With Google and Microsoft-backed OpenAI fighting it out for AI supremacy, early-stage backers in this space globally find themselves to be a hot sector.
ET spoke with multiple venture investors, and funds are taking different approaches towards AI.
Some funds are looking for potential bets on startups with proprietary training models, while others are evaluating potential investments in the applied-AI space where companies build on top of large language models like the generative pre-trained transformer (GPT).
Investors said they were looking at potential investments in companies that are building application layers on top of existing areas such as edtech, healthtech, job-tech, etc.
Indian startups offering AI or machine learning-based products have so far raised $270 million this calendar year, according to data provided by Tracxn – a platform for data on privately held startups.
Discover the stories of your interest
Will Poole, cofounder and managing partner of Unitus Ventures, which has backed companies like Masai School, Cuemath, GigForce, and Awign, told ET that he is on a three-week trip to India to interact with eight of the early-stage fund’s portfolio companies about incorporating generative AI tech into their business and product strategies.He believes the impact of generative AI on businesses would be as big as that of the internet in the 1990s.
“I was raising money for an ecommerce company in the mid-1990s. In 1993, the Mosaic (internet browser) was announced, and by 1994, if you did not have a strong internet strategy and a strong prototype of what you’re doing, you could not raise a penny. Same phenomenon is happening here,” Poole said. “There’s a new technology that’s changing how users can interact with computing just like the browser did in 1993. In our case, there are several portfolio companies that are already building (products) using GPT-4, and some will be shipping products by the end of this month,” he added.
Early-stage venture capital fund pi Ventures said it has raised Rs 22 crore ($2.7 million) from Colruyt Group India last month. The fundraise has helped the fund focused on AI and deep-tech startups nearly hit its final close in the range of Rs 675 crore to Rs 750 crore.
Speciale Invest, another deep tech-focused investment firm, last week launched its Speciale Invest Growth Fund I to make follow-on investments in “winners” from its early-stage portfolio companies. “In all this upheaval of valuations going up and down, nothing has changed much in deep tech. We are still making the kinds of things we were doing before and with a similar sort of pace,” Manish Singhal, founding partner, pi Ventures, told ET.
The crossroads
Investors said companies operating in the AI space could broadly be categorised into two buckets – those building and training their own models, and those building application layers on top of existing models.
“Internally, this is what I tell my team… There are companies in India, which will be funded, which are either building AI or they are using AI. That’s the easiest way to think about it,” Prayank Swaroop, partner at Accel, told ET.
Unitus’ Poole said, “There will be investors that are specialists in the ‘pick and shovel’ part, but that’s not our strategy because it’s going to produce a lot of zeroes as well as big winners.”
Venture investors refer to the strategy of building tools to assist the process of software development by making it more efficient as ‘pick and shovel’.
Poole said it’s a highly competitive space where investment has been going on for a long time in the Bay Area and other tech centres of the world, and even tech giants investing in the core large language models (LLM) could build pick and shovel AI.
“The area we’re particularly interested in is applied AI,” Poole said.
Many investors see greater value creation in companies that build their own proprietary models and datasets.
“We are particularly looking closely at applications that are built on their own proprietary models and datasets, which we believe will be a source of competitive advantage,” said Ishavasyam Dash, director at early-stage venture capital firm Antler in India. “It’s crucial for founders and investors to ask hard questions on overcrowded categories, GPT-wrappers and whether an area warrants a completely new AI-native product, or whether an incumbent is better off adding it as a feature,” he added.
While meaningful products built using AI might take time to come to India, there would be companies building ‘me too’ products, investors said.
“Anything new takes time to trickle down from the US to the Indian market,”
Said Rajiv Mehta, general partner at early-stage firm Athera Venture Partners that has backed startups such as HealthifyMe, PolicyBazaar, and Euler. “The largest market will continue to be the US. What we see in India is that it will be a little while before meaningful companies start getting built in the generative AI space. There will be ‘me too’ companies in India as well but I don’t think they will be meaningful. Right now, things are in an early interest phase,” he told ET.
Sector intelligence
Mehta said Athera Venture Partners sees two primary problem statements for the Indian market – in edtech in rural parts of the country and in healthcare for areas with higher spending power.
“There are two kinds of problems that we see AI solving, which entrepreneurs should look at,” he said. “Generative AI and machine learning can potentially become a teaching tool for people who have access to the internet but don’t have access to real-life schooling.”
Mehta also expects use of AI in healthcare in markets where spending power is comparatively higher. “Both predictive and preventive analysis in healthcare using millions of data points already available should be something that finds some interest,” he said.
Unitus Ventures is looking at companies deploying generative AI tools to transform product front-ends to conversation-based user interfaces.
“Generative AI is simply changing the interface on a mobile app and website from one of point and click to one of conversation,” Poole said. “The front-end of all kinds of products and services ranging from ecommerce and travel to financial services can move to conversations,” he said.
In the job-tech space, helping potential employees become more productive and interfacing with them in a way that’s more efficient for an employer are among areas “that can be directly supported by generative AI”, Poole said.
For all the latest Technology News Click Here
