What the shift from fixed hardware to portable intelligence means for your organization
Most discussions around AI focus on models. Most discussions around IoT focus on devices. Both miss the more consequential shift happening beneath both industries: the decoupling of supervisory software from proprietary hardware ecosystems.
Across industrial automation, edge computing, and distributed AI systems, organizations are deploying vendor-neutral orchestration layers that coordinate workloads across heterogeneous infrastructure — not because the hardware changed, but because the coordination logic no longer has to be bound to it. The result is software-defined operational infrastructure: architectures where deployment logic, analytics, and control are more portable than the physical systems they govern.
This shift is most visible where distributed autonomy must operate under real-world constraints — cybersecurity requirements, regulatory frameworks, interoperability limits, latency boundaries, and the physical realities of industrial environments. These aren’t edge cases. They are the environment.
The real infrastructure shift isn’t in the hardware, it’s in the coordination layer above it. When supervisory software decouples from proprietary ecosystems, the entire operational stack becomes composable.”
Simultaneously, AI research is moving beyond isolated models toward networked ecosystems of cooperating agents. Rather than a single model responding to prompts, these architectures involve distributed populations of specialized agents capable of delegation, negotiation, memory exchange, tool use, and coordinated execution across cloud and edge environments. The MIT Media Lab’s NANDA framework sits at the intersection of this work – treating agents not as assistants but as interoperable operational services that can discover, coordinate with, and supervise other agents across distributed infrastructure layers. The framework can be defined as an “Internet of AI Agents” -this is more than a metaphor. It describes a topology.
This is the moment of convergence for IoT. Traditional architectures depend on tightly coupled device logic, vendor-specific integrations, and centralized supervisory platforms. Agent-oriented orchestration replaces that rigidity with adaptive supervisory layers that coordinate heterogeneous devices, data streams, control systems, and edge workloads through higher-level semantic and operational abstractions. The hardware doesn’t have to change. The coordination model does.
The Shift From Fixed Hardware to Portable Intelligence
For decades, industrial and operational systems were built around tightly integrated hardware stacks. PLCs (Programmable Logic Controllers), SCADA (Supervisory Control and Data Acquisition) systems, embedded controllers, and industrial gateways were often deeply tied to specific vendors and deployment models. Expanding or modernizing those environments typically required significant infrastructure replacement and operational disruption.
That model is evolving.
Modern edge platforms are moving toward software-defined operations where orchestration, supervision, and intelligence exist independently from the underlying hardware layer. Applications are increasingly containerized. Workloads are portable. Infrastructure is abstracted. Supervision is decoupled from hardware.
This creates operational flexibility that most older architectures were never designed to support.
An intelligent workload that once depended on a specific physical appliance can now move between hardware environments with minimal reconfiguration. AI inference can execute locally at the edge rather than relying exclusively on centralized cloud infrastructure. Operational systems can scale horizontally across fleets of devices rather than vertically through increasingly expensive proprietary infrastructure.
In industrial environments, this transition is already being operationalized through virtual PLCs, edge-native SCADA, and software-defined automation frameworks. In AI infrastructure, the same architectural logic is driving the shift toward distributed agentic systems — where intelligence, like workloads, is becoming something that can be coordinated across environments rather than anchored to any single one.s virtual PLCs, edge-native SCADA, and software-defined automation.
The Convergence Between Industrial Edge and Agentic AI
Industrial automation and agentic AI appear to belong to separate categories. They are beginning to solve the same problems.
Modern agentic systems operate as distributed execution environments, not standalone applications. Multiple agents coordinate asynchronously across systems and contexts. Some workloads execute locally. Others route through centralized orchestration. Human approval gates, telemetry, policy enforcement, audit trails, and workload supervision are operational necessities – not optional features.
This is industrial infrastructure logic applied to AI.
The challenge is no longer generating outputs from a model. It’s supervising a distributed network of intelligent processes across heterogeneous environments while maintaining reliability, governance, and operational traceability. That is precisely the problem edge orchestration platforms were built to address – managing software updates, security policies, telemetry, workload deployment, fault monitoring, and rollback capability across thousands of distributed nodes, regardless of underlying hardware vendor or device architecture.
Agentic AI systems are encountering the same operational realities. The nodes are no longer just physical machines – they are inference workloads, autonomous agents, localized automation systems, and policy-bound execution environments.
Industrial edge infrastructure is becoming software-native. Agentic AI is becoming infrastructure-native. The gap between them is closing faster than either community has recognized.
