Experts say Sarvam AI hosting Chinese models like DeepSeek on its Token Factory is a smart business move. While some critics worry, industry leaders argue it helps Indian firms stay competitive. "From business perspective, it's a really good effort," said Hastimal Jangid, noting it supports growth while building domestic AI.

Sana Afreen and Hastimal Jangid explain why Sarvam AI's Token Factory could help Indian businesses access global AI technology while supporting the development of domestic AI capabilities.

Industry experts have pushed back against criticism surrounding Bengaluru-based Sarvam AI's decision to offer Chinese open-weight artificial intelligence models, including DeepSeek, Moonshot AI, and Z.AI, on its commercial Token Factory platform, calling it a pragmatic move that prevents Indian businesses from falling behind in AI adoption while funding homegrown foundational research.

Speaking to CNBC-TV18, Sana Afreen, Founder and CEO of Beyond the Loop, and Hastimal Jangid, Founding Member of Coozmoo, argued that evaluating the move purely through national sentiment misses the commercial, technical, and data-sovereignty realities facing Indian enterprises today.

The two experts believe Sarvam AI's decision to offer access to Chinese AI models through its Token Factory platform should be viewed from both commercial and technological perspectives, rather than purely through the lens of national sentiment.

Sarvam AI has faced criticism for providing access to models such as DeepSeek, Moonshot AI and Z.AI, despite being seen as a key player in India's sovereign AI ambitions. The company has maintained that Token Factory is a commercial venture and does not signal a change in its core mission of developing homegrown AI models.

Jangid said the move could help Indian businesses and developers gain access to advanced AI models while strengthening the country's technology infrastructure. He believes using existing open-weight models can help bridge the gap as Indian companies continue working on developing their own solutions.

"From sentimental perspective, it's not something which should be considered. From business perspective, it's a really good effort," Jangid said, adding that the initiative could face some initial criticism but benefit the Indian economy over time.

Afreen also backed the move, pointing out that Indian businesses cannot afford to lag behind global competitors while domestic AI capabilities are still being developed.

She said building frontier AI models requires significant investment, research and computing power. In the meantime, enterprises, banks and regulated industries need access to global AI technologies to remain competitive.

Afreen also highlighted the importance of data residency, saying that hosting models on Indian infrastructure could address concerns around sensitive data being processed overseas.

"We cannot afford to be two years behind because we are adopting AI while also trying to become 10 years ahead building it," she said.

Jangid also dismissed concerns that the platform could pose a security risk, arguing that Indian businesses already rely on foreign AI models through third-party cloud infrastructure. Hosting these models within India could offer an alternative while helping Sarvam generate revenue to support its broader AI ambitions.

Afreen said the availability of an Indian platform, including the option to pay in INR rather than US dollars, was another advantage.

Both experts stressed that developing homegrown AI capabilities will take time and sustained investment. They believe commercial initiatives such as Token Factory could help Indian companies access existing technology while continuing to build their own AI ecosystem.