OneQode signs for 14 MW of AI data-center capacity in Chennai

OneQode signed for an initial 14 MW at an AdaniConneX facility in Chennai.

OneQode announced on October 28 that it had signed an agreement with AdaniConneX for 14 megawatts of data-center capacity at Siruseri Sipcot in Chennai. The company describes it as its first major AI infrastructure deployment in India. The parties are continuing discussions on additional capacity, with OneQode citing a path to as much as 114 MW. The signed agreement covers the initial 14 MW. [OneQode announcement](https://en.prnasia.com/releases/apac/oneqode-signs-14-mw-ai-capacity-agreement-with-adaniconnex-with-a-path-to-114-mw-553497.shtml)

Capacity under contract is one milestone

The company says the site will support AI training, inference and enterprise and sovereign workloads. Those customers may want different things from the same location. A training team might need many accelerators and high bandwidth between them. An inference buyer may care more about latency, predictable availability and where data can be processed. A capacity figure gives neither team enough information to plan a deployment.

Signed data-center capacity also differs from live accelerator inventory. Hardware has to arrive, be installed and connected, and pass commissioning. Customers then need service terms and a way to confirm when resources are available. The announcement describes the agreement and intended uses. It does not supply a live inventory or customer start dates. [OneQode announcement](https://en.prnasia.com/releases/apac/oneqode-signs-14-mw-ai-capacity-agreement-with-adaniconnex-with-a-path-to-114-mw-553497.shtml)

Expansion remains a separate decision

The proposed route to 114 MW points to a larger ambition for the India business. The company also puts the Chennai agreement in the context of a wider Asia-Pacific expansion that includes Thailand. Beyond the initial tranche, additional agreements would provide evidence that plans had moved from discussion toward construction or operation.

For potential customers, the useful checkpoints are the first service start, the types and quantities of accelerators made available, and what the provider can demonstrate about reliability and data handling. Those facts can be compared with the original agreement without confusing the initial commitment, a possible expansion and resources customers can actually use. [OneQode announcement](https://en.prnasia.com/releases/apac/oneqode-signs-14-mw-ai-capacity-agreement-with-adaniconnex-with-a-path-to-114-mw-553497.shtml)