IBM Partners With IIT Bombay, IISc on Agentic AI, Sovereign AI, Quantum Computing

IBM has expanded its research collaborations with IIT Bombay and the Indian Institute of Science (IISc) to work on agentic AI, sovereign AI, quantum computing, and next-generation computing systems.

The collaboration will cover sovereign and Indic language model adaptation, multimodal AI, AI infrastructure, and knowledge retrieval. IBM and IIT Bombay have been working together since 2018.

The organisations will explore techniques to adapt and optimise sovereign and Indic language models for Indian languages. The research will also cover multimodal AI systems for software programming education, human-AI collaboration and operations across hybrid cloud environments.

The partners will also work on knowledge retrieval systems that can provide faster access to enterprise information. The research will include AI runtime optimisation and distributed inference to improve the performance and scalability of large language models.

With IISc, IBM will continue research in agentic systems, AI applications, and quantum computing. The collaboration began in 2021.

“Our collaborations with IIT Bombay and IISc bring together world-class academic research and IBM's scientific expertise to solve some of the most complex challenges in these areas,” Amith Singhee, Director, IBM Research India and CTO, IBM India and South Asia, said in a statement.

Prof Ganesh Ramakrishnan, Institute Chair Professor at IIT Bombay’s Department of Computer Science and Engineering, said the collaboration will cover multimodal AI models, knowledge retrieval, optimisation techniques, agentic AI workflows, and AI-assisted software engineering and education.

“By bringing together complementary strengths in AI, language, knowledge and intelligent computing, we envision trustworthy, inclusive, and scalable AI technologies that can transform how people learn, work, and access knowledge in India and across the world,” Prof Ramakrishnan said.

The IBM-IISc collaboration will involve lightweight time-series foundation models based on IBM’s Granite model family for energy analytics. The models will be used for forecasting, anomaly detection, load disaggregation, and energy optimisation. The partners will also contribute benchmark datasets to support further research.

In quantum computing, IBM and IISc will explore orchestration between quantum and high-performance computing systems. The research will include quantum-HPC algorithms and quantum subspace iteration methods using approximation-tolerant classical diagonalisation algorithms.

“IISc and IBM bring complementary research strengths to jointly explore intelligent, cloud-native platforms that can reason, adapt and orchestrate quantum-classical, edge AI and agentic workflows at scale,” Yogesh Simmhan, Professor in the Department of Computational and Data Sciences at IISc, said.