Providing a nuanced view of India’s evolving startup landscape in FinTech, AI, and consumer internet sectors, Harshjit Sethi, Managing Director at Peak XV Partners, shares his perspective with Fortune India on the impact of rising capital costs, which have driven a sharper focus on profitability and cash efficiency across the board. Sethi also highlights the potential of Indian language LLMs to unlock growth in customer-centric sectors, while also noting how AI-driven automation could reshape the workforce in financial services and telecom. Edited excerpts.
There was a lot of talk about a “funding winter” recently, with companies finding it harder to raise capital. As an investor, how are you viewing this? Are you seeing more discipline or rational behaviour among founders regarding capital? Is there a self-realisation, or are investors prompting them to scrutinise every dollar spent?
It’s a bit of both. Some founders are becoming more cautious on their own, while in other cases, it’s a push from the board. Across the ecosystem, there’s a much greater appreciation for building profitable businesses today than three years ago, when capital was more readily available. Now, there’s a shared understanding that capital efficiency and profitability are essential.
A key driver behind this shift is the performance of Indian public markets, which have shown that profitability is rewarded. Many founders aspire to go public, and they realise that to achieve strong valuations from public investors, they need to prioritise profitability. So, the increased cost and scarcity of capital, combined with the desire for a successful IPO, have led both founders and investors to focus on building more capital-efficient companies.
That said, it’s important to strike the right balance. Early-stage companies need to be prudent, but being overly conservative can limit growth potential. Venture-backed companies are expected to scale with a certain level of non-linearity, and if that is constrained too much, the growth model can falter. So, finding a balance between profitability and growth is essential.
With public market corrections underway, we’re seeing significant markdowns in private portfolios ahead of IPOs. Do you think this trend of markdowns will become more mainstream? Will retail investors and Qualified Institutional Buyers be the ones who will act as VCs and wait for profitability rather than investors themselves?
In my experience, particularly from speaking with investment bankers, it’s challenging for a company to go public without a clear path to near-term profitability. In our portfolio, we wouldn’t encourage a public offering unless the company is either already profitable or has a defined path to profitability. Going public without this would make for a tough journey as a public company.
For any of us to exit successfully, there needs to be a buyer on the other side who values profitability. So, I don’t foresee a scenario where mutual funds or retail investors essentially take on the role of venture capitalists.
However, what is happening more frequently is secondary transactions. If companies show progress toward profitability or at least slowing losses, early investors may exit through a sale to another private investor who’s willing to hold the investment for the next 5-7 years, rather than an IPO. This allows early-stage investors to partially exit without relying on public markets.
So, instead of a direct shift from funds to the public market, it’s more a transfer from one private fund to another in cases where companies aren’t yet fully profitable. But for companies with a clear path to becoming large, profitable businesses, the public markets remain the goal.
So, it’s crucial to find the right balance between optimising for profitability and pursuing growth. There’s a trade-off here, and this balance varies for each company based on factors such as its stage, the capital used thus far, its pace of growth, and the right unit economics.
With the end of the era of cheap money, how does this shift impact your approach to investment and return expectations? Are you going hard on evaluation before cutting a cheque?
I don’t know if underwriting itself has changed significantly, but the amount of capital companies need to reach a large outcome has. For early-stage investors, the difference between investing at a $10 million or $15 million valuation isn’t as critical because the real success depends on whether the company can grow to a half-billion, billion, or multi-billion-dollar valuation. For an early-stage investor, the real question is whether you’re investing in the right company with the potential to generate multi-fold returns. You’re not going to make substantial returns simply by optimising between a $10 million and $15 million valuation. However, for late-stage investors — those who underwrite based on future public market multiples and have traditionally provided substantial capital to fuel growth — there’s been a notable shift toward prudence. This segment of the market is now more conservative, which means we’ll likely see a new wave of companies that consume less capital to reach significant outcomes.
Additionally, as AI drives efficiency, we may see companies reducing operational costs and building leaner teams. For instance, companies adopting AI solutions, such as Copilot for engineering or automation in sales, can streamline operations and reduce headcount. If you’re promoting these efficiencies to customers, it makes sense to adopt them internally as well. So, while it’s hard to envision a return to low-cost capital, we’re likely to see businesses that are more capital-efficient and adaptable to the current economic environment.
Have you observed a noticeable shift in the mindset of Indian entrepreneurs regarding business and capital, compared to five or six years ago?
Really, if you look at what companies like Zomato, Nykaa, and Policy Bazaar have managed to do, they’ve shown that you can create billions in value even in the public markets. They’ve demonstrated the path, and I think we have enough positive role models now to see what that journey looks like.
Then, you have a number of companies like Ixigo, Truecaller, and Five Star Finance—venture-backed companies that went public and have done well in the public market. So I’d argue that we now have many good examples of companies that have continued to build and thrive.
We’ve got a pretty significant group of companies here showing that if you create value, you can build something large. Freshworks, for example, went public, and it’s been talked about a lot. I think we now have enough positive examples of companies that have gone from zero to IPO, showing what that journey involves. Two or three years back, there were very few examples, hardly anything to look forward to—it was all a bit of a guess. But now, I’d say public markets have shown that a lot is possible.
