After China’s recent push with Kimi K3, India has entered the global AI conversation with one of its most ambitious announcements yet. At its Epoch Builder Edition event held on 30 July in Bengaluru, Sarvam AI unveiled a roadmap that goes far beyond launching another chatbot.
The company announced plans to build a trillion-parameter foundation model while expanding its AI ecosystem with new speech, coding and multimodal models, a developer platform, inference infrastructure and plans to establish a presence in the United States.
The announcements mark a significant step for India’s AI ecosystem. While many Indian startups have focused on building applications on top of existing AI models, Sarvam is investing in the technology stack itself. Its vision includes building large language models, AI infrastructure and developer tools that support Indian languages and enterprise use cases at scale.
If executed successfully, Sarvam’s strategy could position the company as India’s answer to OpenAI rather than simply another AI application company.
India’s AI Ambitions Just Became Much Bigger
India has traditionally focused on adopting AI rather than building frontier models. Most startups develop applications using APIs from OpenAI, Anthropic or Google instead of creating their own foundation models.
Sarvam AI is taking a different path. The company aims to build foundation models, AI infrastructure, sovereign AI capabilities and developer tools designed for Indian languages from the ground up.
This reflects a global trend. Countries are increasingly investing in sovereign AI to reduce dependence on foreign technology.
The UAE, for example, has backed companies like G42 and launched major AI infrastructure projects, including Stargate UAE, to build domestic computing capacity. China has supported companies like Moonshot AI and DeepSeek, while France and Saudi Arabia are also investing heavily in homegrown AI ecosystems.
Against this backdrop, Sarvam AI’s latest announcements show that India is also aiming to build its own AI infrastructure instead of relying entirely on global providers.
Expanding Beyond India With a San Francisco Office
Sarvam AI also announced plans to open an office in San Francisco, signaling that its ambitions extend well beyond the Indian market. The move follows a growing trend of AI startups establishing a presence in the United States to be closer to enterprise customers, investors, strategic partners and top AI talent.
Several fast-growing AI companies, including Harvey AI, have expanded their operations in San Francisco as the city continues to strengthen its position as the global hub for AI innovation. By establishing a presence there, Sarvam will be better positioned to build international partnerships while taking its AI products to a global market.
A New Developer Platform for Building AI in Indian Languages
Alongside its new models, Sarvam AI introduced a developer platform that supports AI applications across more than 10 Indian languages. The platform allows developers to build AI agents, integrate speech capabilities and deploy multilingual applications using Sarvam’s models and APIs.
Most frontier AI models deliver their best performance in English, creating challenges for businesses and developers serving India’s diverse population. Sarvam aims to bridge that gap by supporting languages including Hindi, Tamil, Telugu, Kannada, Bengali, Marathi, Malayalam, Gujarati, Punjabi and Odia. This allows developers to build AI experiences that feel more natural for Indian users instead of relying on translated responses.
As AI adoption continues to grow across industries, better support for Indian languages could help businesses, government organisations and developers create AI products that are accessible to a much larger section of the population.
