1Mby1M Virtual Accelerator AI Investor Forum: With Krishnakumar Natarajan, Co-Founder of Mela Ventures (Part 4)

Sramana Mitra: Yes, definitely. I also want to ask you about your current strategy of investing in AI companies. It sounds like this is a team that you really liked; and you invested in that team and the concept. How often do you do that? Do you invest in concepts much?

Krishnakumar Natarajan: We start off with the concept. It was an outcome of our thesis that the next stage of AI application would be in computer imaging. Often, we also build such ideas with entrepreneurs. The application isn’t just limited to software and enterprise applications.

One of our recent investments is in a company in the earth observability market. It’s a data business as it involves taking images from satellites and selling them. That business has always been there, but then it moved from just high imaging cameras to high imaging spectrometry cameras. However, there’s a constraint. If there’s a cloud cover or a bush or some foliage, you could never get accurate images.

This startup has developed a technology that launches dual payloads: one with high spectral imagery and another with radar-based technology. Both send out signals, and the images are fused together by software, with AI applied to improve the accuracy of the image. This data becomes valuable for customers.

A use case for this technology is in India’s seafood export industry. India has a vast coastline of 7,500 Km, and seafood exports are significant. Importers want traceability to ensure proper processes and quality. These high-accuracy images, taken from space, help establish the quality of the products being shipped, accelerating trade.

There are many interesting use cases evolving, not just in the software arena but also in the tight integration of hardware and software. In areas like space tech, new use cases will emerge through the application of AI.

Sramana Mitra: So, KK, last question. How early are you and your colleagues in the venture capital industry in India going into AI startups? Let me qualify the question by saying that AI technology is a little bit more expensive to build, right? You need high-end engineers. You need data, you need compute. So, it’s not as easy to scrap together a lean startup and get an MVP out.

AI is a bit more expensive. Are you guys going pre-seed and concept stage and all that? What’s happening in your assessment in the industry with AI startups?

Krishnakumar Natarajan: See, fortunately, Sramana, I think there is a lot of focus from the government. What the Indian government is realizing is that unlike US or China, they’ve been little late in the race.

So, there is a lot of infrastructure which is being made available, particularly at the pre-seed level for people to experiment and at least prove their initial concepts. So, funds like us get into what we call early stage where there is some level of proof of concept and probably one or two customers. That’s the stage at which we get in and really help them scale the business.

But you’re right. I don’t think the infrastructure is developed to such a level. If you honestly ask me if is this a VC play, in our fund, we are a little more patient capital, we are not looking at exits in three to four years. We probably would take eight to years to exit many of these companies.

So, what is required is a very dependable core infrastructure, which today is still work in progress, but at least at the basic level, some of the essential infrastructure’s available.

The second thing is risk capital ecosystems, which can be innovative and bring in patient capital to wait for a much longer time. That is still not very well developed in India.

Sramana Mitra: Yes, in India and in a lot of the United States ecosystems as well. The United States does have high risk capital, infrastructure in particular, that is usually extended to serial entrepreneurs who have experience. First-time entrepreneurs don’t get access to that kind of capital. At least now there are repeat entrepreneurs in India as well. So, there is the luxury of betting on people who have some more experience, but I think there is a little gap in the AI financing ecosystem, whereby it’s not so easy to get something started because there are more costs involved than doing a lean startup.

In the 2005-2016 period lean startups boomed. We still believe very much in bootstrapped startups, but this is why one of the tracks we use in 1M by1M is Bootstrapping using Services, which is much more relevant in the context of AI because you need some way to get things moving and you need cap for that.

Krishnakumar Natarajan: The other thing is the traditional VC model of a fund life of eight years does not work in the case of AI or deep tech investments. They certainly require at least a 10-12 year fund cycle, which currently many of the traditional VCs wouldn’t have in the structure of their funds.

Sramana Mitra: VCs cannot change the structure of their funds to have a much longer lifespan just because limited partners won’t allow that.

The fallout of that is that the early-stage VCs exit into the late stage VCs. Instead of trying to stay 8-10 years, or 12-14 years, they stay up to a point and then they exit into the series C VCs or Series D VCs if they’re going for a large play. Otherwise they exit the company.  

Well, KK, we’d a very interesting conversation. We will continue in the future when I have you back. Thank you for coming.

This segment is part 4 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: With Krishnakumar Natarajan, Co-Founder of Mela Ventures
1 2 3 4