Autonomous physical systems have been one of the most persistent threads in Gartner’s Top Strategic Technology Trends lists. Intelligent Things appeared in 2018, Autonomous Things in 2019 and 2020, Autonomic Systems in 2022 and Polyfunctional Robots in 2025. The 2026 and 2027 lists continue the thread with Physical AI and World Models. This post follows that path, checked against Gartner’s press releases, and turns it into guidance for operations and technology leaders in regulated EU industries.

The autonomous systems timeline

List	Gartner trend	Shift
2018	Intelligent Things	AI-driven behavior instead of rigid programming
2019	Autonomous Things	From stand-alone devices to collaborative swarms
2020	Autonomous Things	Toward uncontrolled public spaces
2022	Autonomic Systems	Self-managing systems that learn
2025	Polyfunctional Robots; Ambient Invisible Intelligence	One machine, many tasks; cheap sensing everywhere
2026	Physical AI	Machines that sense, decide and act
2027	World Models	AI that simulates physical outcomes before acting

Each list is published in October of the previous year.

2018: Intelligent Things

According to Gartner’s 2018 release, “intelligent things are physical things that go beyond the execution of rigid programming models,” using AI to deliver advanced behaviors and interact more naturally with their surroundings and with people. The examples were autonomous vehicles, robots and drones. Early applications:

• Smart sensors and IoT devices that adapt to conditions
• Autonomous vehicles in controlled environments
• Intelligent industrial equipment

2019: Autonomous Things

In the 2019 release, autonomous things such as robots, drones and autonomous vehicles “use AI to automate functions previously performed by humans.” Gartner expected “a shift from stand-alone intelligent things to a swarm of collaborative intelligent things.” Its example: a drone finds that a field is ready for harvesting and dispatches an autonomous harvester. The main categories:

• Air: drones for inspection, surveillance and delivery
• Land: autonomous vehicles and mobile robots
• Sea: autonomous vessels and underwater vehicles
• Industrial: collaborative robots (cobots) working next to people

2020: Autonomous Things, toward open spaces

Autonomous Things returned on the 2020 list. Gartner expected them to “increasingly be deployed in uncontrolled public spaces” as the technology matures, regulation allows it and society accepts it. In practice, deployment moves through three types of environment:

• Controlled: mines, warehouses and factories
• Semi-controlled: campuses, logistics yards and fixed routes
• Open: streets and other public spaces

Typical applications are delivery on defined routes, security patrols, agricultural robots and warehouse automation.

2022: Autonomic Systems

Gartner’s 2022 release describes autonomic systems as “self-managing physical or software systems that learn from their environments.” This is the step from systems that follow rules to systems that adapt their own behavior, which matters for testing and certification as much as for capability.

2025: Polyfunctional Robots

According to Gartner’s 2025 release, “polyfunctional machines have the capability to do more than one task” and are replacing task-specific robots custom-designed to repeat a single task. They are “designed to operate in a world with humans,” which allows fast deployment and easy scalability. The advantages:

• Efficiency: several capabilities in one system
• Faster ROI: higher utilization across tasks
• Adaptability: reassignment to new tasks without re-engineering the site

Gartner predicts that by 2030, 80% of humans will engage with smart robots daily, up from less than 10% in 2024. The same list included Ambient Invisible Intelligence: ultra-low-cost tags and sensors for large-scale tracking and sensing, a close relative of the 2018 intelligent things.

The 2026 and 2027 lists

Gartner’s 2026 list (October 20, 2025) includes Physical AI, which “brings intelligence into the real world by powering machines and devices that sense, decide, and act,” such as robots, drones and smart equipment. The 2027 list (October 20, 2026) adds World Models: AI that represents physical environments well enough for agents to simulate outcomes before acting. Gartner links its growth to investment in robotics and autonomous vehicles.

