In a move that aligns with the conversations around sovereign AI cloud, Tata Communications is gearing up to revolutionise the Indian and global AI landscapes with the launch of its AI Cloud. The AI Cloud offers a complete stack to train, deploy, and inference AI models for enterprises, startups, and the government.
While speaking with AIM, Neelakantan Venkataraman, Vice President and Global Head Cloud, Edge & AI Business at Tata Communications, explained that this is not a new direction for the company but rather an expansion of their existing cloud portfolio, which includes infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) offerings.
“We’ve been offering GPUs long before the world woke up to generative AI and even before OpenAI and ChatGPT became household names,” he said.
A critical driver behind Tata Communications foray into this space is the need for greater GPU horsepower to power AI models from the country. Partnering with NVIDIA has been a pivotal step in this journey. “We realised that empowering enterprise customers to train, fine-tune, and deploy generative AI models required strategic GPU capabilities. That’s why our partnership with NVIDIA became essential,” Neelakantan explained.
Some say India is late to the AI party. Neelakantan, too, believes there is some truth to this notion, but Tata Communications AI Cloud is well-positioned to address this. He described Tata’s offering as unique due to its “full stack” capabilities. Unlike other cloud providers who focus heavily on training workloads, Tata Communications offers solutions that cater to the entire AI lifecycle, including the often-overlooked inferencing stage.
“We ensure that after the models are trained, they are fine-tuned, quantized, and deployed for inferencing at the edge, where data is produced, ensuring real-time, low-latency responses,” he said.
Building AI for India and for Bharat
The focus on sovereign AI deployment in India is crucial and there’s a strong emphasis on building AI models within India, particularly for the local market. Sovereign clouds are essential for companies that maintain sensitive information, such as those in the telecommunications and financial services industry.
This includes multilingual models, reflecting the diverse linguistic landscape of the country. A key priority is to develop foundational models, and several Indian companies are already making strides in this direction. The sovereign cloud market is expected to grow significantly and reach $250 billion by 2027. In the meantime, the market for foundation models will be $30 billion by 2027.
“We’re proud to be partnering with some of them and will soon be announcing their names,” said Neelakantan.
The goal is to help establish India’s position as an AI Super Factory, which encompasses all segments of the market, from government initiatives to enterprise and foundational model builders. This ambition extends beyond the enterprise level, touching societal, community, and national dimensions.
Moreover, with India’s academic institutions, startups, and other stakeholders contributing significantly to building AI capabilities with local datasets, Tata Communications wants to be the number one partner in these initiatives.
“We’re not viewing this as a race but as a national mission to ensure India stays on par with global AI technologies. Beyond enterprise use cases, AI has tremendous potential in areas like climate research, genomics, healthcare, agriculture, and defence—all critical sectors for India,” said Neelakantan.
The Full-Stack Solution
The full-stack approach allows enterprise customers to build their own AI Super Factories. Neelakantan noted, “Every customer, whether a mid-sized enterprise or a large corporation, should have the ability to create hundreds of AI use cases, not just limited pilots or proof-of-concepts, making all of them AI Super Factories.”
Tata Communications AI Cloud supports a wide range of AI applications, ranging from ML algorithms to LLMs. “We want to empower enterprises to work on everything from complex ML algorithms to multimodal computer vision applications with the right performance and security in place,” he remarked.
To further strengthen this vision, Tata Communications has introduced the AI Studio, a comprehensive platform designed to enable enterprises to build, train, and deploy AI models efficiently. The platform simplifies the model development process and optimises each step, ensuring that businesses can bring AI solutions to market swiftly.
“Our goal is to make AI accessible to everyone, from enterprises to startups, by providing a complete stack of tools that help them harness the full potential of AI,” Neelakantan added.
In terms of infrastructure, Tata Communications is ensuring that it has the full stack to enable this transformation, from cloud to edge deployments. “Our approach is to allow customers, whether large enterprises or mid-sized firms, to build their own AI super factories, without the financial burden typically associated with large-scale AI projects,” he explained and added that providing scalable GPU resources and AI tools make it easier for businesses to go beyond pilot projects and implement AI at scale.
From a technical standpoint, the AI Studio will enable enterprises to build, train, and deploy generative AI models with ease. Tata Communications is offering a fully managed service, ensuring that even if customers haven’t decided on the exact model or use case, they can quickly run evaluations, experiment with various models, and innovate at a faster pace.
“For instance, we provide access to a range of models like Mistral and Llama 2, allowing flexibility in choosing the best fit. Additionally, our intention to integrate with platforms like Hugging Face broadens the range of pre-trained models customers can experiment with,” Neelakantan explained.
For data management, the AI Cloud offers a comprehensive data management system, MLOps, and GenAIOps blocks, which help streamline the development and deployment process, including data preparation, fine-tuning models, and deploying them at scale. They also emphasise responsible AI practices, with built-in guardrails for privacy, bias reduction, and security to ensure reliable AI outcomes.
“One of the key differentiators of our platform is to provide the AI studio anywhere from on-prem to multi-cloud with Hybrid deployments, so that customers can collect, curate and govern data anywhere without worrying about compliance and data privacy and deploy AI models anywhere for optimal low-latency and sovereignty,” Neelakantan explained.
The post Tata Communications is Building Self-Reliant and Sovereign AI Cloud in India appeared first on Analytics India Magazine.
