IBM CEO: India is competitive in AI, strong in deployment

IBM CEO Arvind Krishna said that India is not behind in AI and has an advantage in deployment, as it services many multinational companies.

During his visit to India, Krishna met Prime Minister Narendra Modi and discussed upskilling the country’s IT workforce.

He emphasized that AI will enhance productivity and highlighted IBM’s US$4.5 billion in internal gains from AI-driven efficiencies.

Krishna also addressed legacy systems, noting that mainframes remain critical for high-volume workloads like banking transactions.

He said AI tools can modernize codebases beyond COBOL and predicted that AI will augment many IT services, with about half of current tasks potentially handled by AI agents.

Krishna acknowledged potential job displacement but believes AI will create new opportunities, especially in transforming business models.

He urged India to focus on AI deployment and skill development rather than trying to build numerous models, leveraging its strong back-end and IT services sector.

🔗 Source: The Economic Times

🧠 Food for thought

Implications, context, and why it matters.

India is pursuing sovereign AI alongside deployment

Arvind Krishna urged India to focus on AI deployment, and India is also building its own systems.
IT minister Ashwini Vaishnaw said India plans to build 12 sovereign AI models, meaning AI systems built under the country’s governance and data rules, aimed at national challenges 1.
The plan sits within a five-layer national AI strategy that spans foundational models, infrastructure, and applications 2.
These models target multimodal use, meaning they can take inputs such as text and images, plus multilingual support 3.

AI pilots are giving way to production rollouts

Krishna linked deployment to an industry move away from speculative work toward measurable business gains.
He said 25% of CEOs get the AI return on investment they expect, because many teams build AI apart from core infrastructure 4.
IBM argues that smaller, specialized models trained on enterprise data deliver more value and can cost 30 times less to run 4.
The market is settling around specialized AI agents that fit into existing workflows, work IBM is pursuing with partners such as Pearson 5.