Bajaj Finserv and Pennant Technologies have jointly invested $4 million in Bengaluru-based artificial intelligence startup Soket AI, strengthening India’s push to develop indigenous foundation models and enterprise AI capabilities. The investment comprises $2 million from Bajaj Finance, through Bajaj Finserv Intelligence, and another $2 million from Pennant Technologies, a fintech company serving banks and financial institutions.

The investment will support Soket AI’s development of indigenous large language models focused on mathematics, coding and reasoning, alongside AI agents designed for enterprise applications. The partnership is expected to focus particularly on financial services, where institutions require AI systems that can handle sensitive data, automate complex workflows and operate within tightly controlled environments.

Bajaj Finserv and Pennant Invest $4 Million in Soket AI

The $4 million investment represents a strategic bet on India’s domestic AI research capabilities and the potential to deploy locally developed AI models across regulated industries.

Bajaj Finance, through its applied research initiative Bajaj Finserv Intelligence, has invested approximately $2 million in Soket AI. Pennant Technologies, an investee company of Bajaj Finance, has separately committed another $2 million.

The funding will enable Soket AI to expand its research into foundation models, AI agents, inference systems and enterprise applications. The companies also plan to collaborate on evaluating and deploying Soket AI’s technologies within Pennant’s banking and financial technology platforms.

The investment comes as financial services companies increasingly look to integrate generative AI and agentic AI into business operations, including lending, customer onboarding, document processing, fraud prevention and collections.

For Bajaj Finserv, backing a homegrown AI research company could provide access to emerging technologies that may eventually be integrated into its own businesses and financial services ecosystem.

Soket AI Is Developing Indigenous Foundation Models

Soket AI is an Indian AI research company focused on building foundation models with capabilities in mathematics, coding, reasoning and multilingual applications.

The startup is supported by the Government of India’s IndiaAI Mission, which aims to strengthen domestic AI development, computing infrastructure, research capabilities and AI adoption across industries.

Its flagship initiative, Project EKA, is focused on developing sovereign foundation models that can support applications requiring greater control over technology, infrastructure and data.

The company’s research ambition extends beyond building conventional chatbots. Its models are being developed to support more advanced reasoning, enterprise workflows and agentic applications that can execute tasks with limited human intervention.

Soket AI is also working towards developing larger foundation models, with its roadmap spanning models from billions of parameters to significantly larger systems.

EKA-1: Soket AI’s 26-Billion-Parameter Model

Soket AI has launched EKA-1, a 26-billion-parameter foundation model under Project EKA.

The model focuses on mathematics, coding and reasoning, areas that are increasingly important for AI systems expected to perform complex tasks rather than simply generate conversational responses.

Project EKA is designed to support more than 100 languages, including Indian languages, languages used across the Global South and programming languages.

The initiative also emphasises deployment flexibility, including the ability to operate on a company’s own infrastructure and in isolated environments where data cannot be sent to external cloud systems.

Such capabilities could be particularly valuable for banks, financial institutions, government organisations and businesses operating in sectors where data security, regulatory compliance and operational control are critical.

Pennant Technologies Wants AI Agents for Banking Operations

Pennant Technologies provides technology platforms and AI-powered lending solutions to banks and financial institutions. Its investment in Soket AI is expected to strengthen its efforts to develop agentic AI capabilities for the banking industry.

The companies intend to collaborate on evaluating Soket AI’s foundation models, agentic harness and inference runtime, with applications across Pennant’s banking agents, document processing systems and collections intelligence platforms.

The partnership could also support the development of specialised banking AI models and agents for several functions.

Planned AI applicationPotential use
KYCCustomer identity verification and onboarding checks
Customer onboardingDocument processing and workflow automation
CollectionsCustomer engagement and collections intelligence
ComplianceRegulatory checks and operational monitoring
Fraud detectionIdentifying suspicious transactions and activity
Document processingExtracting, analysing and organising financial documents

The focus on specialised applications reflects the growing demand for AI systems capable of handling industry-specific tasks with greater reliability and operational control.

