Smarter AIoT at Scale: What’s New in Cumulocity Q2 2026
As AIoT deployments scale, the challenge is no longer just connecting devices; it’s ingesting, processing, and acting on data in real time. In the Q2 2026 release, Cumulocity introduces key innovations to support this shift, from high-throughput data ingestion to more intelligent analytics.
At the core is the new Cumulocity MQTT Service, enabling secure, scalable data ingestion at massive scale. Combined with AI-powered data preparation and enhanced analytics capabilities, this release helps organizations turn AIoT data into actionable outcomes faster.
Let’s dive in.
Application Enablement
AI-Agent Manager – Public Preview
Build smarter AIoT applications with no-code AI orchestration. The AI-Agent Manager enables users to create, configure, and manage multiple AI agents—without writing a single line of code:
- No-Code Orchestration – Design and manage multi-agent workflows visually
- Native Integration – Access inventory and telemetry data, with extensibility via MCP tools
- Model Flexibility – Choose from models such as GPT-4o, Claude, or Gemini
- Dynamic Context & Sandbox Testing – Feed real-time data into agents and validate safely
New in this release, agent configurations can now be exported as JSON—making it easier to move configurations across environments, enable version control, and integrate agent setups directly into application development workflows.
Simplifying Agent Deployment & Configurations
Managing AI agents is no longer limited to the UI. With this update, agent configurations can be exported as JSON—bringing them into the application development lifecycle:
- Portable Configurations – Move agent setups seamlessly across environments and instances
- Version Control Ready – Store and manage configurations alongside application code
- Consistent Deployments – Ensure alignment across development, testing, and production
- Improved Collaboration – Share configurations across teams and projects
- Lifecycle Management – Simplify documentation, backup, and ongoing maintenance
This enhancement makes it easier to operationalize AI agents at scale and integrate them into modern development workflows.
Transforming the Digital Twin Manager Homepage
The Digital Twin Manager (DTM) homepage has evolved from a static overview into a fully customizable dashboard:
- Customizable Dashboard Experience – Add and configure widgets to display relevant operational data
- Domain-Specific Views – Tailor terminology and layout to match your business context
- Consistent UX – Aligns with dashboard capabilities across Cockpit and Device Management
This provides greater flexibility and allows teams to focus on the insights that matter most in their operations.
Unified Terminology: Asset Definitions
To improve consistency across the platform, terminology has been standardized:
- Asset Models → Asset Definitions
- Property Library → Property Definitions
- Navigator Updates – Consistent use of “Definitions” across all views
- API Alignment – Matches terminology used in the Digital Twin Definition API
This creates a more intuitive and unified modeling experience across applications.
Datapoints Library Migration
The Data Points Library has been moved into dtm-plugins as part of a broader shift toward unified modeling:
- Seamless Transition – Delivered as a plugin without structural changes
- Fully Backwards Compatible – Existing functionality remains unchanged
- Default Availability – Accessible across Cockpit applications
- Foundation for Unified Modeling – Aligns all modeling components within DTM
This ensures a smooth transition while enabling a more scalable and consistent modeling architecture.
Sharing Mode for Asset & Property Definitions
A new tenant configuration enables centralized management of definitions across enterprise tenant hierarchies:
- Centralized Governance – Manage definitions in the enterprise tenant
- Reduced Duplication – Eliminate inconsistent copies across subtenants
- Controlled Access – Subtenants inherit definitions when sharing is enabled
- Operational Efficiency – Avoid export/import workflows
This feature introduces stronger governance and consistency for large-scale deployments.
DTM API Enhancements
Enhancements to the Digital Twin Manager APIs improve flexibility, consistency, and performance:
- Asset API Support for Groups
- Query assets and groups together using includeGroups
- Enables gradual transition toward treating groups as assets
- Opposite Relations Support
- Efficient reverse lookup from device → asset
- Optimized for high-frequency queries (e.g. DTM Data Service)
- New endpoint: /assets/linkedSeries/opposites/{deviceId}
These updates simplify data access patterns and improve performance for digital twin use cases.
Enhanced Data Insights with New Aggregation Options
Time-series analysis becomes more powerful with expanded aggregation capabilities:
- New Aggregation Methods – Average, Sum, and Count join existing Minimum and Maximum
- Improved Visualization – Better representation of trends, totals, and overall activity
These updates enable richer insights directly within Data Graph and Data Explorer.
Modernized Dashboard Widgets
A broad set of widgets have been enhanced to deliver a faster, more consistent, and more flexible user experience:
Data Point List Widget
- Angular-based modernization for improved performance
- Decimal precision configuration
- Integration with Global Time Context
- Asset icons for better usability
- Export to CSV and XLSX
- Virtual scrolling for large datasets
Asset Table Widget – Public Preview
- Rebuilt using Angular with improved scalability
- Advanced filtering and configuration capabilities
- Support for sub-assets and flexible hierarchies
- Enhanced column configuration and display options
Pie Chart Widget
- Migrated to Angular with ECharts
- Faster rendering and smoother interaction
- Interactive legend for dynamic exploration
SCADA Widget (Upcoming)
- Built using modern Web Components (LitElement)
- Improved performance and security
- Greater flexibility through reusable components and encapsulated styling
Markdown & Map Widget Enhancements
- Built-in Markdown editor with live preview
- Improved clustering and visualization for map widgets
- Better handling of dense and overlapping data
Data Export Enhancements
Export functionality has been extended to the Data Point Graph widget:
- Export graph data directly from dashboards
- Simplified reporting and sharing
- Improved support for offline analysis
Unified Property Library
A new Property Library introduces a centralized and scalable approach to asset modeling:
- Centralized Management – Define and manage properties across devices, alarms, and events
- Flexible Definitions – Apply reusable property structures across entities
- Advanced Validation – Enforce data quality with regex rules and custom messaging
This provides a strong foundation for a more consistent and extensible Digital Twin experience.