Why Hardware Abstraction Matters to the C-Suite
One of the largest operational challenges in both IoT and distributed AI is fragmentation — and it’s almost never a deliberate choice. Infrastructure accumulates over time. Vendors change. Acquisitions happen. Deployment environments diverge. The result is operational complexity that compounds as organizations scale, and that complexity has a cost: slower deployment cycles, higher maintenance overhead, and technology lock-in that constrains future investment decisions.
Hardware abstraction addresses this at the architectural level. Rather than building operational logic around specific hardware platforms, organizations manage workloads through software layers capable of deploying and supervising across heterogeneous device environments simultaneously. The infrastructure beneath becomes largely irrelevant to the operational layer above it.
For executive decision-makers, this translates into four concrete advantages.
Capital flexibility. Hardware strategies can evolve without rewriting operational systems – reducing the switching costs that typically lock organizations into legacy vendor relationships.
Deployment speed. New edge infrastructure can be provisioned remotely through zero-touch deployment models, compressing rollout timelines and reducing dependence on field engineering resources.
Operational resilience. Workloads are no longer anchored to specific hardware. When devices fail or conditions change, execution shifts – automatically and without manual intervention.
AI at operational scale. Inference can run locally where latency, bandwidth, privacy, or regulatory constraints make centralized cloud execution impractical – unlocking AI deployment in environments where it previously couldn’t operate.
These are not just IT considerations that surface in infrastructure reviews. They are strategic constraints on how quickly organizations can move, how much they pay to maintain what they’ve built, and how much leverage they carry into vendor negotiations.
The Companies Driving the Shift
The market for edge orchestration and distributed operational infrastructure is still forming, and the vendors shaping it are approaching it from distinctly different angles — reflecting the diversity of the environments they were built to serve.
Companies such as ZEDEDA and SUSE Industrial Edge anchor the enterprise end of the market, with platforms built around Kubernetes-native deployment, large-scale fleet supervision, and hardware-agnostic lifecycle management. Their architectures are designed for organizations operating thousands of distributed nodes across complex, multi-vendor environments.
A different set of vendors — including Barbara and Mutexer — are focused on industrial modernization from the operational technology side. Their work centers on OT/IT convergence, software-defined automation, and reducing dependency on tightly coupled legacy hardware. Where the first group abstracts infrastructure for cloud-native operators, this group is meeting industrial environments where they actually are.
Platforms such as Clea by SECO and FairCom Edge occupy a more embedded tier — emphasizing telemetry, OTA lifecycle management, and lightweight edge AI deployment for constrained hardware environments where full Kubernetes orchestration isn’t practical.
Open-source ecosystems are also emerging as a significant force. Projects including KubeEdge, Open Horizon, and EdgeX Foundry are increasingly attractive to organizations prioritizing vendor neutrality, sovereign infrastructure control, or air-gapped deployments — particularly in regulated industries and public sector contexts.
Taken together, these efforts reflect a market converging on a shared architectural premise: that operational intelligence should be portable, distributed, and decoupled from the hardware layer beneath it. The vendors disagree on implementation. They agree on direction.
Leading vendors: hardware-agnostic edge control, orchestration, and supervision software
A few companies consistently emerge as leaders in this space — particularly across industrial automation, IIoT, edge AI, and distributed operations.
Here are some of the current strongest players by category as defined in our research:
| Provider | Core Focus | Strengths | Typical Customers |
|---|---|---|---|
| ZEDEDA | Edge orchestration & lifecycle management | Strong hardware abstraction, zero-touch deployment, Kubernetes/VM support | Industrial, retail, telecom, energy |
| Barbara | Industrial edge AI platform | OT/IT convergence, container orchestration, broad protocol support | Utilities, manufacturing, energy |
| SUSE Industrial Edge | Industrial edge infrastructure | Kubernetes-native, GitOps workflows, scalable fleet ops | Enterprise industrial operations |
| Clea by SECO | Full-stack edge/IoT framework | Hardware-agnostic orchestration, OTA, AI deployment | OEMs, embedded systems vendors |
| Eclipse ioFog | Open-source EdgeOps | Distributed workload orchestration, air-gapped deployments | Defense, industrial, research |
| FLECS | Industrial software layer | Software-defined automation environments | Machine builders, automation OEMs |
| FairCom Edge | Industrial data integration | OT protocol translation, edge persistence, telemetry | Manufacturing, utilities |
| Mutexer | Virtual PLC / SCADA platform | Software-defined controls on generic Linux hardware | Modern industrial automation teams |
How the market is taking shape
The market is converging around Kubernetes-native, containerized edge orchestration — and “hardware-agnostic” has become a term of art with fairly consistent meaning across vendors: ARM and x86 compatibility, support for hardware platforms such as NVIDIA Jetson, Intel, and industrial IPCs, container portability across virtualized environments, and independence from proprietary PLC ecosystems.
