AI is increasingly used in Indian hospitals for documentation, diagnostic support, scheduling, billing and administrative tasks, with reported gains mainly in task completion time and work capacity, according to interviews with five healthcare leaders and a Bain–HealthQuad report. Whether AI will make treatment cheaper remains unproven. Dr Manivannan Selvaraj, founder and managing director of Kauvery Hospitals, said the highest costs are not hardware but usage-based fees, data protection, integration and training, and that savings were not sufficient to reduce prices. “Today the gains show up as greater capacity and better quality rather than lower bills,” he said. The 2026 Bain & Company and HealthQuad report, AI in Indian Healthcare Delivery, identifies workflow automation, clinical documentation, claims processing and selected diagnostic applications as areas where hospitals are already seeing operational gains, but says demonstrating business returns remains difficult. The report says AI model costs fell 92%, from US$36 per million tokens for GPT-4 in March 2023 to US$3 for GPT-5 in August 2025, and H100 GPU rental costs also declined, though hospitals must budget for software licences, system integration, data security, maintenance and staff training. Hospitals say savings are largely absorbed internally rather than directly reducing patients’ bills.