Technical evolution

• Intelligence: from reactive behavior to prediction, from single-task optimization to multi-task coordination, from programmed to learning systems
• Physical capability: better perception, finer manipulation and mobility across terrain
• Environment: from structured cells to dynamic spaces shared with people

Autonomy increases in steps:

1. Assisted: a human controls, the robot assists
2. Collaborative: shared control and decisions
3. Supervised: the robot operates under human oversight
4. Autonomous: independent operation in defined scenarios
5. Fully autonomous: independent operation across scenarios

Industry applications

• Manufacturing: assembly, quality inspection, material handling and predictive maintenance today; reconfigurable lines and one robot across several processes next
• Logistics: warehouse picking and packing, inventory management, autonomous yard and delivery vehicles, last-mile delivery
• Healthcare: hospital logistics, support for patient monitoring, surgical assistance, rehabilitation and elder care
• Security and safety: perimeter patrols, inspection of hazardous sites and support for emergency response

Implementation approach

These phases are maturity steps, not calendar years:

1. Foundation: specific use cases, internal AI and robotics skills, vendor partnerships, and controlled test environments, with proof-of-concept ROI
2. Collaborative integration: cobots and mobile robots integrated with warehouse and manufacturing systems, safety and governance protocols, and successful pilots scaled
3. Polyfunctional deployment: multi-task robots, fleet orchestration across sites and functions, and processes redesigned around human-robot teams

Critical success factors

• Technical: reliable perception and decision models, low-latency connectivity such as private 5G, and fail-safe behavior with emergency protocols
• Organizational: change management, training for human-robot collaboration, process redesign and worker involvement from the start
• Infrastructure: robot-ready floor space, a data platform for robot telemetry and learning, and maintenance and support capacity

Risk management and EU regulation

Operational risks include safety in shared spaces, malfunctions, cyberattacks on connected robots and over-reliance on automation. Strategic risks include workforce disruption, vendor lock-in and fast-moving regulation. The EU rules are now concrete:

• Machinery Regulation (EU) 2023/1230: replaces the Machinery Directive and applies from January 20, 2027; it explicitly covers autonomous mobile machinery and AI-based safety functions
• Cyber Resilience Act: for connected products, reporting obligations apply from September 11, 2026 and the main obligations from December 11, 2027
• EU AI Act: AI used as a safety component of regulated products, such as machinery that needs third-party conformity assessment, is high-risk; under the amended timeline, those obligations apply from August 2, 2028 (see the EU AI Act guide)
• NIS2: medium and large manufacturers of machinery and other equipment are in scope, so robot fleets and their OT networks belong in the risk management program (see the NIS2 guide)
• GDPR: robots with cameras and microphones in workplaces or public spaces process personal data (see the GDPR guide)

Mitigate with documented safety testing and conformity assessment, security by design for robot software and connectivity, phased rollout with human fallback, and ongoing workforce development.

What comes next

• Swarm coordination: fleets of different robots coordinated as one system
• Continuous learning: robots that improve from fleet experience, which raises questions about change control and recertification
• Simulation-first planning: world models and digital twins used to test robot behavior before it reaches the floor
• Human-robot interaction: better reading of human intent and movement; inferring workers’ emotions, however, runs into the AI Act’s ban on emotion recognition in the workplace

Conclusion

From Intelligent Things in 2018 to Polyfunctional Robots in 2025, and now Physical AI and World Models, the direction is consistent: machines that adapt instead of following fixed programs, work alongside people in spaces built for people, and take on more than one task. The business case has shifted from replacing a single repetitive job to raising the utilization and flexibility of the whole operation.

For regulated EU organizations, the next two years add a hard deadline layer: the Machinery Regulation in January 2027, the Cyber Resilience Act through December 2027 and the AI Act’s high-risk rules for machinery in August 2028. Start with pilots in controlled environments, design for human-robot teams, and build safety, cybersecurity and conformity evidence into the program from the first deployment.

Trend names, short quotes and forecasts come from Gartner’s public press releases and are attributed to Gartner. The analysis is independent and not endorsed by Gartner.

Sources

• Gartner Identifies the Top 10 Strategic Technology Trends for 2018 – Gartner
• Gartner Identifies the Top 10 Strategic Technology Trends for 2019 – Gartner
• Gartner Identifies the Top 10 Strategic Technology Trends for 2025 – Gartner
• Gartner Identifies the Top Strategic Technology Trends for 2026 – Gartner
• Gartner Identifies the Top Strategic Technology Trends for 2027 – Gartner
• Regulation (EU) 2023/1230 on machinery – EU-OSHA
• Cyber Resilience Act – European Commission
• AI Act – European Commission

Last reviewed: October 2026.