All Buyer Guides
IoT & EdgeHigh Complexity

Buyer's Guide: IoT Platform & Edge Computing

Evaluate AWS IoT, Azure IoT, PTC ThingWorx, Siemens Insights Hub, Cumulocity, Particle, Samsara, and Telit Cinterion — in a market where Google, SAP, IBM, and Bosch already walked away, so platform longevity is the criterion that outranks features.

16 min read 8 vendors evaluated Typical deal: $100K – $2M+ Updated June 2026
Section 1

Executive Summary

In IoT, the deciding question is no longer which platform demos best — it’s which one will still be accepting your devices in five years, after Google, SAP, IBM, and Bosch already shut their platforms down.

The IoT platform category divides along a clear line: horizontal cloud platforms (AWS IoT, Azure IoT) that hand you device-management and analytics building blocks to assemble; vertical industrial platforms (PTC ThingWorx, Siemens Insights Hub, Cumulocity) tuned for manufacturing and asset monitoring out of the box; connected-operations suites (Samsara) that sell an outcome rather than a toolkit; and connectivity-led enablers (Particle, Telit Cinterion) that bundle the module, the SIM, and the cloud. The hyperscalers win on breadth and cloud integration; the industrial and operations suites win on OT connectivity and prebuilt use cases — and the right choice depends on whether you’re building a platform or buying an outcome.

But the first filter is no longer capability — it is survivability. Google Cloud IoT Core shut down in 2023; SAP, IBM, and Bosch exited the same way; even AWS has quietly retired IoT 1-Click and IoT Analytics; and ThingWorx and Cumulocity have both changed hands. This guide provides a vendor-neutral evaluation framework for 8 platforms — AWS IoT, Azure IoT, PTC ThingWorx, Siemens Insights Hub, Cumulocity, Particle, Samsara, and Telit Cinterion — weighing device management and connectivity at scale, edge-versus-cloud processing, OT integration, and the exit risk of betting a decade-long fleet on the wrong roadmap.


Section 2

Why IoT Platform & Edge Computing Matters for Enterprise Strategy

IoT selection hinges on what happens after the proof of concept: connecting heterogeneous and brownfield industrial equipment, securing and updating fleets across their lifetime, and processing data at the edge where bandwidth and latency demand it. The strategic question is horizontal build versus vertical buy — hyperscaler primitives offer flexibility, industrial platforms offer speed to a specific outcome — and it should follow your use case and engineering capacity.

🎯
Strategic Impact
An IoT platform is a decade-long commitment to a stack you cannot easily rip out — sensors, gateways, firmware, and integrations all bind to it — which makes vendor longevity a board-level risk, not a procurement footnote. Google shut down Cloud IoT Core in 2023; SAP, IBM, and Bosch left the same market; ThingWorx and Cumulocity have both been carved out and sold since 2025. Before you weigh a single feature, ask three questions: (1) Will this platform still exist, and still be invested in, across the life of my fleet? (2) If the vendor exits, how do I get my devices and data off it? (3) Does the architecture lock me in at the device, the protocol, or just the cloud?

Edge AI, tighter OT/IT convergence, and standardized industrial connectivity are reshaping the category and pushing intelligence closer to the device. Weigh each platform on edge capability, fleet security, and how openly it connects existing equipment, because an IoT platform is judged on the production scale it can sustain, not the pilot it can demo.


Section 3

Platform Strategy & Sourcing Decision

The IoT “build vs. buy” choice is really a four-way sourcing decision: assemble your own platform on hyperscaler primitives (IoT Core/Greengrass, IoT Hub/Edge), buy a horizontal application-enablement platform (Cumulocity), buy a vertical industrial suite tied to your OT estate (ThingWorx, Insights Hub), buy a turnkey connected-operations outcome (Samsara), or let a connectivity vendor bundle the whole stack from the module up (Particle, Telit Cinterion). The deciding variables are your engineering capacity, how much of the value is in OT integration versus the application, and — given Google’s, SAP’s, IBM’s, and Bosch’s exits — how much vendor-survival risk you can absorb. Frame the decision around the operating model and the exit path, not the demo.

