'Don't confuse access with ownership': Sarvam CEO on what Anthropic's Fable ban means for India

Sarvam CEO and co-founder Pratyush Kumar on Sunday said the ban on Anthropic's Fable 5 and Mythos 5 models was a reminder that countries should not confuse access to AI with ownership, arguing that the episode strengthens the case for "sovereign AI" in India.

Reacting to the shutdown of Anthropic's Fable 5 and Mythos 5 models for foreign nationals, Kumar said the episode was a reminder that access to cutting-edge technology should not be mistaken for ownership.

"Fable ban is a good instigation for more people to engage in recognising the need for sovereignty," Kumar wrote on Sunday. "But while many opine, Sarvam is at work making the Fable of India's first consequential AI company a reality."

Don't Miss: 'Ultimate weapon...': Sridhar Vembu urges India to build sovereign AI after US Anthropic curbs

Anthropic said on Saturday that the US government had issued an export control directive requiring it to suspend access to Fable 5 and Mythos 5 for all foreign nationals, including foreign-national employees working at the company.

"The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance," Anthropic said, adding that access to its other Claude models remained unaffected.

The Sarvam CEO suggested the development carried implications for AI users, companies, and countries alike.

"For AI users, it is clear that you should not confuse access with ownership, or adoption itself as advantage. And if the most significant tech differentiator you are leveraging has external control loops, then you have to accept you are vulnerable," he said.

He also argued that AI talent would increasingly be viewed through the lens of national interests and predicted that frontier AI systems would become more restricted even as general-purpose AI became widely available.

"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 — what one may call Sovereign AI," Kumar said.

Calling sovereign AI Sarvam's founding thesis, Kumar outlined the company's efforts across compute infrastructure, models, inference, products and deployment.

He said Sarvam was the first Indian team to train sovereign AI models at scale and had trained models using approximately 3,400 H100 GPUs. According to Kumar, the company has also brought India's first Blackwell cluster online and is working towards operating tens of megawatts of AI compute capacity in India by 2027.

On models, Kumar said Sarvam had built "Sarvam 105B", which he described as India's first sovereign model built from scratch, and was now scaling towards trillion-parameter-class systems. He added that a coding-focused model would be released soon.

The executive said usage of Sarvam's hosted models had tripled over the past three months and that the company was preparing to launch a production-grade inference platform designed for banks, governments, enterprises, startups, and developers.

He added that Sarvam's voice AI products now power millions of interactions daily, while its document intelligence offerings were witnessing rapid growth. The company is also expanding into enterprise AI agents and building platforms that allow organisations to customise AI systems for their own use cases.

The CEO further said the company was attracting researchers with frontier AI experience and had begun establishing a San Francisco office to connect India's AI ambitions with global developments.

"India has the 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 said.

Kumar concluded with: "We are hiring!"