When it comes to pioneering AI ventures, few have the portfolio of Dr. Venkat Srinivasan.
As the founder of Innospark Ventures, an evergreen venture fund backing early-stage, AI-led startups, Srinivasan brings his extensive expertise in AI, computational algorithms and natural language understanding to the table.
His career spans multiple successful ventures, including eCredit, Corporate Fundamentals, and Rage Frameworks. More recently, he’s launched projects with a purpose-driven twist: EnglishHelper, aiming to tackle global illiteracy through AI; Gyan.AI, an explainable large language model; Creda Health, a digital health assistant; and PrismX, an automated software development platform.
With a bachelor’s in Commerce from Delhi University and a chartered accountancy degree from The Institute of Chartered Accountants in India, Srinivasan laid the groundwork for his future in finance. Furthering his studies with a doctoral degree from the University of Cincinnati’s Carl H. Lindner College of Business, he soon found himself drawn to the world of AI and intelligent systems, combining technical knowledge with a business acumen that would shape his path as a serial entrepreneur.
Recognized with a Lifetime Achievement award by The Indus Entrepreneurs (TiE) in 2017, Srinivasan is a former associate professor at Northeastern University’s College of Business Administration and has nine patents and over 30 peer-reviewed research papers to his name.
In an exclusive interview with The American Bazaar, Srinivasan discusses investment philosophies, AI boom, current venture capital market and more.
READ: GreenBrilliance CEO Sumit Bhatnagar: Making renewable solar energy accessible worldwide (December 19, 2024)
American Bazaar: Recently, you made a big splash, announcing two investments, in the Welsh startup Antiverse and the Canadian Phaseshift.
Dr. Venkat Srinivasan: We typically do about seven or eight investments a year, and we’re constantly seeing a very full pipeline. We’re very excited about Antiverse and Phaseshift because they fit our thesis very well. We believe that AI will transform every industry, and we’re particularly interested in how it can be used for good. Both of these investments — one related to accelerating the discovery of antibodies and the other focused on discovering new materials using AI — are exciting because they align with that mission.
Discovering new materials is somewhat like discovering new drugs — similar in approach, just not as regulated.
We just made another investment last week that’s somewhat similar to Phaseshift. It’s not public yet, but you’ll see it when it comes out.
It’s interesting that, of the two startups, one is based in Cardiff, UK, and the other is in Canada. Do you look for startups outside of the US?
That’s an interesting point. When we started Innospark, my partner and I decided we weren’t really building a career; we were doing this because we believe that AI is here to stay and that it will transform every field—personal, professional, business, and society as a whole. Given our deep knowledge of AI and the several decades we’ve spent on it, we felt we could actually make an impact by influencing this in a small way, encouraging startups to use AI for good and for positive impact.
Initially, we weren’t interested in going all over the world. Our first goal was to see how many interesting startups emerged within the Boston ecosystem. We didn’t want to travel everywhere; we have done too much work travel in our lives!
Our focus is on very early-stage, IP-driven startups with substantive ideas, not just different business models. For example, we probably wouldn’t have invested in something like WhatsApp or Facebook unless there was deep scientific IP behind it.
Over the last five or six years, Innospark has established its brand, and many entrepreneurs see us as quite knowledgeable about AI. When you’re a startup, you’re not just looking for money; you’re also looking for VCs who are a good fit and understand what you do. That’s what happened with Antiverse and Phaseshift — the teams approached us, not the other way around. Now we receive many inbound inquiries. If the opportunity is interesting, we might engage, even if we’re not physically close by.
READ: How Jagadish Nadig and Wastelink are turning food waste into opportunity (December 2, 2024)
Antiverse, for example, has now relocated to Boston, which makes it easier for us to support them. Toronto is still okay because it’s not that far, but we don’t have an office there. We often suggest that companies consider relocating to Boston because of the strong ecosystem here, which can offer them more resources and support.
Innospark was launched in 2018 after your last exit from Rage Framework. How many companies are in the portfolio right now?
