From Core AI To Deeptech: How Pi Ventures’ Funding Focus Shifted To Bolster ‘Disruption’

Before 2014, India was home to fewer than 100 deeptech startups, and disruptive technology trends like AI/genAI, IoT, blockchain, or cloud computing did not permeate every burger joint and business dinner. Nevertheless, the country’s unprecedented growth in transformative technology post-demonetisation gave it a definitive push and made it one of the most dynamic economies.

It was also an epoch when fast-evolving artificial intelligence (AI) ecosystems (analytics, generative, and all other variants that will come up by and by) emerged as the undisputed leaders of future tech. They enhanced efficiency and created value for every industry and business function, ushering in new products, services, markets and business models. In essence, the world was at the intersection of data and technology, value chains were shifting and India remained at the forefront of things that mattered.

Manish Singhal, an IIT-Kanpur alumnus and a hardcore techie who had worked for industry behemoths like Tata Elxsi, Motorola, Ittiam Systems and Sling Media, quickly recognised the underlying tech forces that could shape our world. And he firmly believed that AI would soon be a mainstream technology.

His journey as an investor started in 2013 as a cofounder of LetsVenture, a platform investing in early-stage startups. However, he left the company within a year and set up pi Ventures in 2016. Given his unwavering faith in AI as a futuristic technology that could solve intricate problems, he launched an early-stage and AI-focussed fund with a $30 Mn corpus.

“We realised that disruptive applications could be created with AI. So, in 2016, we went out with that thesis and scouted the market. Unfortunately, we could not find big AI companies. But we found some green shoots and saw how people intended to utilise the technology,” Singhal told Inc42 in a one-to-one interaction as part of Inc42’s Moneyball series.

Fund II followed in 2023 with an $85 Mn corpus. But by that time, pi Ventures had undergone a major shift in its investment thesis. With ‘disruption’ as its theme, it now targets deeptech – focussed ventures and made as many as 22 investments in RagaAI, Pixis, Agnikul Cosmos, Wysa, Niramai, Silence Laboratories and other prominent tech startups.

In a free-wheeling conversation, Singhal shared the reason behind the pivot, his transition from a tech expert to an angel and a VC, and how the ‘innovation’ road ahead may pan out. Here are the edited excerpts.

Inc42: Tell us how the investment thesis has evolved at pi Ventures.

Manish Singhal: Thesis recognition is an ongoing process. When we started investing from Fund I, we were able to identify our strengths. For instance, we were good at understanding technology and business problems slightly on the disruptive edge of AI rather than just dealing with incremental interplay. When our views turned out to be right, it gave us a major advantage in the market. Gradually, our focus also shifted from AI to disruptive AI.

When we were happily doing all that, we came across a spacetech startup called Agnikul Cosmos. It was building on-demand launch vehicles capable of taking micro and nanosatellites to low Earth orbit (LEO). We were floored by what it was doing. The only problem was that it did not align with our thesis. So, how would we look at this investment?

We parked our thesis for a bit to explore Agnikul. And it turned out to be very good for us. When we looked at Agnikul with an open mind, we invested there and realised that if we explore ‘disruptive’ as a theme, we would find many other disruptive technologies beyond AI.

Inc42: So, what were the new core areas you identified at the time?

Manish Singhal: When we started working on this thesis – in 2019 and early 2020s – we thought about science as a core area. It’s because materials typically had a defining impact on the various ‘ages’ throughout human history. From the Stone Age to the Iron Age [and later on], materials were discovered and utilised to help us progress. These materials are the essential foundation for innovation.

Earlier, traders used to mine materials or combine them [to create products]. But now we have technologies like synthetic biology and even AI, which could design materials in the lab, adding a whole new dimension to what you can do. That’s why our thesis evolved from AI to disruptive AI to disruptive technologies.

When investing from Fund II, we look at all disruptive technologies, whether AI, other digital technologies or materials science. So, we are a broad deeptech fund now.

Inc42: What are the key pillars of your deeptech investment thesis?

Manish Singhal: In  2016-17, we essentially focussed on AI. But these days, we are looking at three major verticals. The first continues to be AI, but within AI, the focus has shifted from applied AI to the infrastructural side because there are deeper problems to solve in infrastructure.

The second focus area is materials science, where we explore very different things functioning across sectors. In biotech, you can create new molecules and new antibodies. Then there are the alternative foods, alternative batteries, alternative fabrics and semiconductors – we can keep going on.

The third vertical covers technologies like cybersecurity, quantum computing, etc, which may or may not use AI. But their fundamental algorithm is based on some hard engineering and software problems.

These are the broad areas we are looking at today. But it will keep changing with time as technology evolves.

Inc42: Deeptech is becoming broader, more sector-agnostic and more complex with time. But what’s the core definition of deeptech at pi Ventures?

Manish Singhal: Deeptech implies technology going deep, with fundamental innovation and IP (intellectual property) matters. And yes, you are right about being sector-agnostic. If we are investing in deeptech, we are probably investing in a whole bunch of sectors.

