IndiaAI Mission has released three indigenous foundation models – EKA-1, Dhīr 2B and Dhīr 5B. The launch is important because these models have been developed in India with a focus on Indian languages, local requirements and secure deployment.

The models are not consumer apps in the same way as a chatbot website. They are the technology layer on which developers, government departments, schools, banks and businesses can build their own AI tools.

EKA-1 was developed by Soket Labs Technology and Research, while Dhīr 2B and Dhīr 5B were developed by ZenteiQ AiTech Innovations. The projects received support under the IndiaAI Innovation Centre pillar of the IndiaAI Mission reported PIB.

What is a foundation model

A foundation model is an AI system trained on a very large collection of text, code and other information patterns. It learns how language works, how questions are usually answered and how ideas relate to each other.

Developers can then adapt the model for specific tasks. For example, a bank may use a model to help employees find answers in policy documents, while a school may use it to create study support in an Indian language.

A foundation model does not simply store a ready-made answer for every question. It predicts useful words, numbers or code based on patterns learned during training. This is why AI can write, summarize, translate and answer questions. It is also why important outputs still need human review.

EKA-1 – the larger multilingual model

EKA-1 is Soket Labs’ open, sparse hybrid language model. It has 26.3 billion parameters and was trained on about 7.5 trillion tokens. Parameters are the internal settings an AI model uses to recognize patterns, while tokens are small units of text that the model processes during training.

According to the IndiaAI announcement, EKA-1 supports all 22 scheduled Indian languages along with English. It is designed for mathematics, coding, reasoning and agent-based tasks, where AI can complete steps within defined instructions.

Soket Labs has positioned EKA-1 for areas such as banking, financial services and cybersecurity. A practical example could be an internal banking assistant that helps staff search policy documents in Hindi, Tamil or English without moving sensitive information to a public AI service.

Dhīr 2B and Dhīr 5B – smaller models for secure use

ZenteiQ’s Dhīr 2B and Dhīr 5B are open-source multilingual models. The “2B” and “5B” refer to their approximate model size in billions of parameters.

They support English and 10 Indian languages and are designed for on-premise and offline-capable use. On-premise means the model can run inside an organization’s own computer systems instead of sending sensitive documents to an outside cloud service as per Newsonair.

This can be useful in sectors such as defence, government, education and industry. For example, a district office could use a Dhīr-based tool to search circulars and answer routine citizen queries even where internet connectivity is weak. An organization handling confidential documents could keep data within its own network.

Smaller models may not match a large cloud-based model on every complex task. Their advantage is that they can be more affordable to run, faster for focused tasks and easier to use in controlled environments.

Why these Indian AI models matter

The biggest benefit is language relevance. Many global AI systems work best in English and may not understand Indian languages, local terms or administrative context equally well. Models trained with Indian-language support can make digital services more useful for a wider population.

The launch also supports data control. Organizations in sensitive sectors may prefer AI systems that can run locally, especially when their documents include financial, personal or government information.

Another benefit is access for researchers and startups. IndiaAI has said that the models can be made available through AIKosh, helping students, innovators and developers experiment with Indian AI tools.

How they compare with global AI models

EKA-1 and Dhīr will compete indirectly with international models from companies such as OpenAI, Google, Meta and Anthropic. But the goal is not only to build the largest model.

Their stronger focus is on Indian languages, local deployment and sector-specific needs. In many cases, a smaller model that works reliably with a company’s own documents may be more practical than a much larger public AI model.

India is preparing a new step in its artificial intelligence policy journey. Union Electronics and IT Minister Ashwini Vaishnaw has said that the government will release a consultation paper on AI regulation within a month.

The document is expected to invite views from technology companies, startups, researchers, legal experts, civil society groups and citizens. It will not immediately become a law. Instead, it will help the government understand the risks and decide what kind of rules India may need.