Flexible Asset Modeling Capabilities
Digital Twin Manager now supports more adaptive modeling approaches:
- Flexible Hierarchies – Allow any sub-asset relationships without strict type constraints
- Dynamic Properties – Add data even if not predefined in the model
A new default asset model enables these capabilities out of the box, allowing teams to build and evolve digital twin structures more iteratively.
DTM Data Service – Private Preview
A new data ingestion service connects device data with asset models:
- Real-Time Processing – Built on Notification 2.0 and WebSockets
- Multi-Tenant Architecture – Dedicated processing per tenant
- Smart Processing Logic – Support for dynamic transformation via scripts or external services
- Reliable Delivery – Queue-based system ensures processing integrity
This enables more accurate and enriched digital twin representations from streaming device data.
Continuous Developer Enablement
- A new in-platform REST API explorer allows developers to explore and test API endpoints directly within their Cumulocity instance
- A new feature toggle view in the Administration application provides visibility and control over enabled and disabled features
And Many More Improvements
- “Now” button added to global time context for faster time selection
- Dashboard import/export now generally available
- Ongoing UX and performance enhancements across applications
Device Management & thin-edge. io
Device Info Dashboard per Device Type
Managing dashboards at scale is now easier:
- Create reusable dashboard templates
- Apply dashboards across all devices of a specific type
- Reduce manual configuration effort
Updated Device Management Homepage
A redesigned homepage improves usability and onboarding:
- More consistent and informative layout
- Built-in thin-edge.io enrollment widget
thin-edge. io 2.0 release
The thin-edge.io 2.0 release is a milestone update that simplifies on-device data (pre-)processing and cloud connectivity:
-
On-device Logic with tedge flows: A lightweight JavaScript runtime for writing custom data mapping and local analytics/filtering
-
Universal MQTT Connectivity: Configurable bridges allow you to connect to any MQTT-broker/platform (e.g., HiveMQ, Azure EventHub, ThingsBoard).
Extensible Configuration: A new plugin API supports custom file types and automates service restarts after updates to apply the new configuration.
Data Management
Introducing the evolved Cumulocity MQTT Service
The Cumulocity MQTT Service is a major addition to the platform, designed specifically for AIoT scale and performance:
- Built for AIoT Scale – Optimized for high-throughput, fan-in architectures handling millions of devices and messages
- Flexible Data Ingestion – Devices can send data in any format
- Device Isolation by Design – Each device operates within its own secure topic space
- Cloud-Native Architecture – Combines an MQTT broker with a shared messaging backbone
Security is built in by default, with support for certificate-based authentication and strong device isolation without complex configuration.
For more information please read our white paper or watch the webinar
Data Preparation – Private Preview
Transform raw device data into usable insights without custom development:
- AI-Powered Transformation – Generate transformation logic using natural language
- Built-In Data Quality – Automatic validation, enrichment, and harmonization
- Scalable Processing – Designed for high-volume MQTT data streams
This enables faster time-to-value by eliminating the need for custom ingestion pipelines.
General Improvements
- API performance improvements
- New confirmation dialog for optional session termination when updating SSO configuration
- Enhanced Single Sign-On configuration for Azure AD)
Find more details here - Improved role mapping and administration experience for tenants with a large cohort of users
Find more details here
AI & Analytics
Five new blocks expand the power and flexibility of Analytics Builder:
Smart Function Block: No-Code Meets Pro-Code
Combine the simplicity of visual modeling with the flexibility of code:
- Write custom logic in JavaScript within a single block
- Reduce complexity in visual models
- Leverage existing developer skills without proprietary syntax
Streaming Analytics AI
A new approach combining machine learning and AI reasoning:
- Detect anomalies using high-speed ML models
- Filter and manage alerts intelligently
- Generate contextual diagnostics using AI agents
This shifts analytics from simple alerting to actionable insights and guided decision-making.
-
ONNX Block (The Watchdog): Runs deployed ML models natively against the high-speed data stream to detect subtle, complex anomalies at machine speed.
Find more details here -
Rate Limiter Block (The Shield): Acts as the crucial traffic cop. It prevents alert storms by suppressing duplicate anomaly triggers, protecting your expensive AI agents from being overwhelmed.
-
AI Agent Block (The Investigator): Queries your AI by injecting live operational context into a prompt, returning a human-readable diagnostic dossier / briefing on what to do
Find more details here
Logger Block – Enable structured debugging and traceability
The Logger Block enhances debugging by enabling structured traces directly within your microservice flows:
- Direct Log Tracing – Write trace messages straight to microservice logs
- No Database Overhead – Avoid temporary test data and manual cleanup
- Easy Filtering – Use log levels and tags to pinpoint issues quickly
- Simple Toggle Control – Turn logging on/off instantly, even in production
Security & Trust
Cumulocity continues to invest in security and compliance:
- Trust Center – Access compliance reports, certifications, and policies
- Regulatory Hub (in development) – Guidance on frameworks such as CRA, NIS2, and the EU Data Act
Find more details here
And Much more
As always, this release includes a wide range of smaller enhancements and optimizations driven by customer feedback—ensuring continuous improvement across the platform.