The key theme is that deeptech should be able to solve a hard problem in a much more disruptive way. At times, one will come across unsolved issues. At other times, it is about providing a much better solution.

Take, for instance, Niramai, a startup specialising in breast cancer detection. It’s not that breast cancer detection did not exist before. But now you can detect it much sooner. It’s a new product in an old market. There have been solutions which are working; it’s not an unsolved problem. But now you have a new technology which can do it much better.

On the other hand, you will come across situations where the issue has yet to be solved, like Agnikul. Nobody has designed a rocket that can take 30 to 100 kg to space economically. Today, there’s no other alternative to that.

Inc42: What are the key parameters for potential investments into deeptech startups? 

Manish Singhal: When we talk with startup founders, one of the key things we look at is whether they are solving a problem/pain point that needs to be solved. It is crucial from the perspective of business, application and geography. We don’t like localised problems and look at a solution more from a global perspective. Like everybody in the world needs an easier solution to detect breast cancer.

Next comes the deeptech angle. Startups should solve problems using a different approach. Their fundamental innovation should have a 10x technology edge.

Finally, there’s the team that can pull it off. A great team is an essential element in decision-making.

Inc42: What do you think about India’s fast-growing genAI landscape?

Manish Singhal: It’s an interesting space now that it has gone mainstream. But genAI has been here in many forms for five to six years. We invested in genAI when we funded Pixis and Wysa in 2019.

Nowadays, genAI is more like a parallel world to relevance-based LLM (large language model) applications. [LLM is an AI programme that can recognise and generate text, among other tasks, and is trained on huge data sets.]

GenAI is not broader than LLM, but when people generally talk about genAI, they are talking about LLM and thinking that LLM is a part of genAI. But based on conventional knowledge, I would say LLM-based applications instead of genAI.

Of course, there are several startups in the LLM applications space. However, we stick to our core investment thesis and often find that many ventures fall short of the second criterion. We don’t see enough disruption or 10x edge in the solutions. But that’s the whole point of LLM – it democratises technology adoption.

Inc42: How does pi Ventures differ from other VCs? What’s the ‘X’ factor on offer for your portfolio startups?

Manish Singhal: You won’t find too many deeptech funds in India and not many of those folks have actually built deeptech products. We are proud to be an operator-led fund – all our team members have led products and tech in different capacities. This gives us an advantage in connecting with founders, understanding their ideas early on, anticipating the challenges they may face and offering support from an operational perspective. Plus, we always stay up-to-date with the latest developments in technology.

Our founders like us because we can help them solve any technical or operational problem. It can be a straightforward thing like how to pick a salary structure or as deep as how to take the technology to the market. It’s a continuous building process. So, we partner with our founders early in the company-building process with a cofounder mindset.

Inc42: The last two years were difficult from a fundraising perspective. What was your experience when you reached out to LPs while raising Fund II?

Manish Singhal: It took us nearly two years and a half to close Fund II ($85 Mn) starting from 2020. One good thing was that everybody stayed put due to the health crisis and Zoom became an accepted mode for doing business. So, we managed to talk to people all over the globe and held seven or eight meetings every day.

As we have a differentiated thesis and focussed on frontier technologies, we found three buckets of investors. Those who understood deeptech and were excited about new technologies invested readily, while others quickly declined. But there was another investor group somewhere in the middle. They were excited about this domain but didn’t know much about the risks and the opportunities. We had the onus to educate them about frontier technology and deeptech.

Also, our Fund I was doing very well. There were startups in which we invested very early and their growth spoke in our favour. It was an advantage.

Fundraising is tough, irrespective of a global health crisis like Covid-19 or macroeconomic factors leading to a funding winter. For pi Ventures, a self-made fund with an entrepreneurial mindset, it takes two to two years and a half to get through at any time.

Inc42: How did Fund I and Fund II experience differ when it came to the domestic LP landscape?

Manish Singhal: There was a huge difference. Fund I primarily spanned 2016-2019 and we did not get much domestic investment. But when Fund II came along (2020-2022), domestic capital flowed from family offices in India. It was not just our story but a trend across the ecosystem. Hence, we raised a significant amount from India during Fund II.

This shift happened due to many reasons. There are many successful startups in India and the ecosystem is growing at a fast pace. So, investing in a startup fund makes sense. Family offices are also looking to diversify their portfolios. Although they are getting good returns from capital markets, people know a lot about startup success stories and want to explore them further.

Again, some family offices invested in us because they want to stay in touch with the latest technologies. So, there has been a lot more appreciation for this asset class [startups] across the family offices.

Inc42: What should be the focus areas of Indian deeptech startups in FY25?

Manish Singhal: Capital is there if you are working on an interesting idea. However, converting technologies into products is a challenge that founders must undertake. That’s where we come in to help them with the transition.

Profitability is not a concern for us when we invest in deeptech startups, as most of these ventures have a longer gestation period. If a Saas company can start earning revenue in a year, a deeptech company can take three to five years, depending on how big a problem it is solving.

But if they are in the right place and solving the right problem at the right time, their revenue will accelerate faster despite the comparatively long gestation period.