bottleneck. “Any business that requires you to hire 20 PhDs in India is very, very difficult to build. Where will you find 20 PhDs with the kind of expertise that is required?” Banglani questioned.
Sethi highlighted the specific challenge: “If you looked at the number of AI PhDs that exist in India, vis-à-vis the valley or China, we just have a smaller number of those.”
The capital requirements for foundational models present an even starker reality. “I don’t think any of us has the kind of capital. Even all of us combined don’t have the kind of capital that is required to fund a foundational model business from scratch,” Banglani admitted. “The number of GPUs that DeepSeek has, our entire country doesn't have.”
“Everybody wants to build AI; somebody has to build the tooling”
Despite these challenges, both firms are placing concentrated bets across specific verticals. Sethi outlined three priority areas: consumer AI, where Indian founders’ “consumer DNA” provides a competitive advantage; vertical software, enabling industries to “leapfrog” traditional software adoption through agentic AI behaviour; and developer tools, recognising that “everybody wants to build an AI, somebody has to build the tooling for it.”
Banglani’s portfolio strategy aligns closely, emphasising vertical enterprise applications, AI-led services that can disrupt traditional IT and business process outsourcing models, and consumer products where AI enhances both service creation and distribution.
Perhaps most tellingly, both investors confirmed that AI has become the dominant conversation topic in boardrooms across their portfolios. “AI is the most consensus view anybody has,” Sethi observed. “It is impossible; forget scaled-up startups, I would be surprised if, like our banks and public sector institutions, even they were not having conversations on AI.”
Yet consensus on importance doesn’t translate to execution clarity. “I think it will be really hard for incumbents to be fast in executing here,” Banglani warned. “And by incumbents, I of course, mean large companies, but also possibly large tech companies.”
This execution gap creates what both investors see as a generational opportunity for new entrants. “Whenever there is a tech inflection, there is an opportunity for startups to sell the exact same thing in a new and more efficient way that the incumbents are selling,” Banglani concluded.
The panel’s insights reveal an ecosystem in transition; one where the traditional playbooks are being rewritten not by those with the deepest pockets or largest teams, but by those willing to rebuild from first principles. For founders and investors navigating this landscape, the message is clear: the AI wave isn’t just about technology adoption alone; it’s also about fundamental business model innovation. The question isn’t whether AI will reshape industries, but whether established players can adapt quickly enough to compete with those building AI-native solutions from day one.
360 ONE is the Presenting Partner of the ET Soonicorns Summit 2025, with Shiv Nadar University as the Ecosystem Partner, Raymond as the Wardrobe Partner, Pi42 as the Gold Partner, Bank of India as the Banking Partner, Tracxn as the Knowledge Partner, and K-Tech Startup Karnataka as the State Partner. The Gifting Partners of the Summit are The Mind & Company, Plum, Clinikally, EM5, and True Elements.
Add Now!
Contrary to widespread belief that artificial intelligence will make company building cheaper and faster, leading venture capitalists (VCs) are discovering the opposite: AI startups are proving more capital-intensive than their traditional counterparts.
Banglani’s revelation emerged from a panel discussion on ‘AI as Catalyst: Early Bets & Growth Levers,’ where he was joined by Harshjit Sethi, Managing Director at Peak XV Partners, who sits on the board of Sarvam AI, a company focused on Indic language models.
Speaking at the panel moderated by Pratik Bhakta of ETtech.com, Sethi stated: “I don’t necessarily believe it’ll be cheaper,” addressing the common assumption that AI will reduce startup costs. Featuring a day-long lineup of sessions navigating AI and the Indian startup ecosystem, the ET Soonicorns Summit 2025, India’s largest congregation of soonicorns, returned for its fourth edition to Bengaluru on August 22.
The capital reality check
The enthusiasm for AI investing in India is undeniable, but context matters. Sethi offered a sobering perspective: “Indian startup funding is probably two orders of magnitude lower than what we’re seeing in the valley (Silicon Valley). Some of that, of course, is driven by the large lab that exists in the valley.”
Despite these higher capital requirements, investor conviction remains strong. Banglani's portfolio demonstrates this commitment: “Over the last two years, we have invested in 12 AI companies. And just to give you a sense, that is roughly, easily, 60% of all our investments in the last two years have been in AI companies.” This portfolio spans five AI software-as-a-service (SaaS) companies, two consumer applications, two development software firms, and three AI services companies—all AI-native from inception.
Sethi concurred with the sector’s potential, noting that “as we move beyond the model layer to the application layer,” India’s opportunity becomes more pronounced. “The hope is that we will start to see more companies play here, and we'll see sort of people solving real problems, population, scale, applications, if you will, that will emerge from India.”
The economics of building smarter: Value over cost reduction
A persistent narrative suggests AI will make company building cheaper and faster. Both investors challenged this assumption, particularly regarding capital requirements. “If you believe software is a commodity, which a lot of coding agents will lead you to believe that at least it’s becoming significantly easier, then the question is, you’re still competing for the same share of wallet.”
