OneQode has signed an agreement with AdaniConneX for 14 megawatts (MW) of data center capacity at Siruseri SIPCOT in Chennai, marking the infrastructure provider’s first major AI deployment in India. The facility is intended to support large-scale AI training and inference, enterprise AI and sovereign AI workloads, expanding OneQode’s presence in Asia-Pacific as demand for compute-intensive applications grows.
OneQode Expands AI Infrastructure Into India
The agreement gives OneQode an initial 14 MW footprint in Chennai as it expands its infrastructure network for artificial intelligence workloads. AdaniConneX, a data center provider backed by the Adani Group and EdgeConneX, will supply the capacity under the agreement.
The deployment is designed to support several categories of AI computing, including training large models, running inference workloads and hosting enterprise AI applications. It also targets sovereign AI workloads, where organizations may require infrastructure that aligns with local data governance, regulatory or operational requirements.
OneQode said the agreement represents its first major AI infrastructure deployment in India. The company is also discussing additional capacity opportunities with AdaniConneX as it looks to expand its footprint in the market.
According to OneQode Chief Growth Officer and President for EMEA and India Johnny Blackfield, the discussions could involve up to 114 MW of additional capacity in India. That figure represents a potential expansion under discussion, not capacity confirmed by the initial 14 MW agreement.
The distinction matters as AI infrastructure providers compete to secure power, data center space and high-performance computing resources ahead of anticipated demand from model developers and enterprise customers.
Chennai Becomes Part of OneQode’s Asia-Pacific Strategy
The Chennai deployment forms part of OneQode’s broader Asia-Pacific expansion, which includes an AI infrastructure deployment in Thailand. The company says its position between Europe, the Middle East and Africa (EMEA) and Asia-Pacific makes India strategically useful for serving workloads across multiple markets.
AI infrastructure requirements vary by application. Model training can involve large-scale, sustained computing workloads, while inference—the process of generating outputs from trained models—can be more sensitive to latency, particularly for interactive applications and real-time services.
OneQode combines compute infrastructure with its own global carrier network, which it says is designed to support low-latency connectivity. Integrating the Chennai site into that network could give customers another location from which to run AI workloads and connect with users or systems across regional markets.
However, actual performance will depend on factors such as network routes, connectivity between facilities, available compute configurations and the location of end users. The announcement does not disclose the Chennai site’s planned GPU count, power usage effectiveness, commissioning schedule or expected network latency.
AMD Hardware Supports the Wider Infrastructure Roadmap
The agreement also builds on OneQode’s collaboration with AMD, announced in May 2026, to deploy AMD Instinct GPUs. OneQode said it plans to incorporate AMD’s Helios rack-scale solution as part of its broader AI infrastructure rollout.
AMD Instinct accelerators are designed for demanding computing workloads, including AI model training and inference. Rack-scale systems extend beyond individual accelerators by integrating compute, networking and other components into a larger deployment architecture.
For infrastructure providers, the challenge is not simply acquiring accelerators. AI deployments also require sufficient power and cooling, high-bandwidth networking, storage, software support and operational expertise. These elements influence how efficiently a data center can support large models and enterprise applications.
OneQode says its infrastructure spans more than 30 data centers across five continents. The Chennai capacity would extend that network into another major AI market, although the company has not disclosed whether all 14 MW will be available at once or how capacity will be allocated among customers.
India’s AI Infrastructure Market Attracts Investment
India has become an important market for data center development as cloud services, digital businesses and AI adoption increase demand for computing infrastructure. AI workloads can require substantially different infrastructure configurations from conventional enterprise hosting, particularly when large accelerator clusters and high-speed interconnects are involved.
The Chennai agreement reflects a wider industry focus on securing capacity in regional markets rather than relying exclusively on established data center hubs. Local infrastructure can support regional demand, while international network connectivity can help connect deployments to customers and services elsewhere.
Sovereign AI requirements may also influence infrastructure decisions. Organizations can face specific expectations around where data is stored, how it is processed and which entities control the infrastructure. The precise requirements vary by customer and workload, and OneQode has not identified the customers expected to use the Chennai deployment.
Power availability, construction timelines, network access and the economics of accelerator utilization will remain important factors in converting planned capacity into operational AI services.
Strategic Outlook
The 14 MW agreement gives OneQode an initial entry point into India’s AI infrastructure market and provides a potential basis for further expansion with AdaniConneX. Its combination of data center capacity, AMD-based computing plans and global network connectivity is intended to support customers running AI workloads across regions.
The next milestones will be the deployment’s operational availability, the specific computing configurations offered and whether discussions over additional Indian capacity lead to further agreements. Until those details emerge, the 14 MW commitment is the clearest confirmed measure of the company’s announced expansion in India.
Market Landscape
AI data center development increasingly combines high-density computing, accelerator-based servers, advanced cooling and high-performance networking. Providers are expanding geographically to serve regional AI demand, reduce connectivity constraints and address enterprise requirements. OneQode’s Chennai agreement with AdaniConneX adds planned capacity to this market, while the proposed expansion of up to 114 MW remains subject to further discussions.
Top Insights
- Initial 14 MW agreement: OneQode has secured data center capacity from AdaniConneX at Siruseri SIPCOT in Chennai.
- Multiple AI workloads: The deployment is intended for AI training, inference, enterprise AI and sovereign AI applications.
- Potential future expansion: The companies are discussing additional opportunities of up to 114 MW; this is not part of the confirmed initial commitment.
- AMD infrastructure roadmap: OneQode is building on its AMD Instinct GPU collaboration and plans to use the AMD Helios rack-scale solution in its wider rollout.
- Global connectivity strategy: The Chennai site is expected to connect with OneQode’s network spanning more than 30 data centers across five continents.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