Why Sovereign AI Matters to Financial Services

Financial institutions handle sensitive information involving customer identities, financial transactions, credit histories, lending decisions and business operations.

Deploying AI models through external infrastructure can create concerns around data exposure, governance, security, regulatory compliance and vendor dependence.

Sovereign AI refers broadly to the ability of an organisation or country to develop, control and deploy AI systems using infrastructure, models and governance arrangements that meet its strategic and regulatory requirements.

For banks, locally developed AI models that can run on private infrastructure could help organisations retain control over customer data, system access, operational processes and AI-generated decisions.

Soket AI’s focus on on-premises and air-gapped deployments could therefore make its technology relevant to institutions that cannot freely use externally hosted AI services.

However, sovereign AI does not automatically guarantee security or accuracy. Institutions will still need to evaluate model performance, cybersecurity, reliability, auditability and governance before deploying AI in sensitive operations.

Bajaj Finserv Expands Its AI Investment Strategy

The Soket AI investment is part of Bajaj Finserv’s broader strategy to explore artificial intelligence and build capabilities that can be applied across financial services.

The group is increasingly focused on AI as a potential driver of operational efficiency, customer experience, risk assessment, digital services and product innovation.

Investing in AI startups gives large companies an opportunity to access research and technology that may be difficult or time-consuming to develop internally.

For Bajaj Finserv, the Soket AI partnership combines financial services expertise with foundational AI research. Pennant’s experience in banking technology could also provide an environment for testing and adapting AI models for real-world financial workflows.

The partnership’s success will depend on whether the technology can deliver measurable improvements in efficiency, accuracy, customer service and operational costs while meeting the strict requirements of financial institutions.

India’s Indigenous AI Ecosystem Gains Corporate Backing

The investment reflects a broader shift in India’s AI ecosystem, where large corporations, startups, research institutions and government programmes are increasingly collaborating to build domestic AI technologies.

India’s AI ambitions extend beyond using foreign-developed models to developing foundation models, computing infrastructure, research capabilities and specialised applications that can serve Indian businesses and international markets.

Indigenous AI development could create opportunities in multilingual AI, enterprise automation, financial technology, cybersecurity, healthcare, manufacturing and government services.

Funding from established financial services companies could help startups access commercial use cases, industry expertise, enterprise customers and long-term development opportunities.

The Bigger Picture

Bajaj Finserv and Pennant Technologies’ $4 million investment in Soket AI highlights how Indian financial services companies are becoming participants in the development of AI technology rather than remaining only consumers of external platforms.

The partnership is particularly significant because it connects foundation-model research with practical banking applications. Developing AI models is only one part of the challenge; deploying them reliably in regulated environments requires testing, infrastructure, governance, specialised workflows and enterprise integration.

If Soket AI can translate its research into commercially useful systems, the partnership could help strengthen India’s domestic AI ecosystem while giving financial institutions additional options for deploying AI under their own security and governance requirements.

Looking Ahead

Soket AI is expected to continue developing Project EKA, expanding its foundation-model capabilities and improving systems designed for reasoning, coding, mathematics and multilingual use. The collaboration with Bajaj Finserv and Pennant could accelerate the evaluation and deployment of its AI models and agents across financial services applications. The immediate focus will likely be on building practical solutions for banking workflows, including document processing, customer onboarding, collections, compliance and fraud detection.

The investment also reflects a growing interest in India’s sovereign AI capabilities, with financial institutions increasingly seeking models that can operate securely within their own infrastructure. As AI adoption expands across regulated industries, the ability to combine strong model performance with privacy, control, reliability and cost efficiency could become a competitive advantage. Soket AI’s progress in turning its research into scalable enterprise applications will determine how significant the investment becomes in India’s rapidly developing AI market.

Get the day’s top stories in your inbox

One concise email. No spam, unsubscribe anytime.