What will the consultation paper examine

The proposed paper is expected to focus on AI safety, cybersecurity, deepfakes and a human-first approach. It may also examine how companies should explain the use of AI, protect personal information and respond when an automated system causes harm as reported in News On AIR.

For example, if an AI tool creates a fake video of a public figure, it should be possible to identify the content as synthetic and trace responsibility for its distribution. Similarly, if an AI system is used in lending, recruitment or healthcare, people may need safeguards against unfair or incorrect decisions.

Why India needs clearer AI rules

AI is already being used in banking, agriculture, education, healthcare, customer service and government services. It can reduce costs, improve productivity and help people access services more quickly.

At the same time, AI can create false information, copy personal data, generate harmful content and make decisions based on biased information. Deepfakes can damage reputations, influence public opinion and create confusion during elections or emergencies.

Clear rules can help define who is responsible when an AI system fails. They can also encourage companies to test their products, maintain records and offer a way for users to complain or seek correction.

Balancing safety and innovation

India will need to strike a careful balance. Very strict rules may increase costs for small startups and slow down useful experiments. Very weak rules may expose citizens to fraud, privacy violations and unsafe automated decisions.

A practical framework could use different rules for different levels of risk. A simple AI writing tool may need basic transparency, while an AI system used in medical diagnosis, insurance or public welfare may require stronger testing and human supervision.

The government has already indicated that AI should remain accessible to startups, researchers, students and small businesses. The consultation process can help create rules that protect users without making innovation impossible as per the Economic Times.

What it could mean for companies and users

Startups may receive clearer guidance on data use, model testing, content labels and customer consent. Larger technology companies may face stronger expectations around safety reporting and accountability.

For users, good regulation could mean clearer notices when they are interacting with AI, better protection from fake content and stronger remedies when automated decisions cause harm.

India’s AI Governance and Economic Group, created by MeitY in 2026, is expected to support policy coordination. The government had also released broader AI governance guidelines in 2025, making the upcoming paper part of an evolving policy process reported PIB.

Conclusion with key takeaways

The consultation paper is important because AI is spreading faster than formal rules. India needs a framework that supports innovation while protecting privacy, safety, fairness and public trust.

The next stage will depend on the quality of the consultation. Industry participation, public feedback and clear explanations will be essential before any final AI regulation is introduced.

Image

Disclaimer

The consultation paper has not yet been released. Its final scope, recommendations and legal impact may change after stakeholder feedback.

Anthropic has begun offering in-country inference for its Claude AI models in India through Amazon Web Services’ Amazon Bedrock platform.

In simple terms, requests made by Indian organizations can now be processed within AWS infrastructure in India instead of being routed to overseas locations. The move could make Claude more suitable for banks, hospitals, government departments and large companies that handle sensitive information.

What has Anthropic launched?

Amazon Bedrock now provides access to Anthropic’s Claude Opus 5, Claude Sonnet 5 and Claude Haiku 4.5 through an India geographic cross-Region inference profile.

The system uses AWS regions in Mumbai and Hyderabad. Requests can move between these two Indian regions, allowing AWS to balance workloads and maintain performance during periods of high demand. AWS says the traffic remains within India and moves through its secure network with encryption while in transit as per AWS announcement.

This is different from simply allowing Indian users to access Claude through the internet. The important change is where the AI model processes the prompt and generates its answer.

However, local inference does not mean Claude runs offline on a user’s laptop or phone. It remains a cloud-based service managed through Amazon Bedrock.

Why local inference is important

Many Indian organizations work with information that cannot easily be sent outside the country. This may include bank records, medical information, government documents, customer identities and internal business data.

For example, a bank could use Claude to summarize loan documents or assist its customer-support team. A hospital may use an AI system to organize medical records. A government department could use it to search large document collections.

Local processing may make it easier for these organizations to meet internal security policies, contractual requirements and data-handling rules. It can also reduce the distance data travels, which may help lower response times for some applications.

