Tata Communications is gearing up to launch trials for its AI Cloud platform, developed in partnership with Nvidia, by the end of the year, while actively engaging with enterprise customers to secure initial deals. Managing Director and CEO AS Lakshminarayanan revealed that interest is already emerging from sectors like manufacturing and retail, with cautious exploration from banking firms focused on back-office AI applications.
The AI Cloud will be powered by Nvidia’s GH200 Grace Hopper Superchip, part of a larger collaboration announced in September between Tata Group and Nvidia to deliver advanced AI infrastructure. Tata Communications is not just acquiring GPUs but is also undertaking significant engineering work to establish data centers, architect the platform, and build the necessary infrastructure.
The Nvidia GH200 Grace Hopper Superchip will power the AI Cloud, part of a broader agreement between Tata Group and Nvidia announced in September likewise focused on bringing high-level AI infrastructure to the table. What Tata Communications is doing is not just buying GPUs; they are really engineering serious setup of data centers, designing the platform, and creating the infrastructure needed for it to function.
Lakshminarayanan said enterprise spending on AI today is directed at training and not on inference. He added this trend will change in the next few years since companies are now moving from the proof-of-concept projects to production deployment of AI. He expects the AI Cloud to be adopted in a progressive manner, first in model training and then in inference.
While India and the world as a whole are interested in AI, this is likely to be a low-scale PoC. Any enterprise will try to validate business cases, model trustworthiness, and data privacy before going full-scale into AI adoption.The Nvidia GH200 Grace Hopper Superchip will power the AI Cloud, part of a broader agreement between Tata Group and Nvidia announced in September likewise focused on bringing high-level AI infrastructure to the table. What Tata Communications is doing is not just buying GPUs; they are really engineering serious setup of data centers, designing the platform, and creating the infrastructure needed for it to function.
Lakshminarayanan said enterprise spending on AI today is directed at training and not on inference. He added this trend will change in the next few years since companies are now moving from the proof-of-concept projects to production deployment of AI. He expects the AI Cloud to be adopted in a progressive manner, first in model training and then in inference.
While India and the world as a whole are interested in AI, this is likely to be a low-scale PoC. Any enterprise will try to validate business cases, model trustworthiness, and data privacy before going full-scale into AI adoption.
Do you have a news tip for Lakshmishree reporters? Please email us at media@lakshmishree.com
Source: Moneycontrol
News Desk
