Manufacturing IoT Platforms: AWS IoT vs Azure IoT Hub vs Self-Hosted

Industrial IoT platforms diverged in 2026. The three credible options for connecting plant floors to data platforms.

Manufacturing IoT Platforms: AWS IoT vs Azure IoT Hub vs Self-Hosted

Industrial IoT platforms diverged substantially in 2026. The three credible options for connecting plant floors to data platforms — AWS IoT, Azure IoT Hub, and self-hosted (typically based on open-source MQTT brokers plus custom platform) — have substantially different operational profiles. The choice substantially affects deployment cost, operational burden, and vendor lock-in for decade-plus equipment lifecycles. This post walks through what’s actually deployed.

What manufacturing IoT does#

Manufacturing IoT platforms handle substantial scope:

Device connectivity. PLCs, sensors, gateways from various manufacturers (Rockwell, Siemens, Schneider, plus the various) connecting to cloud or edge platform.

Protocol translation. OPC-UA, Modbus, Profinet, MQTT — various protocols need bridging.

Edge compute. Processing at plant floor before sending to cloud. Bandwidth, latency, reliability considerations.

Data ingestion. Time-series data at substantial high frequencies (sometimes sub-second).

Analytics. OEE (Overall Equipment Effectiveness), predictive maintenance, quality analytics.

Integration. MES (Manufacturing Execution Systems), ERP, maintenance systems.

Security. OT (Operational Technology) security has substantially different concerns than IT security.

AWS IoT#

AWS IoT provides broad capability:

Strengths:

  • Managed service: broker, device registry, rules engine.
  • Greengrass for edge compute.
  • SiteWise for industrial-specific use cases.
  • Broad AWS integration with S3, Kinesis, Lambda, plus the various.
  • Strong security capabilities.

Trade-offs:

  • Substantial cost at large scale.
  • AWS-anchored — substantial lock-in.
  • Substantial pricing complexity.

Best for: AWS-anchored enterprises with AWS skills.

Azure IoT Hub#

Azure IoT Hub provides comparable capability:

Strengths:

  • Managed service with broad capability.
  • Azure IoT Edge for edge compute.
  • Digital Twins for digital-twin modeling.
  • Deep Microsoft enterprise integration — natural fit for Microsoft-anchored manufacturers.
  • Strong OPC-UA support.

Trade-offs:

  • Substantial cost at scale.
  • Microsoft-anchored — substantial lock-in.

Best for: Microsoft-anchored manufacturers; common in Europe and established discrete manufacturing.

Self-Hosted#

Self-hosted IoT platforms typically build on:

  • Eclipse Mosquitto, HiveMQ, or EMQX as MQTT broker
  • Apache Kafka for event streaming
  • TimescaleDB or InfluxDB for time-series
  • Node-RED or custom services for protocol translation
  • Custom platform for device management

Strengths:

  • Substantial cost advantage at scale (infrastructure cost rather than per-message pricing).
  • Portability — run anywhere.
  • Substantial flexibility.
  • No vendor lock-in.

Trade-offs:

  • Substantial operational burden — DIY operations.
  • Substantial skill requirement.
  • Substantial integration work with cloud services.

Best for: larger manufacturers with substantial OT/IT capability; multi-cloud or hybrid deployments.

The decision framework#

For most manufacturers in 2026:

Pick AWS IoT when AWS-anchored and AWS skills available.

Pick Azure IoT Hub when Microsoft-anchored and Microsoft skills available. Frequently natural fit for manufacturing.

Pick self-hosted for large-scale deployments where cost savings justify operational burden.

Pick hybrid — edge self-hosted plus cloud-managed — for common middle ground.

Pick OEM platforms (PTC ThingWorx, GE Digital, Siemens MindSphere, Schneider EcoStruxure, plus the various) for integrated industrial offerings.

What we typically see at clients#

Common patterns:

AWS or Azure at smaller-to-mid manufacturers. Managed service simplicity wins.

Self-hosted at larger enterprises. Cost economics justify operational investment.

OEM platforms at heavy industrial deployments.

Hybrid — increasingly common pattern.

Where pdpspectra fits#

Our data engineering practice builds production manufacturing IoT platforms with appropriate platform selection.

Related reading: the mining heavy equipment telematics post, the field service post, and the maritime port operations post.


Manufacturing IoT platform choice substantially matters. Talk to our team about your industrial IoT platform.