TATC focuses on advanced computing service options for Southeast Asian workloads, including model-access integration and reference bare-metal GPU configurations. Delivery details are defined for each customer engagement.
TATC separates unified upstream model access through Token API from dedicated physical infrastructure through Bare Metal.
Token API is a unified gateway for supported upstream model providers.
One access layer routes API requests to supported upstream model providers, cloud vendors and model-service channels.
The gateway experience covers API keys, access permissions, routing, usage records and billing visibility, while the selected upstream route provides model inference.
Model capability, version, processing route, availability and pricing depend on the selected upstream provider and are reflected in the model catalogue and console.
TATC rents dedicated, single-tenant physical GPU servers to customers for remote use. TATC owns, deploys and maintains the GPU equipment; the customer rents the full server under the agreed service terms.
At colocation sites, TATC rents rack space, power, cooling and related facility services from the data-centre operator. A TATC-owned data centre is a separate future plan.
Displayed GPU-hour rates and whole-server specifications are references for comparing possible configurations. Actual hardware, deployment, access, billing method and price depend on the customer's requirements and are specified in the customer contract.
Data governance and operational responsibilities are confirmed for each service arrangement.
Operational responsibilities and service levels are defined by product and signed agreement:
Note: Service features, allocation, hardware availability, processing location, limits and support arrangements are confirmed in the applicable quotation and signed agreement.