We have about 30 companies in the portfolio, give or take. There were four that we incubated before formally starting Innospark. Now we have a team of about seven or eight people and a well-defined process. When we started, the main tasks were to figure out our sourcing strategy, establish our investment process, and recruit a team.
What types of companies are there in the portfolio? In other words, what’s the portfolio composition like?
There are essentially three types of companies. The first ones are solving a very specific problem — I would call them vertical point solutions — a third of our portfolio, I would say, is in those types of companies. Generally speaking, these companies will have a trade exit, meaning they will get bought by somebody because they are part of a bigger solution. Some of them, if they do really well, might start to add adjacent capabilities and then become a bigger solution themselves, but that’s not the initial reason to invest in them. The reason to invest is that they are solving a really hard, interesting problem that has a significant impact on society, and we felt compelled by the team.
The second third of the portfolio is what we would call platform companies. These companies can have multiple products within a specific field. For example, we have a company called Olaris, a metabolomics company that creates precision diagnostics to help physicians determine which medications will work best for each individual patient. In certain diseases, this can be a life-or-death scenario. Olaris can produce a precision diagnostic for any metabolic disease, so it will have multiple products. These companies generally carry more risk because you don’t know if their first product will be successful, so they have to experiment. In contrast, a point solution is only solving one problem, so it’s easier to evaluate. We have many examples of such companies – Javelin, Neurable, Phaseshift, to name a few. Phaseshift is a platform company that can discover new materials — the potential is vast. Generally, we expect these companies to be larger in terms of size and value, though they carry more risk.
READ: From corporate narratives to children’s stories: Monica Saigal on crafting impactful tales (November 23, 2024)
The third category includes companies that are sector-agnostic; they are fundamental technologies. About a third of our portfolio is in this category, where the greatest risk lies. You often hear the phrase, “a technology in search of a problem,” and many of these companies fit that description. They are fundamental technologies, like Gyan, or another company called Prism, which is also in this category. These companies are utility technologies, so they carry the highest risk and have the longest timelines; generally it takes time for these companies to find the ’killer’ use case.
That’s our portfolio composition today. As we move forward, we anticipate more opportunities in platform companies, though we are cautious with horizontal technologies, selecting only a few and acknowledging the high risk involved.
Innospark Ventures is an evergreen fund. Our goal is to build a legacy that will continue beyond our tenure. we don’t plan to increase our size in terms of capital. As we exit investments, the capital will simply be reinvested, allowing it to continue indefinitely. Over time, we may refine our focus as new opportunities arise, especially with AI startups evolving rapidly over the next 5-10 years.
How do you source companies and figure out what’s a good investment? What’s the process like?
Our thesis was clear from the start — AI had to be central to a startup’s business for it to interest us. Many startups don’t fit our criteria because AI is an afterthought. We want AI to be at the heart of the business, so we focused on finding startups that fit this model. We leaned heavily towards innovation, especially from universities in the Boston area like MIT, Harvard, Northeastern, Boston University and also broader on the east coast like Dartmouth or Princeton. We knew many people in the ecosystem, but we had to establish these connections in the context of Innospark.
We are very comfortable taking lots of risks. If a concept had breakthrough potential, it is more interesting to us. We weren’t looking for safe investments because, frankly, there is no such thing. Also, we wanted to invest very early, and we’re comfortable with that. Thirty years ago, when I became an entrepreneur, there was very little early-stage venture capital on the East Coast, unlike the West Coast. Nowadays, there are several early-stage VC options, and we’re one of them.
But isn’t there more risk in looking primarily for breakthrough companies?
Absolutely. Conventional wisdom suggests that early-stage venture firms lose up to 50% of their portfolio. We approached this knowing that and formed our own strategy based on experience and intuition. Our approach is to find companies with defensible technology — whether it’s proprietary IP or something they have that others don’t. Our thesis is that even if a company doesn’t go all the way to a successful exit, the defensible technology might still have value, creating non-zero “salvage value”.
