What’s Holding Back VCs From Betting Big On Indic LLMs?

Why hasn’t India made major strides in the deeptech space yet? The oft-quoted question recently gained steam when the Minister of Commerce and Industry, Piyush Goyal, pulled up Indian startup founders for wasting their talent, time and capital on delivering groceries or making fancy ice cream brands rather than doing real-world innovation in the realm of deeptech.

While the minister’s concerns were genuine, the reality is that India’s deeptech sector is plagued with capital constraints and a risk-averse funding ecosystem. According to the Inc42 data, the deeptech sector was the least favoured investor destination in the first three months of 2025. 

The sector, which houses robotics, frontier models in AI, semiconductor, EV infrastructure, next-gen healthcare, defence tech, aerospace, drones, etc., could only raise $84 Mn in 21 deals. 

Compared to ecommerce and consumer services (quick commerce is a major part of this sector) sectors, which have raised $35.8 Bn and $14 Bn, respectively, since 2014, the Indian deeptech startups have only been able to raise $2 Bn+ since 2014. This is not even half of what Zomato has raised since its inception. For those wondering, the foodtech major has raised more than $5 Bn to date.    

With such limited capital, India’s deeptech ambitions currently seem to be in a precarious state of affairs, especially when the creation of a sovereign large language model (LLM) forms the core of India’s GenAI vision.

This brings us to the most crucial part of the equation — What do investors want? Well, it is unfortunate, but only a handful of investors see Indic-language LLM as a good business proposition. 

According to Inc42’s latest “The Pulse Of Tech Inc42’s Investor Survey, Q1 2025”, only 13% of the 75+ Indian startup investors surveyed believe investing in startups working on building Indic-language LLMs is a good opportunity. 

Meanwhile, about 21% and 24% of investors surveyed by Inc42 are willing to invest in crucial building blocks of LLMs. These crucial blocks comprise data sets, data centres, GPUs, etc.  

Access Free Report

Why Are VCs Bearish On Indic-Language LLMs? 

Well, the answer is multifaceted. For starters, developing Indic-language LLMs is a capital-intensive prospect and takes years to come to fruition. Despite this, success is not certain.

From the lens of SaaS-focussed VC firm NuVentures’ founder Venk Krishnan, it is tough to make bets on a new-age tech, which requires an exceptional amount of capital and whose success, in turn, continues to remain variable. Plus, he is of the opinion that India lacks the prowess to create foundational models. 

“If you look at foundational model companies worldwide, they are doing very advanced research on the back of advancements in compute power. Replicating that in India would require a massive investment,” he said.

For context, compute power means the hardware and software needed for AI systems to process data, train models, and make predictions. It’s a key part of the AI infrastructure. Like Indic LLM startups, this area also hasn’t attracted much investor interest so far.

Another blazing reason why VCs think that Indic LLMs may not see a massive success is because of the ocean of languages and dialects that our country harbours.

With close to 400 languages, the country has vast amounts of data but lacks high-quality digital datasets, which are imperative to building strong LLMS. This poses a big challenge when it comes to building effective LLMs for India.

Access Free Report

Another reason why VCs hesitate to back startups working on long-term deeptech projects like Indic LLMs is because of their low appetite for risk capital. This makes sense, especially when they are bound to give exits to their LPs in a fixed time, which in the case of developing deeptech may extend indefinitely.

“Maybe the government or bigger industry players can back such innovations, but a VC will always refrain from putting somebody else’s money at risk. It is a moonshot is what I believe,” Krishnan said.

Despite the VC’s bearish outlook towards indigenous LLMs, hordes of VC dry powder has been set aside for AI innovations. For instance, in the first three months of 2025, five VC firms Accel, Bessemer, Eximius Ventures, Prime Venture Partners and 2am VC launched funds cumulatively worth $1.13 Bn to invest in AI innovations.

In sync with this, the Inc42 survey highlights that a majority of investors are bullish on India’s sovereign AI future. 

So, Where Are VCs Looking To Invest In AI?

According to the Inc42 report, 53% of the investors are looking to invest in startups building Agentic AI, while 47% are optimistic about investments in application layer AI.

Requesting anonymity, a partner at a VC firm said that many Indian VCs are actively looking for innovations in the space that can achieve product-market fit by building agents on top of existing AI foundational models.

“We’re particularly bullish on agentic AI because it moves beyond demos and into deployment. The startups that can seamlessly integrate these agents into enterprise workflows are the ones poised to define the next generation of AI applications,” the investor said. 

What’s India’s Stand On Its LLM?

It’s worth noting that the Indian government is actively working to build an Indic foundational model through a public-private partnership (PPP) approach. 

Recently, IT minister Ashwini Vaishnaw said that the government is in the final stages of evaluating applicants for building an AI LLM under the IndiaAI Mission. The mission, approved by the government with a budget of INR 10,372 Cr (about $1.2 Bn), aims to make investments over the next five years.

According to the minister, the evaluation of the AI LLM applications is in its final stage. The government should be able to award the first few LLM model teams, which will then start getting funding under the AI mission.

However, when compared to the investments made by other major economies in AI, the IndiaAI Mission’s scale appears quite small. 

For instance, the US government recently announced a private sector investment of up to $500 Bn to support infrastructure for Stargate — a joint venture between OpenAI, SoftBank and Oracle. Earlier this year, China also launched an $8.2 Bn National AI Industry Investment Fund to strengthen its AI ecosystem.

Recently, Info Edge founder Sanjeev Bikhchandani highlighted the importance of government and industry support in developing Indic LLMs. Citing the example of OpenAI, which has raised nearly $60 Bn to date, Bikhchandani noted that the company’s early backers were not VC funds, but well-funded individuals from within the industry.

All in all, India’s ambitions to build a sovereign AI future — anchored in Indic-language LLMs — face steep challenges. A risk-averse VC ecosystem, long gestation periods, and limited government funding have created a tough environment for deeptech startups to thrive.  

Now, as the global AI race intensifies and countries like the US and China ramp up efforts in sovereign AI, chips, and quantum tech, India needs to act now to realise its Indic LLM dream.

Access Free Report

[Edited By Shishir Parasher]