From Smart Cameras to Autonomous Security Robots: Why Public Safety Is Becoming Mobile

During my time working with Intel's Video & AI Cities and RealSense business units, including exposure to edge compute, IoT, and computer vision, I saw firsthand how cities, enterprises, and public agencies were trying to use technology to improve situational awareness.

Most of that conversation centered on fixed infrastructure: cameras, sensors, video analytics, access control, command centers, connected city platforms. Important work. But inherently limited.

That conversation is evolving.

The next phase of physical security is mobile. Autonomous mobile robots (AMRs), quadruped robots, and drones are beginning to operate as roaming sensor platforms: patrolling, observing, detecting, documenting, and escalating events across physical environments in ways that fixed cameras simply cannot.

The Gap That Fixed Security Cannot Close

For twenty years, physical security has been built on fixed systems. Organizations deployed cameras, access control readers, alarms, intercoms, intrusion sensors, and video management systems. Over time, those systems became more intelligent. Video analytics improved, cloud-managed platforms gained adoption, access control and video became more unified, and AI began helping operators reduce false alarms, search footage faster, and identify events of interest.

That trend is accelerating. Genetec's 2026 State of Physical Security report, based on responses from over 7,300 security professionals worldwide, found that AI ranked alongside access control and video surveillance as a top project priority for 2026. Interest in AI adoption among end users more than doubled compared with the prior year, and more than 70% of respondents reported running unified or integrated security systems.

But there is still a structural gap. Fixed cameras observe from mounted positions. They depend on placement, lighting, field of view, and operator attention. A fixed camera can show what is in front of it, but it cannot move toward an alarm, reposition itself, inspect a fence line, patrol a parking deck at 2 a.m., or act as a visible deterrent in changing locations.

That is the gap mobile security platforms are designed to close.

Three Types of Mobile Security Platform

Security robotics is not a single product category. It is a set of mobile sensing platforms, each suited to different environments and operational needs.

Integration Is the Make-or-Break Issue

Here is where my perspective diverges from most of the marketing in this space.

Security robots cannot scale if every robot provider expects customers to operate a separate, robot-only dashboard. Security teams already manage cameras, access control, visitor management, alarms, dispatch tools, radios, mobile apps, and incident reporting platforms. Adding another disconnected interface creates friction, not value.

This is already happening.

The lesson is clear: the market is not just moving toward robots. It is moving toward robots that integrate.

What Matters Is the Workflow, Not the Hardware

One of the recurring mistakes in the robotics market is treating the robot as the product. In security, the robot is part of the solution. The actual product is a safer, more efficient, better-documented security operation.

A hotel does not need a robot because robots are interesting. A hotel may need better after-hours visibility, safer parking areas, faster incident documentation, and reduced pressure on security staff. A warehouse may need better perimeter monitoring, restricted-area detection, and yard visibility. A city may need more coverage in parking garages, public parks, and pedestrian infrastructure.

Dublin, Ohio offers a concrete example. The city deployed its Knightscope K5 autonomous security robot, nicknamed "DubBot," to the Rock Cress Parking Garage in the summer of 2025 as part of a two-year pilot program. A second robot, "Bridge Walker," patrols Riverside Crossing Park and the Dublin Link pedestrian bridge. The deployment is part of a broader public safety technology strategy that includes drones, body-worn cameras, traffic monitoring tools, and license plate readers.

The important thing about Dublin's approach is that it frames the robot as one component of a broader safety infrastructure, not as a standalone technology showcase.

Where Mobile Security Platforms Make Sense First

Not every environment is ready for security robots or drones. The strongest early markets are places with repetitive patrol needs, defined routes, labor constraints, visibility gaps, high-value assets, or measurable incident costs.

The pattern is consistent: the best early use cases combine physical scale, operational repetition, and measurable security gaps that mobile platforms can address more consistently than manual patrol alone.

The Edge AI Layer: Why Silicon Matters to the Stack

The article so far has focused on the robots, drones, VMS platforms, and integrators that make up the visible layers of the mobile security stack. But underneath all of it is a layer that is easy to overlook and hard to replace: the edge AI silicon that gives these machines the ability to perceive, process, and act.

Every security camera running on-device analytics, every AMR navigating a parking garage, every drone identifying an intrusion. They all depend on purpose-built AI processors capable of running deep neural networks, vision-language models, and sensor fusion at low power, in small form factors, without constant cloud connectivity.

Several semiconductor companies are shaping this layer of the stack.

