The BharatGen Technology Foundation signed a landmark five-year Memorandum of Understanding with Larsen & Toubro’s subsidiaries on April 29, 2026, to develop India’s first end-to-end sovereign AI compute platform. This partnership between the government-backed initiative and the private sector aims to create a fully indigenous ecosystem encompassing specialized AI silicon, cloud infrastructure, and foundational models. By integrating these components, the collaboration seeks to ensure national data sovereignty and reduce India’s reliance on global technology supply chains.
The Three Pillars of the Sovereign AI Stack
The partnership leverages the unique strengths of each entity to build a unified, nationally deployable AI stack. This end-to-end platform is designed to be highly energy-efficient and optimized for the specific linguistic and cultural requirements of India.
1. Indigenous AI Silicon by LTSCT
L&T Semiconductor Technologies (LTSCT), a fabless semiconductor subsidiary of Larsen & Toubro, will lead the architectural design and development of custom AI ASIC (Application-Specific Integrated Circuit) and xPU silicon platforms. These chips will be specifically tailored to accelerate BharatGen’s AI workloads, such as multilingual Large Language Models (LLMs) and multimodal AI systems. By focusing on high performance-per-watt and low latency, LTSCT aims to provide the hardware backbone necessary for large-scale AI deployment without relying on imported high-end processors.
2. Sovereign Cloud Infrastructure by L&T-Vyoma
L&T-Vyoma, the digital infrastructure and data center arm of L&T, will provide the hyperscale, AI-ready data center environments required to host the platform. This includes leveraging its state-of-the-art facilities, such as the 30 MW data center in Kanchipuram, Tamil Nadu, to ensure that all data and computational processes remain within Indian geographical and regulatory boundaries. The infrastructure will feature advanced cooling technologies and high-density configurations optimized for intense AI training and inference tasks.
3. Foundational AI Models by BharatGen
BharatGen, as the foundational AI initiative, will define the AI workloads and develop the underlying models. These include Large Language Models (LLMs), Small Language Models (SLMs), and multimodal systems designed to handle India’s vast linguistic diversity. By co-optimizing the models with the underlying silicon and cloud layers, the initiative ensures that the entire stack is fine-tuned for maximum efficiency and national relevance.
Strategic Importance and National Sovereignty
The concept of a sovereign AI compute platform is central to India’s digital future. It ensures that the critical infrastructure required for AI—including the chips, the data centers, and the foundational models—is designed, built, and controlled within the country. This reduces the risk of data breaches and ensures that sensitive national data is not stored or processed on foreign servers.
Furthermore, this partnership aligns with the IndiaAI Mission, a comprehensive national initiative aimed at creating a robust AI ecosystem. By developing indigenous hardware, India can mitigate the risks associated with global supply chain disruptions and reduce the high costs of importing proprietary AI accelerators. The project is expected to set new benchmarks for domestic technological self-reliance, positioning India as a global leader in responsible and sovereign AI development.
The signing of the MoU took place in the presence of Prof. Ajay Kumar Sood, the Principal Scientific Adviser (PSA) to the Government of India. His presence underscores the strategic weight of this initiative in achieving India’s goals under the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS).
Profile of the Collaborative Entities
Understanding the background of the partnering organizations provides context for the scale of this project.
BharatGen Technology Foundation
BharatGen is India’s sovereign AI initiative focused on building inclusive and multilingual foundational models. Officially launched on September 30, 2024, it is spearheaded by IIT Bombay through the TIH Foundation for IoT and IoE. The initiative is supported by the Department of Science and Technology (DST) and is part of a national consortium that includes several Indian Institutes of Technology. Its primary mission is to create AI tools that cater specifically to the diverse linguistic needs of India.
L&T Semiconductor Technologies (LTSCT)
Incorporated on November 29, 2023, L&T Semiconductor Technologies is a wholly-owned subsidiary of Larsen & Toubro. Headquartered in Mumbai, it operates as a fabless semiconductor company. LTSCT focuses on the design and development of advanced semiconductor products, aiming to strengthen India’s position in the global chip design market.
L&T-Vyoma
L&T-Vyoma is the digital infrastructure and AI cloud business of L&T. Formerly known as L&T-Cloudfiniti, the business was rebranded in November 2025. In April 2026, L&T incorporated Vyoma.AI Ltd as a dedicated subsidiary to handle its expanding network of hyperscale data centers. These facilities are designed to support high-density AI workloads using advanced liquid cooling technologies.
Key Takeaways
- BharatGen Technology Foundation and L&T subsidiaries signed a 5-year MoU on April 29, 2026, to develop India’s first end-to-end sovereign AI compute platform.
- The partnership involves L&T Semiconductor Technologies (LTSCT) for designing indigenous AI ASIC chips and L&T-Vyoma for sovereign cloud infrastructure.
- BharatGen, launched on September 30, 2024, is a government-backed initiative hosted at IIT Bombay and supported by the Department of Science and Technology (DST).
- The project aims to foster technological self-reliance and ensure data sovereignty by keeping AI models and infrastructure within Indian regulatory boundaries.
- The collaboration is a strategic part of the IndiaAI Mission and the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS).
- Prof. Ajay Kumar Sood, the Principal Scientific Adviser (PSA) to the Government of India, was present at the signing of the agreement.
