Once a policy ambition, sovereign cloud is gradually emerging as a market reality in India, with home-grown vendors beginning to secure large orders from banks, insurers and government agencies. Among them, Tata Communications is positioning itself at the centre of this shift with a sovereign, full-stack cloud platform designed to meet compliance, security and the demands of AI workloads.
“Our aim is to offer a platform that is sovereign by design and fully integrated across infrastructure, data and AI layers,” Neelakantan Venkataraman, Vice President and Global Head of Cloud, AI and Edge Computing, Tata Communications, told TechObserver.in’s Mohd Ujaley in an exclusive interview.
Edited Excerpts:
What drives Tata Communications’ cloud strategy in the BFSI sector?
We have been in this space for over ten years, serving more than 350 customers. We run one of the largest cloud footprints in India with multiple cloud nodes globally. In India, our nodes are located in Delhi, Mumbai and Bangalore.
Recently, we added a general-purpose compute cloud node in Chennai. We had already launched our AI cloud node there and have now expanded it with a general-purpose compute node as well.
We take pride in having built a full-stack cloud platform, from infrastructure as a service to a rich platform as a service layer. We are not just an infrastructure provider; we also enable platform components such as data management, databases and key management services. We help customers publish APIs and operate securely on our scalable infrastructure.
We offer both general compute and GPU-based infrastructure. Our GPU stack is based on Grasshopper architecture, built in partnership with NVIDIA, making us an NVIDIA Cloud Partner. Our compute is available as virtual machines or as Kubernetes clusters for modernised, container-based applications.
FinTech thrives on modular, service-oriented, API-driven architectures. We provide the containerised infrastructure for startups and FinTechs to build, test and publish APIs. Our full DevSecOps suite supports the entire development lifecycle, enabling secure, scalable deployments.
We have built strong resilience with multiple firewall layers and SIEM (Security Information and Event Management). As a multi-tenant public cloud provider, we cater to both core BFSI applications and emerging FinTech workloads that demand modernisation.
How do you help large financial institutions modernise legacy systems?
This is a key differentiator for us. We do not believe in a one-size-fits-all approach. Legacy applications cannot be re-platformed or re-architected overnight. They often involve decades of code and petabytes of data. SBI and LIC are examples of that scale.
We promote a hybrid cloud model, giving customers flexibility in where to run workloads, on dedicated on-premise infrastructure or modernised, cloud-native environments. This allows them to benefit from cloud scalability, resilience, flexibility and cost efficiency at their own pace.
We also provide professional services to re-platform and migrate applications gradually. We are not a full system integrator and do not rewrite code, but we enable transitions from legacy to cloud-based environments, even when applications are not yet containerised.
This approach helps customers move step by step, charting a path over five to ten years to modernise selectively. We provide end-to-end orchestration through a single management interface, a single pane of glass, enabling seamless movement of workloads across environments.
With Tata Communications and a preferred system integration partner, customers can manage complex IT landscapes through a single unified architecture.
How does your platform support testing and DevSecOps for FinTechs?
We have built a comprehensive DevSecOps platform tailored for FinTechs and independent software vendors. It includes development tools, testing environments, GitHub as a service, Git registries and CI/CD pipelines through Jenkins, all delivered as a platform as a service.
This setup allows developers to code, test, secure and deploy rapidly with continuous iteration and full security integration. It significantly reduces time to market and operational headaches.
From a total cost of ownership perspective, we are setting new industry benchmarks. Our pricing is predictable and affordable, making it easier for small and medium FinTechs to adopt our platform. While other providers may offer free credits initially, costs can escalate quickly. Ours remain transparent and sustainable.
What makes your AI strategy distinct for BFSI clients?
Two aspects make us stand out: sovereignty and our full-stack approach.
First, our entire cloud infrastructure is sovereign. No customer data, metering or telemetry leaves India. Many claim to be sovereign but true sovereignty means full compliance with Indian laws and zero exposure to foreign jurisdictions.
Second, we are a full-stack provider. We cater not only to infrastructure teams but also to application, data and AI teams, including emerging Chief Data and AI Officers.
From an infrastructure standpoint, our AI cloud is built on a highly resilient GPU architecture. GPUs are prone to failure, which can disrupt costly training jobs. We have addressed this through engineering innovation, introducing Direct Liquid Cooling (DLC), a first in India, at our Chennai data centre. It cools GPUs efficiently during heavy workloads, ensuring performance and reliability.
We have also built AI Studio, an end-to-end workspace for enterprises and startups. Data scientists can sign in, experiment using pre-integrated tools and data pipelines, and build AI/ML models without setup delays. It supports notebooks, fractional GPU allocation, data integration, guardrails, governance and full AI/ML Ops.
For customers who prefer simplicity, we offer a serverless option that automatically manages GPU allocation and billing control. We are also introducing an AI Agent Platform, currently in beta, and actively partnering with FinTechs, manufacturers and LLM providers to develop real-world use cases.
Through AI Studio and our marketplace, customers can test, deploy and scale ready-made AI models and APIs quickly, realising business outcomes faster.
What is the level of AI and cloud deployment among BFSI clients?
We already have major BFSI deployments. One of India’s leading payment companies is training a large language model for fraud detection on our GPU infrastructure. We are also working with insurance companies to enhance customer experience, for example, AI-driven voicebots supporting call centre agents.
Banks are proceeding carefully with AI because of regulatory sensitivities but they are actively exploring its potential. Most of these deployments are still in early production stages and are not yet mission-critical. Therefore, large-scale rollout is still evolving, but interest and experimentation are growing rapidly across the sector.
How are you addressing disaster recovery for AI workloads?
At present, we have not seen large-scale deployments for inferencing workloads in India. Most current implementations are in early production, which means they are not running core applications.
For instance, consider an insurance company using a voicebot to remind customers about premium renewals. If that bot goes down for a few hours, there is some productivity loss but no business disruption or financial impact. As AI adoption matures and such workloads become business-critical, DR will become essential.
However, DR for AI will not follow traditional data-centre models. The architecture will differ significantly, focusing on distributed model checkpoints, GPU workload resilience and automated recovery to ensure minimal loss during training or inferencing. Over time, we expect DR to be built natively into AI infrastructure design rather than treated as an external add-on.