NVIDIA leads the market through its Jetson platform, which powers a wide range of autonomous robots, drones, smart cameras, and industrial edge devices. The Jetson AGX Orin delivers up to 275 TOPS of AI performance and has become a common compute platform for AMR developers who need high-throughput perception and path planning. In early 2025, NVIDIA introduced the Jetson Thor platform, its most advanced edge AI module to date, targeting robotics, autonomous machines, and industrial edge computing with over 800 TOPS of AI performance.

Qualcomm brings a different strength: integrated 5G connectivity alongside edge AI processing. The Robotics RB5 platform delivers 15 TOPS through Qualcomm's AI Engine while supporting up to seven concurrent camera inputs and low-latency 5G communication. At CES 2026, Qualcomm launched the Dragonwing IQ10, positioning itself as the power-efficient alternative for machines that need always-on connectivity without the compute overhead of a full GPU stack.

Ambarella has built a strong position in edge AI vision processing, with over 400 million SoCs shipped worldwide, including more than 42 million edge AI SoCs deployed in production devices. The N1-655 edge GenAI SoC is designed for on-premise AI boxes, AMRs, and smart-city security applications, decoding 12 simultaneous video streams at under 20 watts. Ambarella's partnership with e-con Systems also produced a Robotics Computing Platform based on the CV72S SoC, giving OEMs a developer-ready path into the AMR and warehouse robotics market.

NXP Semiconductors focuses on reliability, security, and real-time control at the edge. In March 2026, NXP announced robotics solutions developed in collaboration with NVIDIA, integrating NVIDIA Holoscan Sensor Bridge with NXP's SoCs to enable sensor fusion, machine vision, and precision motor control for next-generation robots. NXP's edge processors, combined with its automotive-grade networking and functional safety expertise, position it as the foundation for robot "body" systems that connect perception to physical action.

The broader point is this: the mobile security stack does not run on software alone. It runs on purpose-built silicon from companies like NVIDIA, Qualcomm, Ambarella, and NXP. At ISC West 2026, multiple silicon vendors showcased edge AI platforms targeting security and robotics, reinforcing one of this article's core arguments: the mobile security stack is not just about robots. It is about the full technology ecosystem, from chip to SOC operator.

The Trust Question

Any deployment of AI-powered robots, drones, and cameras in physical environments where people live, work, shop, and gather must address trust directly.

People will want to know what is being recorded, whether facial recognition is in use, who has access to the data, how long footage is stored, and whether the system is there to help or surveil. The Dublin deployment already prompted public discussion about privacy, data retention, and citizen rights, which is exactly the kind of conversation that should happen before and during any rollout.

For cities, campuses, hospitality properties, and other public-facing environments, transparency is not optional. It is part of the deployment model. Clear communication about purpose, data collection, retention policies, AI boundaries, human oversight, and escalation procedures builds the trust that sustains adoption. The best deployments will position robots and drones as visible, accountable, human-supervised tools, not mysterious surveillance devices.

The Winning Model Is Ecosystem-Led

The security robotics segment will not scale as a collection of isolated robot products. It will scale when robots, drones, cameras, access control, VMS platforms, AI analytics, edge silicon providers, integrators, and human operators work together as a coherent system.

The partnerships that matter span every layer of the stack: VMS platforms like Genetec and Milestone, access control providers like Brivo and ASSA ABLOY, SOC and GSOC software, managed security service providers, facility management systems, telecom and connectivity providers, cloud platforms, edge AI semiconductor companies like NVIDIA, Qualcomm, Ambarella, and NXP, robot manufacturers like Cobalt, Knightscope, and SMP Robotics, drone providers like Skydio, and local integrators and service organizations like RobotLAB.

From a business standpoint, this matters because customers do not want science projects. They want fewer blind spots, faster verification, fewer false alarms, better documentation, safer staff and guests, predictable operating costs, and tools that fit into the systems they already use.

Security robots and drones can deliver real value. But only when deployed as part of a complete, integrated security stack, not as standalone technology.

The Bottom Line

The future of physical security will be layered. Human guards, fixed cameras, access control, alarms, AI analytics, VMS platforms, command centers, edge AI processors, autonomous mobile robots, quadrupeds, and drones will each play a role.

The question is not whether robots will replace people. The better question is where robots and drones can extend people, strengthen visibility, improve response, and create better security workflows.

For cities, hotels, warehouses, campuses, logistics sites, and critical infrastructure operators, the mobile security layer can add meaningful value when it is tied to real operational needs and integrated into real security ecosystems. For robot providers, silicon companies, and integrators, the path to scale depends on those ecosystem partnerships.

The companies that win in this market will not simply place robots in the field. They will make mobile security platforms useful, integrated, trusted, and operationally relevant.