Soket AI Raises $4 Mn From Bajaj Finserv To Expand Its AI Model Horizon

Startups & Funding

Soket AI Raises $4 Mn From Bajaj Finserv for India-Built AI Models

Soket AI raised $4 Mn from Bajaj Finserv's Finserv Intelligence and Pennant Technologies to scale EKA-1, its 26-billion-parameter sovereign foundation model for India.

By
Nathan Brooks
Filed
Channel
Startups & Funding
Read
2 min read

AI startup Soket AI has raised $4 Mn (around ₹37 Cr) from Bajaj Finserv's applied research and innovation initiative Finserv Intelligence, and from Pennant Technologies. Bajaj Finance invested approximately $2 Mn, while Pennant Technologies, itself an investee company of Bajaj Finance, put in the other $2 Mn.

The startup builds indigenous foundation models focused on mathematics, coding and reasoning. Backed by the government's IndiaAI Mission, Soket AI recently launched EKA-1, a 26-billion-parameter foundation model under Project EKA, its flagship sovereign AI initiative.

What will the money fund?

The $4 Mn will support applied AI research and enterprise deployments, particularly across financial services, cybersecurity, defence and enterprise technology. The company also intends to scale its foundation model research and develop AI systems for those same sectors.

Beyond the cash, the deal carries a strategic layer. The investment is expected to open opportunities for applied AI research, enterprise applications and capability building by combining Bajaj Finance's industry expertise with Soket AI's foundational research capabilities. In effect, a lender known for consumer credit is buying a seat next to frontier research.

Bajaj Finance Vice Chairman and MD Rajeev Jain said: "With Soket AI, we look forward to enabling India's indigenous AI ecosystem, engaging with frontier research at an early stage, and bringing our scale and domain expertise to turn this research into practical solutions."

Who is Soket AI?

Founded in 2019 by Abhishek Upperwal, the startup is building a sovereign open-source AI stack that serves a range of Indian languages. Speaking to Inc42 earlier this year, the founder said the idea is to crack the AI model for India first.

Soket belongs to a limited set of Indian AI startups developing indigenous foundation models with support from the IndiaAI Mission. That is a short list by any measure, and state backing narrows it further.

How crowded is the field?

Two other moves this month frame the competitive picture:

  • ZenteiQ AiTech Innovations launched Dhīr 2B and Dhīr 5B, two open-source multilingual models supporting English and 10 Indian languages. Unlike Soket's reasoning-focused EKA-1, these are lightweight assistants that run on-premises and on edge devices, including smartphones, without continuous cloud connectivity.
  • AI unicorn Sarvam AI expanded its portfolio with Prism, an open-weight decision model built on its Sarvam 105B model. Prism targets enterprise workflows such as customer support routing, security checks and evaluating AI agent activity.

The three companies occupy different niches: Soket chases reasoning, ZenteiQ chases the edge, Sarvam chases enterprise decisioning. Whether one stack can serve all of them remains an open question in the Indian market.

For Bajaj Finance, the wager is that financial services will be the first place sovereign models earn their keep. The round is modest by foundation-model standards, yet it pairs a bank's distribution with a research team's models before either has proven the combination works.

Source: inc42.com

Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

Senior reporter covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,070 articles

Related articles

« Previous articleNext article »