AWS Is Trimming Its IoT Portfolio: The 2026 Retirement List and What It Means for Your Platform Strategy

WZ-IT operates thingshost. Current hosting starts from €159.80 net/month; setup and service level follow the proposal. Former package prices, location lists and cost comparisons in this dated article describe the previous offer, not current WZ-IT terms. View current service scope.

Within twelve months, AWS has ended or frozen six IoT services and features: AWS IoT Analytics has been shut down since December 15, 2025, AWS IoT Events since May 20, 2026, SiteWise Monitor and FleetWise no longer accept new customers, Device Defender Detect follows on August 31, 2026, and Greengrass V1 reaches its end of support on October 7, 2026. AWS IoT Core itself is not affected, and that belongs to any honest assessment. Still, the accumulation raises a question every IoT decision-maker has to answer: how much of your architecture depends on services whose lifecycle is determined solely by the vendor? This article documents the complete retirement list with all deadlines and primary sources, puts the pattern behind it into context, and provides a strategy framework: when you remain in good hands at AWS, and when an open-source platform like ThingsBoard hosted in Germany is the sturdier path.

Also worth reading: AWS IoT Events Is Dead (EOL May 20, 2026): Migrate Detector Models to ThingsBoard · ThingsBoard vs. AWS IoT Core: Open Source vs. Amazon Cloud

The 2026 Retirement List at a Glance

Six Services, Six Deadlines: What AWS Is Winding Down in 2025/2026

The list is not an interpretation but the documented state of official AWS announcements. Two services are already fully shut down, three are frozen, and one still has its cutoff date ahead of it.

IoT Analytics and IoT Events: The Completed Shutdowns

AWS IoT Analytics, the managed pipeline service for IoT data preparation, reached its end of support on December 15, 2025; new customers had already been turned away since July 24, 2024 [2]. AWS IoT Events, the service for state logic and detector models, followed on May 20, 2026: new customers were excluded since May 20, 2025, and after the cutoff there is no more access to the console or resources [1].

One detail shows how hard this "end of support" is meant: the original end-of-support page has since disappeared from the AWS documentation itself, and the former product page redirects to the general AWS IoT overview. The only reliably citable version left is the archived state of the documentation [1]. Anyone who wanted to export definitions or look up configurations after the cutoff was too late.

These two cases are not a first. Fleet Hub, the fleet dashboard of AWS IoT Device Management, already ended on October 18, 2025 [11], and smaller services such as AWS IoT 1-Click (late 2024) had been discontinued before that. How to transfer existing detector models onto ThingsBoard in practice is documented in our dedicated migration guide.

SiteWise Monitor, FleetWise, Device Defender Detect: Frozen

SiteWise Monitor, the dashboard frontend of AWS IoT SiteWise, and the SiteWise Edge Data Processing Pack were moved to maintenance mode with the service availability update of October 13, 2025 and have not accepted new customers since November 7, 2025 [3]. As a replacement for the Monitor dashboards, AWS recommends migrating to Amazon Managed Grafana; the SiteWise core service and Managed Grafana remain explicitly fully supported, only the visualization layer is affected [4]. If you are rethinking your dashboard stack anyway, you will find the trade-off between an IoT platform and a pure visualization tool in our comparison ThingsBoard vs. Grafana.

AWS IoT FleetWise, the vehicle data service, was moved to maintenance mode with the announcement of March 31, 2026; it has been closed to new customers since April 30, 2026 [5]. AWS stresses that it continues to provide security updates and maintain availability for existing customers; no shutdown date has been announced [6].

At the security service AWS IoT Device Defender, the wind-down affects only one part: the Detect feature, the anomaly detection on device behavior, was announced for maintenance mode on June 30, 2026 and will no longer be available to new customers starting August 31, 2026 [7]. The audit feature is not affected; AWS states verbatim that there is no change to IoT Device Defender audit availability [8].

Greengrass V1: Cutoff October 7, 2026

The next hard deadline belongs to the first generation of the edge runtime. AWS ends support for AWS IoT Greengrass Version 1 on October 7, 2026; after that, the V1 console and resources are no longer accessible, and the official path is migration to Greengrass V2 [9]. This is not new: the V1 software has been in its extended life phase since June 30, 2023 [10], and the October 2025 update officially listed Greengrass V1 as "Entering Sunset" [3]. More than three years of lead time is fair, but the end now has a date.

What Is Explicitly Not Affected

A clean assessment includes what is not on the list. AWS IoT Core, the MQTT broker and centerpiece of the portfolio, does not appear in any of the three service availability updates from October 2025, March 2026, and June 2026, and has no lifecycle announcement [3][5][7]. The SiteWise core service with asset models and data ingestion remains fully supported [4], Greengrass V2 is the actively recommended migration path [9], and the Device Defender audit feature continues unchanged [8]. If you use only IoT Core for connectivity and device management today, there is no acute pressure to act.