For instance, we invested in a company called Encora, which created a non-invasive wristband device using AI to help Parkinson’s patients control tremors. It was a breakthrough idea with huge impact potential. If we could create a non-invasive device that could give Parkinson’s and essential tremor patients control of their lives again, the impact would be phenomenal.
You’ve been involved in ed-tech for a long time. Can you talk about your education-related companies?
Yes, while AI will impact every sector, education is one that will see a huge transformation. Education is all about knowledge, and so is AI. I founded English Helper about 10 years ago, which was an impact-first enterprise aiming to help people think, read, and comprehend in English. Unlike language learning technologies like Duolingo, our goal was deeper — enabling people to understand work or study related complex material written in English . English Helper now serves millions of students globally.
In terms of higher education, we’ve also launched a rapid content course creator leveraging Gyan, our explainable large language model I mentioned earlier. Long-duration degrees are increasingly being challenged by employers who now prioritize employable skills over academic degrees. The course generator can help universities quickly generate curriculum that meets current industry needs.
READ: Cross-continental love: Rediscovering love in the valley of Kashmir with Monica Saigal’s latest novel (March 17, 2024)
Finally, what’s your assessment of the current venture capital market?
I would say these are good times for VCs, despite the challenges in raising later-stage capital in a depressed market. For strong companies, there’s never really a bad time to raise capital. You may need to knock on more doors, but solid companies are generally more immune to market sentiment.
Right now, there’s a lot of hype around AI, and while there is money in climate tech too, the broader market is slower.
Talking of hype, there’s been a lot of irrational exuberance in AI recently, with $70 billion pumped into the field last quarter. What’s your take?
The hype is real, and it’s reminiscent of the dot-com bubble in the late ’90s. However, the scale of this AI hype cycle dwarfs past ones. I believe there’s an architectural flaw in current black-box AI models, which some investors are pouring billions into. Eventually, I think we’ll see a shift toward more rational approaches. It may take a few years, but we’re already seeing companies adapt and pivot. Once the irrational exuberance subsides, I believe the real advancements in explainable, sustainable AI will come through.
You started Innospark in 2018 as a $100 million fund? Now you have doubled that, right?
Yes, you could say that. We have two funds now; the first one has completed its investment phase, and the second one is ongoing at the same size. I expect that pattern to continue.
A common question many investors get asked is how closely they work with founders. Are you directly involved in the companies?
We are not directly running any of the portfolio companies, of course. Each of them has excellent teams running them. But we work closely with the founding teams. Innospark is a very early-stage investor, so our involvement can feel almost like founding. We have already created a lot of value for our portfolio companies with our close involvement. This is why we have a steady inflow of high-quality opportunities—because of the interaction we have with founders. We’re known for being founder-friendly. Having been operators ourselves, we understand the challenges founders face. Unlike many VCs, we’ve been in the trenches with AI and have actually operated AI companies, so we relate closely to founders’ challenges.
We tell founders we’re there [for them] as much as they need us. Some founders, would come to our office every week to learn and seek guidance, and we welcomed that. That company, by the way, is doing very well. Other companies have experienced CEOs who only reach out when necessary — at board meetings, for instance. We’re available, but we’re not intrusive. We’re there as a sounding board, to help guide, and to tell them if something’s not working, but we respect that it’s the CEO who ultimately calls the shots.
You are among the first Indian Americans to work in AI. Can you share your journey in AI?
My interest has always been in problem-solving. During my doctoral work, I didn’t take the conventional path most finance PhDs did. My background in computing led me to explore AI. I focused on building intelligent systems — like one that could help companies set credit limits based on consistent knowledge. The system could generate a written report, similar to what ChatGPT does today.
Over the years, I’ve delved deeper into various forms of AI and believe stongly in explainable and tractable AI. Intelligence acquisition for machines need not be only based on data; it can also rely on human experts. I believe that anything important needs to be explainable,. Ultimately, I think explainable AI will be what endures.