Image courtesy: Getty Images
Nvidia, a global leader in chip technology, has introduced a new, lightweight AI model designed specifically for Hindi, India’s most widely spoken language. The company aims to capture a bigger slice of the growing AI market in India by making this technology accessible to a broader audience.
The launch comes just ahead of a high-profile meeting where Nvidia’s CEO, Jensen Huang, is set to discuss future AI developments with Mukesh Ambani, the chairman of one of India’s largest conglomerates, at a business conference in Mumbai.
The new AI model, named Nemotron-4-Mini-Hindi-4B, is a compact version of Nvidia's language models. With four billion parameters, the model allows companies to develop and customize their own AI systems tailored for Hindi-speaking users. “This model has been optimized and trained on a blend of real and synthetic Hindi data, along with an equal amount of English data,” Nvidia stated. This bilingual training enhances the model's efficiency and responsiveness in Hindi, making it a valuable tool for various business applications.
The first Indian company to adopt this AI model is Tech Mahindra, a major player in IT services and consulting. Tech Mahindra is using Nvidia’s technology to create "Indus 2.0," an AI model designed to support Hindi and numerous regional dialects spoken across India. With this development, Tech Mahindra seeks to offer businesses improved tools for reaching and engaging Hindi-speaking audiences, enhancing customer experiences with virtual assistants, content translation, and more.
Despite India’s vast population of 1.4 billion people, only about 10% of the population speaks English. Recognizing this, Indian companies, from large corporations to new startups, have started investing in language-specific AI to better connect with local audiences. Language-based AI solutions can be especially effective for customer service, where conversational agents in the local language are more relatable and useful.
Unlike advanced AI systems like GPT-4 from OpenAI, which require extensive data and computing resources, smaller models like Nvidia’s Nemotron-4-Mini-Hindi-4B are easier and more affordable to train. This makes them appealing to companies with limited resources, as they can still harness AI for various applications without massive investments.
This move from Nvidia also coincides with India’s broader push to build a strong semiconductor industry. Major global tech firms are investing in India to set up production and research facilities as the country aims to position itself as a key player in the semiconductor industry alongside established hubs like Taiwan. However, experts suggest that India’s journey toward a mature semiconductor industry will take time.
Nvidia has a longstanding presence in India, having established its footprint nearly two decades ago. The company operates design and engineering centers in multiple locations, including Bengaluru, India’s tech capital, and Hyderabad. By launching a Hindi-focused AI model, Nvidia is strengthening its relationship with India’s tech industry and supporting the country's efforts to integrate AI and semiconductor technology into its economy.