Fairness also applies to the how: AWS announces these steps with documented deadlines, publishes migration guides [4][9], and keeps security updates running in maintenance mode [6]. This is an orderly wind-down, not an overnight shutdown. "AWS is abandoning IoT" would simply be wrong. The strategic question arises nonetheless, because it does not hinge on the tone of the announcements but on the structure: the vendor alone decides the lifecycle.

The Pattern Behind It: Hyperscalers Consolidate on the Core

A look beyond the AWS fence shows this is not an Amazon quirk.

Google set the precedent: Google Cloud IoT Core was shut down completely on August 16, 2023, existing connections included [12]. No maintenance mode, no remaining core, but a complete withdrawal from the direct IoT platform business.

Microsoft delivered an episode in 2024 that shows how nervous the market has become. In February 2024, a message appeared in the Azure portal saying Azure IoT Central would be retired on March 31, 2027 [13]. Microsoft withdrew the message shortly afterwards and declared it a system error [14]. So there is no confirmed retirement, and that is exactly how it should be phrased. The episode remains remarkable nonetheless: a single portal message was enough to trigger migration discussions in many companies, because hardly anyone considered such an end unthinkable.

In parallel, the specialist market is consolidating: on February 19, 2026, SUSE acquired the proprietary IIoT platform Losant and announced plans to open source the technology [15]. The direction is telling: a proprietary standalone product is to become an open stack, because openness has become a selling point in the industrial space.

The pattern: hyperscalers retreat to core infrastructure, meaning connectivity, data streaming, and storage, and leave the vertical functionality layer above it to partners or to self-builds. For you as a decision-maker this means: the lifecycle risk of proprietary IoT services rests with the vendor. With open source it rests with the user, and there it is plannable.

What Maintenance Mode Means for Your Planning

Maintenance is not shutdown, and this distinction matters to avoid both dramatizing and downplaying. Using FleetWise as the example, AWS documents what the status means in practice: existing customers keep using the service, security updates continue, availability remains; no shutdown date is named [6]. The same applies to SiteWise Monitor and Device Defender Detect: no end date, and none should be claimed.

For your planning, the status still means three things. First, it is an investment signal: the roadmap is frozen, and features you are missing today will not arrive anymore. Second, the new-customer stop closes the door on growth: the building block is no longer available for new projects, new tenders, and new setups. Third, maintenance defines your exit window: as long as the status holds, you migrate without deadline pressure and on your own terms. Whether an end date eventually follows maintenance is the vendor's decision. The IoT Events and IoT Analytics cases show what an end date looks like when it comes: hard, including loss of access to console and resources [1][2]. That is not a forecast for FleetWise and friends, but it is a reason not to leave the timing of your own migration to the vendor.

The Strategy Question: Keep Building the Kit or Switch Platforms

When AWS IoT Remains the Right Choice

Honesty goes both ways. The AWS IoT kit remains a good choice if your IoT data is already deeply integrated into AWS workloads, for example a data lake or ML pipelines that deliver the actual business value. It also remains one if you need extreme elasticity because device counts and load fluctuate heavily, or if a seasoned cloud team can and wants to carry the discontinued building blocks as a self-build from Kinesis, Lambda, and friends. And anyone using only IoT Core and Device Management is simply not affected by the retirement list. Coexistence is legitimate too: IoT Core as the ingest layer, a platform like ThingsBoard as the application layer on top.

When an Open-Source Platform Is the Sturdier Path

The calculation flips as soon as dashboards, alarm logic, and device management are supposed to be product features rather than construction projects. Exactly this layer is what AWS thinned out in 2025/2026, and exactly this layer is the core of a platform like ThingsBoard. It also flips if your GDPR assessment should work without third-country complexity, if data location in Germany is a requirement, or if vendor independence is a strategic demand: nobody can switch off Apache 2.0 code on you. Honestly added: an open-source project can change direction too, which is why a documented LTS policy with public support windows matters [17]. We walk through the fundamental comparison of the two worlds in ThingsBoard vs. AWS IoT Core.

Decision Matrix: AWS IoT Kit vs. Managed ThingsBoard

Whether a managed instance, the vendor's public cloud, or self-hosting fits your team is broken down separately in ThingsBoard Hosting: Cloud vs. Self-Hosted vs. Managed.

The Concrete Way Out: ThingsBoard 4.3 and the Migration Path

The technical basis for the switch is more mature today than ever. ThingsBoard 4.3, released on January 20, 2026, ships Alarm Rules 2.0, new Calculated Fields types for geofencing, propagation, and aggregations across related entities and time series, API keys for REST access without user credentials, and system-wide enforceable two-factor authentication [16]. The 4.3 line is Active LTS with documented support until July 20, 2027; the current patch 4.3.1.3 dates from July 1, 2026 [17]. That is the contrast to the retirement list: the lifecycle is published in a release table, not in a deprecation email.