Banglani reinforced this point with ground-level evidence: “I don’t know if you have tracked the cost of tokens on LLMs. I think a lot of the expenses of AI companies today are going in that direction. So as of today, at least, I don’t think AI native companies consume any less capital.”
More intriguingly, Banglani shared insights from enterprise conversations that reveal a fundamental shift in AI adoption drivers: “Just yesterday, I was speaking to a bank CEO, and he said that their intent in adopting AI has to do a lot more with improved business outcomes, has to do with increased business velocity, and they are not able to underwrite this on the basis of cost reductions. They are not even looking at cost reductions.”
The disruption advantage: Speed over scale
Where AI’s true power emerges is in enabling new entrants to fundamentally reimagine established business models. Banglani illustrated this with a compelling example from the travel sector: a startup using AI not for trip planning—“the easiest thing in the world”—but for selling holiday packages through voice bots that qualify leads and determine customer readiness.
“Now, can MakeMyTrip do it? Absolutely. Will MakeMyTrip do it tomorrow? I highly doubt it. They have to change all their processes. They will have to—what will they do with the 800 people who are selling holidays?” Banglani asked, highlighting the coordination costs that plague incumbents.
Sethi echoed this sentiment: “The one good thing that we’ve seen in some of the AI-native companies is that, in general, the teams tend to be a lot smaller. So coordination costs, the cost to actually sort of get feedback from a user, understand what to build, and launch quickly. All of that, we see that cycle being significantly faster.”
The fintech sector exemplifies this transformation potential. Banglani outlined how AI is revolutionising the entire stack: “voice bots for distribution, for lead qualification, funnel management, AI for funnel management, and onboarding. KYC processes are being completely transformed by AI, backend processes... collections are getting much more efficient with the use of AI.”
The talent and capital conundrum
Despite the optimism, both investors acknowledged significant structural challenges. The talent shortage at foundational layers presents a particular bottleneck. “Any business that requires you to hire 20 PhDs in India is very, very difficult to build. Where will you find 20 PhDs with the kind of expertise that is required?” Banglani questioned.
Sethi highlighted the specific challenge: “If you looked at the number of AI PhDs that exist in India, vis-à-vis the valley or China, we just have a smaller number of those.”
The capital requirements for foundational models present an even starker reality. “I don’t think any of us has the kind of capital. Even all of us combined don’t have the kind of capital that is required to fund a foundational model business from scratch,” Banglani admitted. “The number of GPUs that DeepSeek has, our entire country doesn't have.”
“Everybody wants to build AI; somebody has to build the tooling”
Despite these challenges, both firms are placing concentrated bets across specific verticals. Sethi outlined three priority areas: consumer AI, where Indian founders’ “consumer DNA” provides a competitive advantage; vertical software, enabling industries to “leapfrog” traditional software adoption through agentic AI behaviour; and developer tools, recognising that “everybody wants to build an AI, somebody has to build the tooling for it.”
Banglani’s portfolio strategy aligns closely, emphasising vertical enterprise applications, AI-led services that can disrupt traditional IT and business process outsourcing models, and consumer products where AI enhances both service creation and distribution.
Perhaps most tellingly, both investors confirmed that AI has become the dominant conversation topic in boardrooms across their portfolios. “AI is the most consensus view anybody has,” Sethi observed. “It is impossible; forget scaled-up startups, I would be surprised if, like our banks and public sector institutions, even they were not having conversations on AI.”
Yet consensus on importance doesn’t translate to execution clarity. “I think it will be really hard for incumbents to be fast in executing here,” Banglani warned. “And by incumbents, I of course, mean large companies, but also possibly large tech companies.”
This execution gap creates what both investors see as a generational opportunity for new entrants. “Whenever there is a tech inflection, there is an opportunity for startups to sell the exact same thing in a new and more efficient way that the incumbents are selling,” Banglani concluded.
The panel’s insights reveal an ecosystem in transition; one where the traditional playbooks are being rewritten not by those with the deepest pockets or largest teams, but by those willing to rebuild from first principles. For founders and investors navigating this landscape, the message is clear: the AI wave isn’t just about technology adoption alone; it’s also about fundamental business model innovation. The question isn’t whether AI will reshape industries, but whether established players can adapt quickly enough to compete with those building AI-native solutions from day one.
360 ONE is the Presenting Partner of the ET Soonicorns Summit 2025, with Shiv Nadar University as the Ecosystem Partner, Raymond as the Wardrobe Partner, Pi42 as the Gold Partner, Bank of India as the Banking Partner, Tracxn as the Knowledge Partner, and K-Tech Startup Karnataka as the State Partner. The Gifting Partners of the Summit are The Mind & Company, Plum, Clinikally, EM5, and True Elements.
Add 
as a Reliable and Trusted News Source
as a Reliable and Trusted News Source Add Now!