Your Situation Recommended Path Rationale
Strong cloud engineering team, custom data and ML needs, already standardized on one hyperscaler Build on cloud primitives AWS IoT or Azure IoT give you device management, rules, and edge runtime as composable services that drop straight into your existing data and ML stack — maximum control, at the cost of owning the integration and the roadmap yourself.
Brownfield factory or plant with mixed PLCs, OPC–UA, Modbus and legacy controllers Vertical industrial platform ThingWorx (with Kepware connectivity) and Siemens Insights Hub ship the OT protocol drivers, asset models, and prebuilt manufacturing apps that turn months of integration into configuration — the value is in the OT plumbing you don’t have to build.
Want device management and app enablement but no allegiance to one cloud or one automation vendor Independent horizontal platform (Cumulocity) A cloud- and OT-neutral platform avoids hyperscaler and automation-vendor lock-in and runs on-prem or in any cloud; weigh its newfound independence after the management buyout as both flexibility and a vendor-scale question to diligence.
Fleets, vehicles, or distributed field assets and a small IT team that needs results now Turnkey connected-operations or connectivity PaaS Samsara delivers safety, telematics, and equipment monitoring as a managed outcome with its own hardware; Particle bundles the cellular module, global SIM, and cloud — both trade platform flexibility for time-to-value and a single throat to choke.
Long-lived fleet where a vendor exit would strand thousands of devices Buy for exit-ability, not just features Favor open protocols (MQTT, LwM2M, OPC–UA), standard data formats, and a documented data-export path so a future migration is an inconvenience, not a re-deployment — Google IoT Core customers learned this on a hard deadline.
⚠️
Common Pitfall
The most common IoT mistake is celebrating a successful pilot with no plan for production — underestimating fleet security, OTA firmware lifecycle, and OT integration, then stalling in proof-of-concept purgatory with no measurable value. The second-most-common is ignoring exit risk: betting a ten-year sensor fleet on a platform that may not survive the decade, with no answer for how devices and data get out. Start from a measurable business outcome, design for scale, security, and a migration path from the first pilot, and prefer a vertical platform only when it reaches that outcome faster than assembling cloud primitives yourself.

Section 4

Key Capabilities & Evaluation Criteria

Weight these domains against your use case and your tolerance for vendor risk. The traditional RFP over-indexes on dashboards and feature checklists; in IoT, the deciding factors are whether you can manage and secure a fleet over a decade-long lifecycle, connect the equipment you actually own, and still get your data out if the vendor walks away. Note that platform longevity carries real weight here in a way it would not for most software categories — a deliberate response to a market where established platforms have been shut down or sold.

Capability Domain Weight What to Evaluate
Device & Fleet Management at Scale 25% Zero-touch provisioning and onboarding, secure over-the-air (OTA) firmware and config updates with staged rollout and rollback, device registry and digital shadow/twin, certificate lifecycle, and proven operation at your target device count, not the demo’s
Connectivity & OT / Protocol Coverage 20% MQTT, HTTP, LwM2M and AMQP support; industrial protocols (OPC–UA, Modbus, MQTT Sparkplug) and PLC/historian connectors; cellular (LTE–M, NB–IoT, 5G), LPWAN, and global SIM/eSIM management; brownfield gateway and edge-broker options
Edge Processing & Autonomy 15% Edge runtime for local compute, stream processing, and ML inference; store-and-forward during disconnection; edge-to-cloud orchestration and remote deployment of edge workloads; footprint from gateway-class down to constrained “tiny edge” devices
Security & Compliance 15% Per-device identity and mutual TLS, hardware root of trust / secure element support, fleet-wide credential rotation, anomaly detection on device behavior, RBAC and audit logging on the console, and SOC 2 / ISO 27001 / IEC 62443 coverage
Platform Longevity & Openness 15% Vendor financial footing and strategic commitment to IoT, install base and roadmap cadence, reliance on open standards over proprietary lock-in, contractual data-export and device-migration guarantees, and a credible answer to “what happens to my fleet if you exit?”
Application Enablement & Analytics 10% Low-code app builder and dashboards, prebuilt vertical/industry apps, time-series storage and rules engine, AI/ML and generative-AI tooling on device data, and clean APIs into your data lake, ERP, and field-service systems
💡
Evaluation Tip
In the POC, run the unglamorous fleet-lifecycle test, not the dashboard demo: register a few hundred devices, push a staged OTA firmware update, force-fail some of them and confirm clean rollback, then pull a device offline for hours and verify store-and-forward and reconnection. Separately, make the vendor walk you through — in writing — exactly how you would export every device, credential, and historical data point if you left. The platform that handles a botched update and a hard question about leaving, not the one with the slickest live map, belongs at the top of your shortlist.