What matters most for the migration: the Community Edition is licensed under Apache 2.0 and speaks MQTT, CoAP, and HTTP [18]. Devices that send to AWS IoT Core via MQTT today do not have to learn a new protocol. The state and alarm logic from AWS IoT Events maps onto Alarm Rules 2.0, documented step by step in our migration guide on the IoT Events end of support. And dashboards, the role of SiteWise Monitor, are a core function in ThingsBoard rather than an add-on product. The limits stay honest: ThingsBoard does not replace the FleetWise-specific vehicle signal decoding (CAN/OBD) or a generic edge runtime like Greengrass 1:1; anyone who needs exactly those building blocks must solve them separately. Whether the Community Edition is enough or you need the Professional Edition with white labeling and SSO is clarified in ThingsBoard CE vs. PE 2026.

thingshost runs your dedicated ThingsBoard instance as a managed service: Standard (CE) for EUR 149.90 per month in the Germany region, EUR 169.90 per month in the seven other regions; Premium with an official PE license from EUR 249.90 per month (Germany region, from EUR 279.90 in the other regions); Enterprise with an HA cluster on request. Cancellable monthly, data processing agreement included, test free for 7 days [19]. The full tiering of the ThingsBoard hosting plans is publicly documented.

GDPR, NIS2, and the Data Act: The EU Perspective

A forced rebuild is the natural moment to decide on data location at the same time. Anyone running US cloud IoT has to assess third-country transfers; the CLOUD Act applies to US providers regardless of which region the data resides in. Companies subject to NIS2 obligations must manage supply chain risks, and the lifecycle risk of a platform service belongs squarely in that assessment; how those obligations translate into practice is shown in our guide to NIS2 in OT and IIoT environments. The EU Data Act, finally, brings access, portability, and switching obligations for connected products and thereby structurally strengthens the case for open, switchable platforms; we break down the details in EU Data Act and IoT Platforms 2026. Important: this article provides technical and economic context, it is not legal advice. Involve your data protection and legal advisors for specific questions.

Conclusion: Lifecycle Risk Belongs in the Risk Register

The 2026 retirement list is documented fact, not alarmism: two completed shutdowns, three frozen services, one scheduled end of support on October 7, 2026. Equally factual is what keeps running: IoT Core, the SiteWise core, Greengrass V2, the Defender audit. AWS is not giving up on IoT, but AWS has shown that the functionality layer above connectivity is not guaranteed forever. So treat the lifecycle of your platform like any other operational risk: take inventory, map deadlines, secure export access, decide deliberately. If you switch, the migration guide on the IoT Events end of support is the concrete starting point. And if you want to take back control over the lifecycle, an open-source platform under Apache 2.0 with managed operations in Germany gives you exactly that: the provider operates the platform, but the provider cannot take it away from you. What that looks like in practice is the option to have ThingsBoard run as a managed service.

Recommended reading: ThingsBoard vs. AWS IoT Core: Open Source vs. Amazon Cloud · ThingsBoard Hosting: Cloud vs. Self-Hosted vs. Managed · EU Data Act and IoT Platforms 2026

Sources

1. AWS IoT Events end of support (archived developer guide, snapshot Jun 4, 2026)
2. AWS IoT Analytics end of support (archived user guide, snapshot Dec 7, 2025)
3. AWS Service Availability Updates (What's New, Oct 13, 2025)
4. AWS IoT SiteWise Monitor availability change (official application guide)
5. AWS Service Availability Updates (What's New, Mar 31, 2026)
6. AWS IoT FleetWise availability change (official developer guide)
7. AWS Service Availability Updates (What's New, Jun 30, 2026)
8. AWS IoT Device Defender product page (end-of-support notice, audit availability)
9. AWS IoT Greengrass developer guide (V1 end of support, migration path to V2)
10. AWS IoT Greengrass V1 maintenance policy (extended life phase)
11. AWS IoT Blog: end-of-life for Fleet Hub effective Oct 18, 2025
12. InfoQ: Google Cloud IoT Core shutdown on Aug 16, 2023
13. The Register: Azure IoT Central retirement message (Feb 15, 2024)
14. Microsoft IoT Blog: Microsoft's Commitment to Azure IoT (correction)
15. SUSE press release: SUSE Acquires Losant (Feb 19, 2026)
16. ThingsBoard 4.3 release: Alarm Rules 2.0, Calculated Fields, API Keys
17. ThingsBoard releases overview (LTS roadmap, version history)
18. ThingsBoard GitHub repository (Apache 2.0, protocols)
19. thingshost.de - Managed ThingsBoard Hosting (regions, pricing, trial)