Contrary to the broader belief that artificial intelligence is building corporate buildings cheaper and faster, leading venture capitalists (VCSs) are discovering the opposition, AI startups have proven more capital-intensive than their traditional counterparts.
Banglani's revelation emerged from a panel discussion on “AI As Catalyst: Early Bets & Growth Levers.” There, Harshjit Sethi, managing director of Peak XV Partners, who will be seated on the board of Sarvam AI, focusing on Indian language models.
Speaking on a panel hosted by Ettech.com's Pratik Bhakta, Sethi said, “I don't believe it's necessarily cheaper,” working on the general assumption that AI will reduce startup costs. It features a lineup of daily sessions that navigate the AI and the startup ecosystem in India. This was the ET Sopenorns Summit 2025, India's largest congregation of SONICORNS, and returned to Bengaluru in the fourth edition on August 22nd.
Capital Reality Check
Although there is no denying the enthusiasm for investing in AI in India, the context is important. Seti offered a calm perspective. “Funding for Indian startups is probably double digits lower than what we see in the valley (Silicon Valley), some of which is driven, of course, by the large labs that exist in the valley.”
Despite these higher capital requirements, investors' convictions are strong. Banglani's portfolio demonstrates this commitment. “In the last two years, we've invested in 12 AI companies. And roughly, it's easy to 60% of all investments over the last two years, which is roughly, 60%.” This portfolio spans five Services as AI Software (SAAS) companies, two consumer applications, two development software companies, and three AI services companies.
Seti agrees with the sector's potential and points out that “as we move beyond the model layer to the application layer,” and the opportunity in India becomes more pronounced. “The hope is to start seeing more companies play here, and if you do, it'll see people like solving the real problems, population, size, applications that come out of India.”
tThe economics of smarter construction: more value than cost reduction
The enduring story suggests that AI will build the company cheaper and faster. Both investors challenged this assumption, particularly with regard to capital requirements. “If you believe software is a product, many coding agents will come to believe that at least it's been significantly easier. The problem is that you're still competing for the same share.”
Banglani strengthened this point with evidence at the ground level. “I don't know if they tracked the cost of LLMS tokens. I think a lot of the costs of AI companies today are moving in that direction.
Even more interesting, Banglani shared insights from an enterprise conversation that uncovers the fundamental changes in AI adoption drivers.
theCrassing Benefits: Scale Over Speed
The true power of AI comes into play to enable newcomers to fundamentally rethink their established business models. Banglani explained this with a compelling example from the travel sector. Startups that use AI rather than travel planning – “The easiest thing in the world” – but to sell holiday packages through voice bots that certify and determine customer preparation.
“Now, can MakemyTrip do that? Banglani asked, highlighting the adjustment costs that are bothering the incumbent.
Seti repeated this feeling. “One good thing I saw with some of the AI native companies is that in general, teams tend to be much less. So the adjustment costs, the cost of actually getting feedback from users, understand what to build right away.
The Fintech sector illustrates the possibility of this conversion. Banglani outlined how AI is revolutionizing the entire stack. “Voice bots for distribution, lead entitlement, funnel management, funnel management, and onboarding. The KYC process is completely converted by AI, backend processes… The collection is much more efficient with the use of AI.”
Talent and capital challenges
Despite optimism, both investors acknowledged important structural challenges. Lack of talent in the basics presents a specific bottleneck. “The business that requires you to hire 20 doctoral degrees in India is very, very difficult to build. Where are the 20 doctoral degrees with the necessary expertise?” Banglani asked.
Seti highlighted a particular challenge. “If you look at the number of AI doctorates in India, the numbers are just small in the Valley and China.”
The capital requirements of the basic model further demonstrate Starker's reality. “I don't think any of us have capital. Even all of us don't have the kind of capital necessary to fund our basic model business from scratch,” Banglani admitted. “The number of GPUs Deepseek has is not our entire country.”
“eVery much wanting to build AI. Someone has to build the tools.”
Despite these challenges, both companies have placed specific vertically concentrated bets. Sethi outlined three priority areas: In Consumer AI, India's founder “consumer DNA” offers a competitive advantage. Vertical software that allows the industry to “leaps” into adopting traditional software through the operation of agent AI. And there are developer tools, realizing that “everyone wants to build AI, someone has to build the tools for it.”
Banglani's portfolio strategy closely aligns, highlighting vertical enterprise applications, AI-driven services that disrupt traditional IT and business process outsourcing models, and consumer products where AI enhances both the creation and distribution of services.
Perhaps most importantly, both investors have confirmed that AI has become the dominant conversation topic in the boardrooms of the entire portfolio. “AI is the most consensus view that everyone has,” Seti observed. “That's not possible. Forget scaled startups. I would be surprised if they weren't having a conversation with AI, like banks and public sector institutions.”
However, consensus on importance does not lead to clarity in implementation. “I think it's really difficult for incumbents to run here faster,” Banglani warned. “And according to the current position, I mean, of course, large companies, but perhaps large tech companies.”
This execution gap creates what both investors consider as a generational opportunity for newcomers. “Whenever there is a technical inflection, there is an opportunity for startups to sell the exact same thing in new and more efficient ways that the incumbent sells,” Banglani concluded.
The panel's insights reveal the ecosystems in the transition. Traditional playbooks are rewritten by people who are willing to rebuild from Indigenous principles, not by those with the deepest pockets or the largest team. For founders and investors navigating this landscape, the message is clear. The wave of AI is not just about adopting technology. It also concerns basic business model innovation. The question is not whether AI will rebuild the industry, but whether established players can adapt quickly enough to build AI-Native solutions from day one.
360 One is the presentation partner for ET Sopnicorns Summit 2025, with Shiv Nadar University as ecosystem partner, Raymond as Wardrobe partner, Pi42 as Gold Partner, Indian banking partner, Tracxn as knowledge partner and K-Tech startup Karnataka as state partner. Summit's gift partners are Mind & Company, Plum, Clinically, EM5, and True Elements.