Section 5

Vendor Landscape

This is a market defined as much by who has left as by who remains. Google shut down Cloud IoT Core in 2023; SAP, IBM, and Bosch wound down their platforms; even AWS has pruned IoT 1-Click and IoT Analytics. The survivors fall into four camps. Hyperscalers (AWS, Microsoft) sell composable cloud primitives and bet on their broader cloud gravity. Industrial/OT platforms (PTC ThingWorx, Siemens Insights Hub, Cumulocity) own the factory-floor connectivity and asset models — though ThingWorx is being sold to private-equity firm TPG and Cumulocity was carved out of Software AG into independence, both in deals that closed or were announced in 2025. Connected-operations players (Samsara) sell a managed outcome with their own hardware. And connectivity-led enablers (Particle, Telit Cinterion) bundle the module, the SIM, and the cloud. Most shortlists end up comparing across these camps; open-source-leaning options such as Losant (acquired by SUSE in 2025) and ThingsBoard sit alongside them for teams that prioritize portability.

AWS IoT Leader — Hyperscaler

Strengths: Broadest cloud-IoT service portfolio — IoT Core for connectivity, Greengrass as an open-source edge runtime with local compute and ML inference, SiteWise for industrial data, FleetWise for vehicles, and TwinMaker for digital twins — backed by enormous scale and seamless ties to the wider AWS data, analytics, and ML stack. 2025 managed integrations added Zigbee, Z-Wave, and cloud-to-cloud connectors to unify heterogeneous fleets. Considerations: It is a toolkit, not a turnkey platform — you assemble and own the application and the integration work. Per-message and per-service pricing gets intricate at scale, the service sprawl has a real learning curve, and AWS’s own retirements of IoT 1-Click and IoT Analytics are a reminder that even the leader prunes peripheral services.

Best for: Cloud-mature teams already on AWS that want to build a bespoke platform on composable primitives with deep analytics and ML
Microsoft Azure IoT Leader — Hyperscaler

Strengths: IoT Hub for secure bidirectional device communication, IoT Edge for containerized edge workloads, Device Provisioning Service for zero-touch onboarding, and Azure Digital Twins for modeling environments — all wired into the broader Azure data platform, Fabric, and Dynamics 365. A natural fit where the enterprise already runs on Microsoft. Considerations: Roadmap signals have been muddled: Microsoft posted, then retracted as “presented in error,” a 2024 notice retiring the low-code IoT Central layer, and pointedly declined to commit to its future while citing its three-year retirement-notice policy — so treat IoT Central as supported-but-uncertain and anchor new builds on IoT Hub. As with AWS, it is a build-it-yourself toolkit, and IoT Edge expects Linux container skills.

Best for: Microsoft-centric enterprises building on IoT Hub and Digital Twins who can navigate the IoT Central question
PTC ThingWorx Strong — Industrial / OT

Strengths: One of the deepest industrial IoT platforms, purpose-built for manufacturing and connected products: rapid application modeling, prebuilt analytics for OEE and root-cause, and best-in-class OT connectivity through Kepware’s 150-plus device drivers. ThingWorx 10 added IoT Streams, TLS 1.3, and generative-AI tooling. Considerations: PTC announced in November 2025 that it is selling ThingWorx and Kepware to private-equity firm TPG (up to $725M, expected to close in the first half of 2026), to be combined with TPG’s GE Vernova Proficy assets. The technology is strong, but buyers must diligence the ownership transition and the new entity’s investment commitment. Premium pricing; industrial-focused, not for consumer IoT.

Best for: Manufacturers and industrial OEMs needing deep OT and PLM integration — who are comfortable underwriting the TPG transition
Siemens Insights Hub Strong — Industrial / OT

Strengths: Industrial IoT-as-a-service (formerly MindSphere), now positioned within Siemens’ Industrial Operations X and the Xcelerator ecosystem, with strong asset analytics, prebuilt industrial apps, and tight integration to Siemens automation hardware, Teamcenter, and Opcenter. A natural extension for shops already running Siemens controls. Considerations: Most valuable inside a Siemens-centric automation estate; less compelling as a neutral horizontal platform. The MindSphere-to-Insights Hub rebrand and folding into Industrial Operations X have created naming and packaging confusion that buyers must cut through; pricing is modular and can be complex.

