Tech • IA • Crypto
Building the Future of Voice-First Sovereign AI: Sarvam & NVIDIA
TL;DR
An India-based AI initiative is building a fully sovereign, open-source AI ecosystem for Indian languages, powered by large-scale data, Nvidia infrastructure, and population-scale deployment.
KEY POINTS
Full-Stack AI Development
The initiative focuses on building AI systems end-to-end, covering datasets, model development, and application deployment. This integrated approach enables tighter control over performance, optimization, and quality across the entire pipeline rather than relying on external components.
Sovereign and Open-Source Strategy
A central goal is technological sovereignty, ensuring that AI systems are developed and owned domestically. The effort emphasizes open-source development, allowing broader access and transparency while reducing dependence on foreign platforms.
Focus on Indian Languages
The project targets the complexity and diversity of Indian languages, addressing challenges such as linguistic variation and underrepresented data. This includes handling long-tail language scenarios that are often ignored in global AI systems.
Massive Data and Training Scale
Large language models have been trained from scratch using datasets comprising tens of trillions of tokens, millions of hours of audio, and billions of images. Extensive data curation pipelines ensure quality, supported by tools like Nvidia NeMo Curator.
Advanced AI Training Techniques
The models leverage pre-training, fine-tuning, and reinforcement learning, with reinforcement learning delivering consistent performance gains at scale. Training and inference are conducted using Nvidia’s Hopper GPU architecture and the NeMo framework.
High-Volume API Deployment
The platform currently handles more than 4 million API calls per day, making it one of the largest AI API deployments originating from India. This reflects both scalability and real-world adoption.
Developer Ecosystem and Compute Shift
With a large developer base, the initiative encourages building AI systems rather than merely consuming them. It highlights the growing importance of expertise in accelerated computing and AI-specific software stacks.
Population-Scale AI Vision
The broader aim is to deliver AI systems that reflect India’s diversity and operate at national scale, moving beyond niche use toward widespread societal impact.
CONCLUSION
The effort signals a push toward sovereign, large-scale AI infrastructure in India, combining open-source principles, advanced compute, and linguistic inclusivity to serve a vast and diverse population.
Full transcript
We always believed that AI is such a critical technology that a country of the size of India, we should be building AI grounds up in India. So the two main focus areas for us are the ability to do full stack. We go from uh the data sets required to the models to be built and also the applications uh that can be built on these models. And the second thing is we want to do it with foundational research uh which can improve each of these layers while being completely sovereign about it. We took up the problem of doing it for Indian languages and doing it completely in open source. We understood the nuances of languages, the long tale of challenges and so on. And that's where Serum was born. We have an API platform today uh which serves more than 4 million API calls a day. By far the largest AI API effort out of India. And we're doing all of this on technologies. We're using the entire Nvidia stack for training these models and inferencing them at scale. So doing this entire thing of the platform, the model and the applications gives us lot more levers of optimization and quality. For all the work that we do at Saram, we start with data. Uh and data requires curation entire pipelines around ensuring that the quality of data is good and we've been extensively using the Nemo curator platform. In fact, we have trained now large language models from scratch and all that data for tens of trillions of tokens for millions of hours of audio, billions of images. All of that has flown through Nemo curator and that tool has really scaled and now we understand that deeply and the value it brings. The training itself we have done with the Nemo framework. The pre-training, the fine-tuning, the reinforcement learning. In fact, reinforcement learning is something that's giving consistent dividends at scale and we've been using the Nemo RL framework for that. And of course we have been um doing inferencing with our models at at some fair scale. So we've been using training and inferencing stacks extensively primarily on the hopper series of GPUs. In India having such a large developer base I think should be building AI not just consuming AI and that requires being expert in the new stack of software for accelerated compute. Nvidia stack is one great example of where people can do this. I genuinely believe developers should think about this as the core of development going ahead because whatever we build will hit these genative AI models. With NVIDIA, we are looking really forward to building models that represent the diversity of India and also serve them at scale so that it actually is a population scale effort rather than a few people just using them.