NextFin News - In a significant leap for healthcare technology in the Global South, Bengaluru-based health AI platform Eka Care announced on February 18, 2026, a strategic collaboration with NVIDIA to develop India’s first offline-capable, unified medical scribe model. The announcement, made during the India AI Impact Summit in New Delhi, outlines a plan to deploy an end-to-end multilingual AI tool designed to automate clinical documentation. Unlike traditional cloud-based AI services, a specialized variant of this model is engineered to run entirely on a doctor’s local device, ensuring real-time transcription and documentation even in areas with limited or no internet connectivity.
According to Analytics India Magazine, Eka Care is utilizing the NVIDIA AI software stack, specifically the NVIDIA NeMo Curator, to process and structure vast datasets of Indian medical conversations. To ensure the model retains its specialized medical knowledge while adapting to new linguistic nuances, the company is integrating the NVIDIA Nemotron-CC dataset during the pre-training phase. This technical approach aims to solve the problem of "catastrophic forgetting," where AI models lose accuracy in core domains as they are fine-tuned for specific tasks. The initiative is led by Vikalp Sahni, Founder and CEO of Eka Care, who emphasized that the goal is to build a "sovereignty-first" model that prioritizes patient data privacy by keeping sensitive information on the edge rather than the cloud.
The development of an offline-capable scribe is a direct response to the unique structural challenges of the Indian healthcare system. While India has seen a massive digital transformation, rural and semi-urban clinics often face inconsistent bandwidth. By moving the compute requirement from the cloud to the device, Sahni and his team are effectively removing the latency and connectivity barriers that have previously hindered the adoption of AI in clinical settings. Furthermore, the model is being trained to handle the complex "code-mixing"—the blending of English with local languages—and the diverse accents prevalent across India’s 22 official languages.
From a technical perspective, the collaboration leverages NVIDIA’s accelerated computing infrastructure to evaluate state-of-the-art architectures like the Parakeet automated speech recognition (ASR) models. The objective is to merge ASR and Large Language Model (LLM) capabilities into a single, high-efficiency architecture. This consolidation is critical for running AI on mobile devices or tablets, where computational resources are far more constrained than in data centers. Tobias Halloran, Director of EMEAI Startups and Venture Capital at NVIDIA, noted that providing startups like Eka Care with direct access to scalable AI infrastructure is essential for accelerating India’s AI ecosystem toward global competitiveness.
The broader economic and policy implications of this partnership are substantial. The project aligns with the IndiaAI Mission, a government-backed initiative to foster domestic AI innovation. As U.S. President Trump’s administration continues to emphasize technological competition and secure supply chains, the development of "sovereign AI" in India represents a growing trend where nations seek to build independent digital infrastructure. By reducing reliance on international cloud providers for sensitive health data, Eka Care is positioning itself at the forefront of a movement toward data localization and security.
Looking ahead, the success of Eka Care’s offline scribe could set a blueprint for other emerging markets. The ability to provide high-precision, low-latency AI tools in low-resource environments is a multi-billion dollar opportunity. As hardware continues to evolve—with specialized AI chips becoming standard in consumer-grade tablets and smartphones—the shift toward edge-based clinical AI is likely to accelerate. For the medical profession, this means a significant reduction in administrative burnout; currently, Indian doctors often see dozens of patients a day, leaving little time for the meticulous documentation required for modern electronic health records. If Sahni’s vision holds, the AI scribe will not just be a luxury for urban hospitals, but a fundamental tool for every practitioner in the country.
Explore more exclusive insights at nextfin.ai.