Best for: Siemens automation customers wanting industrial analytics and apps integrated with their existing controls and PLM
Cumulocity Strong — Independent IIoT

Strengths: A cloud- and vendor-neutral horizontal IoT platform with strong device management, a low-code application builder, edge and on-prem deployment options, and broad protocol support — consistently recognized as an industrial-IoT leader by analysts. Its neutrality is the pitch: it locks you to neither a hyperscaler nor an automation vendor. Considerations: Carved out of Software AG in January 2025 via a management buyout (backed by Avedon Capital Partners, Schroders Capital, and Verso Capital) with founders Bernd Gross and Stefan Vaillant back at the helm. Independence brings agility but also a smaller balance sheet than a hyperscaler — diligence the funding runway and roadmap continuity through the transition.

Best for: Enterprises wanting a neutral, deployment-flexible platform without hyperscaler or automation-vendor lock-in
Samsara Strong — Connected Operations

Strengths: A turnkey connected-operations platform that sells an outcome, not a toolkit: AI dash cams, GPS and ELD telematics, equipment monitoring, and a single cloud across fleets, sites, and frontline workers, with its own ruggedized hardware. Fast time-to-value for physical operations and a public company (NYSE: IOT) with a clear standalone IoT business. Considerations: Optimized for vehicles, equipment, and field operations rather than general-purpose or product-embedded IoT; far less of an open application-development platform than the industrial suites; the appliance-plus-subscription model and proprietary hardware mean a different kind of lock-in than the cloud platforms.

Best for: Fleet-, equipment-, and field-operations organizations wanting safety, telematics, and asset monitoring as a managed outcome
Particle Strong — Connectivity PaaS

Strengths: A full-stack, cellular-first IoT platform-as-a-service: hardware modules with embedded connectivity, a device OS, global multi-carrier SIM/eSIM across hundreds of networks, OTA updates that scale to large fleets, and developer-friendly tooling — collapsing what would be a multi-vendor build into one stack. Strongest where Wi-Fi is impractical: remote, mobile, and field-deployed assets. Considerations: Best suited to greenfield connected-product and remote-asset programs rather than heavy brownfield factory integration; lighter on deep OT/PLC connectivity and prebuilt industrial analytics than the industrial suites; a smaller vendor whose scale and longevity warrant the same diligence applied to any independent.

Best for: Product teams and remote/mobile-asset deployments wanting cellular connectivity, device management, and edge in one integrated stack
Telit Cinterion Niche — Modules + Connectivity

Strengths: Formed by Telit’s acquisition of Thales’ Cinterion cellular business (Thales retains a minority stake), one of the largest Western cellular-module makers, pairing 4G/LTE-M/NB-IoT/5G hardware with a connectivity-management portal, the deviceWISE device-management and edge platform, iSIM, and global eSIM. Strongest when the module, the SIM, and the management layer come from one accountable vendor. Considerations: Hardware-and-connectivity-led rather than an application-development platform — you bring or build the apps and analytics on top; faces intense competition from low-cost module makers; the post-merger Telit Cinterion portfolio still spans several product lines that buyers must map to their needs.

Best for: Connected-product builders wanting cellular modules, global connectivity, and device management from a single Western supplier
🔎
Market Insight
The defining dynamic of this category is platform attrition, and it cuts against the usual instinct to play it safe with a big name. Google, SAP, IBM, and Bosch all exited; AWS retired peripheral IoT services; PTC is selling ThingWorx to private equity; and Cumulocity and Losant were both carved out and sold — even as the hyperscalers’ combined share of the agnostic platform market rose sharply. Brand size is no guarantee of staying power here. The durable hedge is architectural: insist on open protocols (MQTT, LwM2M, OPC–UA), standard data formats, and a contractual export path, so that whoever owns your platform next, your fleet and your data remain portable.

