Primary partner · Deployment platform

AWS + Foundry Cortex — Bedrock AgentCore, SageMaker, IoT SiteWise, deployed for mid-market industry.

AWS is a named primary deployment platform. Where the operation runs on AWS, Foundry Cortex deploys natively — Bedrock AgentCore for agent infrastructure, SageMaker for industrial ML, IoT SiteWise for the historian extension, and S3 / Lake Formation for the industrial data backbone. Across South Africa, Nigeria, Kenya, Zambia, Ghana, India, Brazil, GCC, Australia, and Southeast Asia.

Why we partner with AWS.

  • 1

    Bedrock AgentCore is the 2026 agent infrastructure layer designed for enterprise-grade industrial agents. Native AWS deployment closes the loop between agent reasoning and AWS operational primitives.

  • 2

    SageMaker has the broadest enterprise track record for industrial ML model training, registry, and inference at scale. Time-series, computer vision, and tabular regression workflows are well-trodden.

  • 3

    AWS IoT SiteWise is the industrial historian extension that ingests OPC-UA, MQTT, and Modbus traffic into the AWS data plane — useful where the operation does not run PI System and wants a hyperscaler-managed historian.

  • 4

    AWS has production-ready regions in Cape Town (af-south-1), São Paulo, Mumbai, Sydney, Singapore, and Bahrain — covering most of the Foundry Cortex priority-market footprint with low-latency local deployment.

Where we deploy it.

Agentic operations

Bedrock AgentCore for agent runtime, AgentCore Memory for long-context state, AgentCore Identity for fine-grained tool-use authorisation.

Industrial ML

SageMaker for model training and inference. Feature Store for time-series feature engineering. JumpStart for foundation-model fine-tuning.

Industrial data backbone

IoT SiteWise for historian extension where PI System is not present. S3 + Lake Formation for governed industrial data lake. Glue for cross-source data preparation.

Regional footprint

af-south-1 (Cape Town), sa-east-1 (São Paulo), ap-south-1 (Mumbai), ap-southeast-2 (Sydney), me-south-1 (Bahrain). Low-latency local deployment across the priority markets.

The deployment pattern.

Foundry Cortex deployments on AWS run end-to-end on AWS-native primitives. Historian ingest via IoT SiteWise (or federation from PI System where present), model training and inference on SageMaker, agentic workflows on Bedrock AgentCore, observability through CloudWatch and X-Ray, and the customer's deployment topology kept within their AWS account boundary.

Where the operation already has AWS commitments — common in cross-border logistics, IPP renewables, and cloud-native manufacturing — this maps to existing capex authority and procurement frameworks. Where the operation prefers Azure, Foundry Cortex deploys natively there instead.

When AWS fits.

AWS is the default for: (a) operations with existing AWS enterprise agreements, (b) operations building on Bedrock AgentCore for agentic ops, (c) operations with cross-border logistics or IPP renewables footprints where AWS regional coverage is the differentiator, (d) operations where SageMaker is already established as the ML platform.

Foundry Cortex deploys natively to Azure where the operation prefers it — see the Azure partner page for that path.

Should you deploy on AWS?

The audit answers that question against your specific operation — data maturity, vendor stack fit, total cost of ownership, integration economics.

Run the readiness audit →