Anthropic India managing director Irina Ghose said Indian customers, particularly those in regulated sectors, had requested local processing for critical workloads. India is also described as one of Claude’s most important markets, with software development accounting for a large share of work-related usage.

Benefits for Indian businesses

The main benefit is stronger data control. Companies can design their AI applications with a clearer understanding of where requests are processed.

The India-based cross-Region setup may also improve reliability. If one AWS region faces a capacity problem, traffic can be handled by the other Indian region. This is useful for customer-service platforms, coding tools and business applications that need steady performance.

Amazon Bedrock also gives companies access to security controls, identity management and monitoring tools available within the AWS ecosystem. Developers can use Anthropic’s Messages API, the Bedrock InvokeModel API or the Converse API to build applications per AWS technical guidance.

Impact on India’s AI ecosystem

The move could encourage more Indian companies to shift from small AI experiments to production systems. Local banks, IT companies, healthcare firms, universities and public agencies may find it easier to evaluate Claude for real business tasks.

Anthropic will face competition from Google’s Gemini models on Vertex AI, OpenAI models available through cloud platforms, Microsoft Azure AI services and Indian language-model companies.

The key question will be pricing, performance and compliance. Local processing may help with data residency, but companies will still need strong access controls, careful data policies and human oversight. AI responses can also contain mistakes, so sensitive decisions should not be fully automated without review.

Conclusion with key takeaways

Anthropic’s India launch through AWS Bedrock is mainly an enterprise-focused development. It gives Indian organizations access to Claude models while keeping inference inside AWS regions in Mumbai and Hyderabad.

Key takeaways-

• Claude Opus 5, Sonnet 5 and Haiku 4.5 are available through AWS Bedrock in India.
• Inference requests can be processed between AWS regions in Mumbai and Hyderabad.
• The move may help regulated sectors manage data-residency concerns.
• Local inference can support lower latency, better reliability and easier compliance planning.
• The service remains cloud-based and is not an offline version of Claude.
• Indian businesses will still need security controls, testing and human supervision.

Ref. Claude on Amazon Bedrock

India’s artificial intelligence sector is seeing one of its biggest funding moments yet. As of May 15, 2026, multiple reports indicate that HCLTech is likely to invest $150 million in Sarvam AI and lead a broader $300 million funding round that could value the startup at around $1.5 billion. If the deal closes on these terms, it would mark a major milestone for India’s homegrown AI ecosystem and highlight how traditional IT giants are taking direct stakes in next-generation AI companies.

Why the Reported HCLTech-Sarvam Deal Is Important

The reported structure matters because this is not just a financial transaction. HCLTech’s participation appears strategic as it can gain early access to locally relevant AI capabilities while strengthening its enterprise AI offerings. Sarvam AI, on the other hand, could gain scale capital, market trust, and faster enterprise distribution through HCLTech’s global client network.

This is also significant for India AI startup funding more broadly. For years, large AI model development required deep capital and long timelines, which many startups struggled to secure. A large round led by an established IT company suggests a shift from cautious pilot-stage investments to bigger, long-term bets on AI infrastructure and foundational model development.

What This Could Mean for India’s AI Roadmap

The deal aligns with a wider push toward sovereign AI India initiatives, where local language capability, regional compliance, and domestic compute strategy are becoming more important. Sarvam AI’s positioning in multilingual and foundational AI development fits that direction, especially as enterprises and public institutions demand AI tools built for Indian contexts.

For HCLTech, this move can support its AI strategy by combining services expertise with deeper product and model-level capability. For Sarvam, it can accelerate execution in sectors such as financial services, telecom, healthcare, and digital public platforms.

The broader signal is clear which is that India’s AI market is maturing, and partnerships between large IT firms and specialized startups may become a defining growth model.

In short, while the investment is still being reported rather than officially closed in public filings, the proposed numbers and structure already show where the market is heading i.e. larger rounds, strategic investors, and faster enterprise AI adoption.

Facts Input