Section 6

Pricing Models & Cost Structure

IoT pricing fragments by camp, and the unit of measure matters more than the headline rate. Hyperscalers meter consumption — messages, connected-device-minutes, rules executed, edge deployments — which starts cheap in a pilot and compounds unpredictably as the fleet and message rate grow. Industrial suites license by platform, connected asset, or named developer, often with prebuilt-app modules on top. Connected-operations and connectivity vendors bundle hardware with a per-device or per-vehicle subscription, frequently on multi-year terms. Model cost against your real device count, message frequency, and retention curve at production scale — and price in cellular data, edge gateways, and the integration labor the toolkit platforms push onto you.

Vendor Pricing Model Relative Tier Key Cost Drivers
AWS IoT Consumption (per-message, per-device, per-service) Low entry, scales up Messages and connected-device-minutes, rules/Greengrass/SiteWise usage, data egress, cross-service data and ML spend
Microsoft Azure IoT Tiered per-unit (IoT Hub) + consumption Low entry, scales up Daily message quota and Hub tier/units, IoT Edge and Digital Twins usage, Device Provisioning, downstream Azure data services
PTC ThingWorx Platform + connected-asset / named-user licensing Premium Connected assets, named developers, Kepware connectivity, prebuilt app modules, perpetual-vs-subscription terms
Siemens Insights Hub Modular subscription (assets, apps, capacity) Premium Connected assets, data ingest/retention, prebuilt application packs, edge devices, Xcelerator entitlements
Cumulocity Per-device subscription; cloud or on-prem Moderate Device count, message/data volume, edge and on-prem deployment, application and microservice add-ons
Samsara Hardware + multi-year per-device/vehicle subscription Moderate–Premium Vehicle/asset count, license tier (telematics, video safety, equipment monitoring), hardware, contract length
Particle Per-device subscription + cellular data + hardware Moderate Active device count and tier, cellular data usage, modules/dev kits, premium support and professional services
Telit Cinterion Module hardware + per-SIM connectivity + platform Varies (hardware-led) Module volumes, per-SIM/eSIM data plans, deviceWISE management seats, edge and security add-ons
3-Year TCO Formula
TCO = (Platform/Subscription × 36 months) + Device & Module Hardware + Cellular/Connectivity Data + Edge Gateways + Integration & OT Onboarding + OTA & Fleet Operations + Security − Avoided Truck Rolls & Downtime − Asset-Utilization Gains

Section 7

Implementation & Rollout

Sequence an IoT rollout around the leap that actually kills projects — the jump from a clever pilot to a secured, managed fleet at production scale. Prove connectivity and fleet operations on real devices early, lock down security before you scale, and treat OTA firmware management and edge deployment as first-class from day one, not afterthoughts you bolt on once the device count is in the thousands.

Phase 1
Connect & Prove (Months 1–2)

Stand up the platform, connect a representative slice of real devices (including the awkward brownfield and OT cases), validate protocols and ingestion end to end, and confirm a measurable use case — proof of value, not just a dashboard. Define exit-and-export terms in the contract before signing.

Phase 2
Pilot Fleet & Secure (Months 2–4)

Grow to a managed pilot fleet with per-device identity and certificate provisioning, exercise OTA firmware update with staged rollout and rollback, integrate identity (RBAC/SSO) and security monitoring, and harden the platform console as a high-value target before any scale-out.

Phase 3
Scale to Production (Months 4–7)

Roll out across sites and asset classes, automate zero-touch onboarding, deploy edge runtimes and local processing where bandwidth or latency demand it, wire device data into your data lake, ERP, and field-service systems, and establish fleet-operations runbooks and SLAs.

Phase 4
Operate & Optimize (Months 7–12)

Tune edge-vs-cloud processing and connectivity cost, add analytics and AI/ML on the device data, institute ongoing firmware and certificate lifecycle management, and review consumption and licensing against the original model as the fleet grows.


Section 8

Selection Checklist & RFP Questions

Use this checklist during evaluation to confirm each shortlisted platform covers what actually decides an IoT program at scale — fleet lifecycle, connectivity, security, and a clean way out if the vendor doesn’t survive the decade.


Section 9

Related Resources

Spotlight Listing

Interested in getting featured here?

Put your solution in front of the CIOs evaluating this category.

Learn how
Tags:IoTEdge ComputingAWS IoTAzure IoTThingWorxIndustrial IoT