Following a recent US government export control directive that led to Anthropic suspending access to its Fable 5 and Mythos 5 AI models for foreign nationals, Sarvam AI CEO Pratyush Kumar has underscored the critical importance of India developing its own sovereign AI capabilities. Kumar emphasized that the incident serves as a stark reminder against confusing access to advanced technology with true ownership.
Anthropic's Ban and the Call for Sovereign AI
Anthropic, a leading AI company, announced on Saturday that it was compelled by a US government export control directive to disable access to its Fable 5 and Mythos 5 models for all foreign nationals, including its own foreign-national employees. While other Claude models remain unaffected, the abrupt nature of the shutdown has sent ripples through the global AI community.
Reacting to the development, Pratyush Kumar, CEO and co-founder of Sarvam AI, stated, "Fable ban is a good instigation for more people to engage in recognising the need for sovereignty." He elaborated that for AI users, companies, and nations, it's crucial to understand that "you should not confuse access with ownership, or adoption itself as advantage." Kumar warned that relying on significant technological differentiators with external control loops leaves one vulnerable.
Sarvam AI's Vision for India's AI Autonomy
Kumar passionately argued that the incident strengthens the case for "sovereign AI" in India, a concept he describes as Sarvam's founding thesis. "We need to have more countries and companies owning their own destinies. And in the post-AI world, that means being able to use and improve AI systems within their own perimeters," he explained.
Sarvam AI is actively working to make this vision a reality across several fronts:
- Compute Infrastructure: Sarvam is the first Indian team to train sovereign AI models at scale, utilizing approximately 3,400 H100 GPUs. The company has also brought India's first Blackwell cluster online and aims to operate tens of megawatts of AI compute capacity in India by 2027.
- AI Models: They have developed "Sarvam 105B," touted as India's first sovereign model built from scratch, and are now scaling towards trillion-parameter-class systems. A new coding-focused model is also slated for release soon.
- Inference Platforms & Products: Usage of Sarvam's hosted models has tripled in recent months. The company is preparing to launch a production-grade inference platform for diverse sectors including banks, governments, enterprises, and developers. Their voice AI products currently power millions of daily interactions, and document intelligence offerings are rapidly expanding.
- Enterprise Solutions: Sarvam is venturing into enterprise AI agents and building platforms that allow organizations to customize AI systems for specific use cases.
- Talent & Global Reach: The company is attracting top researchers with frontier AI experience and has begun establishing a San Francisco office to bridge India's AI ambitions with global advancements.
India's Potential in the Global AI Landscape
Kumar expressed strong confidence in India's capacity to lead in sovereign AI. He highlighted India's "talent, the scale of the economy, the ambition amongst people for a better life, and a government that understands the imperative for technology innovation." He concluded his remarks by emphasizing Sarvam's active recruitment efforts, signaling rapid growth and expansion.