Bengaluru-based AI startup Gnani.ai has unveiled its 'Artha' sovereign AI stack, designed to help government agencies and businesses manage AI infrastructure internally. The move aligns with the government's IndiaAI Mission to secure data residency. Please note that Gnani.ai is a private company and is not listed on the Indian stock exchanges, though it is backed by investors including Info Edge Ventures.
Bengaluru-based AI startup Gnani.ai officially launched its 'Artha' sovereign AI stack on August 28, 2026, at Uprashtrapati Bhavan in New Delhi. The launch, attended by Vice President C. P. Radhakrishnan, highlights the growing focus on localizing artificial intelligence infrastructure within India. The company has been selected under the government’s IndiaAI Mission, which aims to foster an indigenous ecosystem for AI development.
Understanding Sovereign AI
For many large organizations, especially in banking, insurance, and government sectors, using public cloud-based AI tools presents a data security challenge. Sensitive information must often stay within the organization's own digital borders to comply with data residency regulations. 'Sovereign AI' refers to a system where a company or government agency keeps full control over its data, model, and infrastructure, rather than relying on external providers.
Gnani’s Artha stack aims to address this by allowing enterprises to run AI models on their own servers, known as 'on-premises' deployment. This approach provides the customization and privacy that public AI models, which are often hosted in external data centers, may not offer.
The Artha Stack Components
The newly launched suite consists of two primary offerings. The first is 'Ivon 3.3,' a large language model with 30 billion parameters that supports 11 Indian languages. It uses a 'Mixture-of-Experts' (MoE) architecture. To understand this in simple terms: instead of running the entire model for every request, the system activates only the specialized parts needed for that specific task. This method significantly reduces computing power requirements and operating costs, making it more efficient for businesses to run high-performance AI on smaller hardware.
The second component, 'Plexis,' is an agentic AI platform. This is designed to act as an autonomous agent that can carry out business workflows, such as processing documents or handling customer queries, with minimal human intervention.
Investment and Market Context
Investors looking for exposure to this development should note that Gnani.ai is a private venture-backed startup and not a publicly traded entity on the NSE or BSE. As a result, there is no direct stock price to track for Gnani.ai itself. However, the company is backed by firms like Info Edge Ventures and Aavishkaar Capital. Info Edge (India) Limited, a publicly listed company, maintains a diverse portfolio of tech investments, which is a key area for those monitoring the startup ecosystem.
The commercial success of Artha will depend on how quickly it can win contracts within regulated industries where data security is the top priority. The company faces stiff competition, as it must prove that its sovereign stack can match the performance and reliability of global AI models provided by large technology corporations. The next monitorable for the industry will be the adoption rate of these sovereign models by public sector enterprises and large financial institutions, as well as the company's ability to maintain a technological edge in a rapidly evolving AI market.