Where do you see the market for Indian language large language models (LLMs)? Do you think building products specifically for India will be a lucrative segment? I recently covered a story on indigenous AI in India. For example, a founder created Kisan.AI, a model specifically designed for Indian farmers, but he finds it more lucrative to work with agri-tech clients in the U.S. and Brazil. The potential in India seems exciting, but how do you view it as an investor?
While it’s still early, and data on the potential market size is limited, I firmly believe there’s a significant business to be built here. Take customer support as an example. Many large call centres in India handle tasks such as customer support and prospecting calls, often through phone or WhatsApp interactions. Much of this work, currently managed by human agents, could be automated with AI, and this change captures both software and human resource expenditures.
When you consider software spend plus the people spend, the market opportunity becomes substantial. Large consumer-facing companies in India — banks, airlines, telecoms, and similar industries — employ tens of thousands in customer-facing roles. Automating these roles with AI could be transformative, especially in a multilingual market like India. So, while its early days, I see a strong potential for Indian language LLMs to create meaningful value in sectors with heavy customer interaction.
Within your own portfolio, what are some companies you’re particularly excited about, and why do you believe they have strong growth potential?
On the AI side, there are several companies we’re excited about. For example, Sarvam, which we discussed earlier, is building multilingual large language models and targeting the voice interaction space. They’ve already secured pilot projects with several top Indian conglomerates across different sectors, which highlights the potential for AI in localised, voice-based solutions.
Another company is Enterpret, which we led the seed round for. Enterpret aims to disrupt traditional SaaS approaches by rethinking customer feedback management. For instance, today, companies like Qualtrics — a $10 billion SaaS company — rely on structured surveys sent to customers. However, Enterpret leverages unstructured data from customer interactions, such as calls, social media posts, or feedback shared directly with customer support. In the LLM era, where analysing unstructured data is more feasible, Enterpret helps consumer-facing companies such as Figma, Notion, and Canva understand customer feedback without relying on traditional surveys. This approach enables companies to gain richer insights from organic customer interactions, moving away from the conventional survey model altogether.
Another thesis we’re exploring is “Service as Software.” Essentially, this involves delivering outcomes with a recurring billing model or per-interaction pricing. AI-driven companies can help bridge skill gaps, such as that between a business analyst who works in Excel and a data scientist who codes in machine learning languages. With generative AI, analysts can describe problems in natural language, and the system translates those into actionable data insights without requiring advanced technical training. This approach effectively upskills people by enabling them to work directly in their area of expertise without learning new technical skills.
And then you have some of our existing companies asking themselves, “What will our product look like in an AI-driven world?” Take Darwinbox, for example—they’re exploring how AI can transform common HR interactions. A typical HR team fields questions like “What’s the expense policy?” or “How much leave do I have?” People want to apply for leave around Diwali, or check their relocation benefits, and all these interactions require processing by an HR business partner. Now, imagine if all of that could be automated so that people simply send an email, and it’s handled instantly by AI.
They’re looking at the same idea on the talent side. Say you’re a company such as Bharti Airtel, and you receive thousands of resumes for a single job posting—most of which aren’t relevant. Traditionally, HR needs to sift through all those applications manually, but this is exactly where LLMs can make a big difference. With LLMs, you can specify what you’re looking for, and the system can filter out the most relevant 10 or 15 candidates from a large pool.
So, there’s a segment of software companies thinking, “How can we use AI to enhance what we already offer?” And then there are newer companies exploring entirely new problems that couldn’t be solved before but are now possible with generative AI.
We’ve seen traditional conglomerates such as Tata and Reliance increasingly acquiring or investing in new-age startups to expand their consumer focus. How well do you think these established, industrially rooted companies can integrate these acquisitions? Are these startups better off going public independently, or can they achieve more as part of a large conglomerate?
For a healthy ecosystem, both paths — acquisitions and IPOs — need to coexist. Some startups will go public, while others will find a natural fit within a larger conglomerate. For the ecosystem to thrive, having established players who can acquire and support new-age businesses is essential. Acquisitions not only provide an exit route but also validate the value these startups create. Zomato showed that with what happened with Blinkit…
But Zomato was always a digital-first company…
Yes, I agree. I’m talking about, for instance, Titan’s acquisition of CaratLane shows the potential of combining traditional brand strength and distribution with a newer, consumer-oriented digital business.
In India, we haven’t seen as many acquisitions as we should, and there’s room for more traditional companies to engage with and acquire independent startups. This creates opportunities for value creation on both sides — the conglomerate leverages fresh innovation, while the startup gains access to larger resources and markets.
At the same time, startups shouldn’t only build with the expectation of an acquisition; they should focus on creating long-term value. If they do, acquisition opportunities or IPOs will naturally become viable paths. Both options are important, and as the ecosystem matures, we’ll hopefully see a balance between independent growth and strategic acquisitions.