The clearest differentiator between platforms is where they sit in the stack. Some — including ZEDEDA and SUSE — focus on IT-style edge orchestration: abstracting heterogeneous hardware and managing large-scale distributed infrastructure. Others, such as Barbara and Mutexer, target industrial OT environments directly, working to replace tightly coupled PLC and SCADA stacks with portable software layers. A third group — including Clea by SECO and FairCom Edge — centers on IoT telemetry, OTA lifecycle management, and lightweight edge AI deployment.
For industrial control specifically, the most consequential trend is the move toward virtual PLCs, software-defined automation, edge-native SCADA, and AI-assisted operations at the edge. This is why vendors like Mutexer and Barbara are drawing attention: they’re not just modernizing the interface to industrial systems — they’re attempting to replace the underlying control architecture entirely.
ZEDEDA occupies a different position: a recognized horizontal platform that abstracts heterogeneous edge hardware and supports distributed management at scale, without being tied to any specific industrial vertical.)
The open-source layer
Open-source ecosystems are playing an increasingly significant role. Projects including Eclipse ioFog, KubeEdge, Open Horizon, and EdgeX Foundry tend to surface in environments where vendor neutrality is a priority, air-gapped deployments are required, or organizations need to avoid cloud lock-in — conditions common in regulated industries, defense-adjacent infrastructure, and public sector deployments.
One way to read the competitive landscape:
| Category | Representative vendors |
|---|---|
| Cloud-native edge infrastructure | ZEDEDA, SUSE, ioFog |
| Industrial automation modernization | Barbara, FLECS, Mutexer |
| Embedded / IoT edge platforms | Clea, FairCom |
| AI-centric edge orchestration | Barbara, Clea, TwinEdge |
The convergence that matters
What makes this moment distinct is that the convergence isn’t just between IT and OT – it’s between physical operational infrastructure and the emerging layer of distributed agentic AI. The platforms being built today to supervise heterogeneous edge hardware are, in many respects, the same platforms that will eventually coordinate heterogeneous agent ecosystems. The infrastructure problem and the AI problem are becoming the same problem.
Velocity Ascent helps organizations think through the infrastructure and AI questions that don’t have obvious answers yet. If that’s where you are, we should talk.
Learn More: Core Concepts — A Plain-English Overview
What Are Industrial Edge Systems?
Industrial edge systems are localized computing environments that sit close to physical operations rather than inside centralized cloud infrastructure.
Rather than routing every sensor reading, command, or signal back to a distant data center, edge systems process information at or near its source. The practical effect is meaningful: lower latency, stronger resilience, and the ability to keep operations running even when cloud connectivity is interrupted or unavailable.
Examples include factory floor automation, utility monitoring infrastructure, logistics and warehouse operations, transportation systems, energy infrastructure, and oil and gas facilities.
These environments typically operate continuously, under strict requirements for reliability, low latency, and operational oversight — conditions that make edge processing not just useful, but necessary.
What Are Agentic AI Systems?
Agentic AI systems are environments where software agents perform tasks autonomously or semi-autonomously on behalf of users or organizations.
Unlike a traditional chatbot that generates a single response, an agentic system can retrieve information, make decisions, coordinate with other agents, trigger workflows, monitor systems, generate outputs, request approvals, and execute operational tasks — often in sequence, and often without direct human involvement at each step.
A mature agentic system behaves less like a standalone application and more like a distributed operational workforce: specialized digital actors operating under defined rules, permissions, and supervisory controls.
What is IoT?
IoT “the Internet of Things” refers to the connection of physical devices to digital networks so they can collect, transmit, receive, and act on data.
The devices themselves span a wide range: environmental sensors, smart meters, connected industrial machinery, surveillance systems, wearables, fleet tracking hardware, and building automation systems, among others.
The core idea is straightforward but consequential: physical infrastructure becomes digitally observable and, increasingly, digitally controllable.
What Is Hardware-Agnostic Edge Control Software?
Hardware-agnostic edge control software is a supervisory layer that manages distributed systems regardless of who manufactured the underlying hardware.
Traditionally, operational systems were tightly bound to proprietary hardware ecosystems with software and hardware sold and maintained together, by the same vendor, on the same roadmap.
Modern orchestration platforms break that coupling. Workloads become portable. Hardware becomes interchangeable. Vendor lock-in becomes a choice rather than a structural constraint.
In practice, this means organizations can deploy workloads across mixed hardware fleets, centrally supervise distributed systems, push software updates remotely, scale without replacing infrastructure, standardize governance and security policies, and run AI workloads across diverse environments – all through a single coordination layer.
In simple terms: software intelligence becomes more portable than the hardware beneath it.
What Are PLCs and SCADA Systems?
PLCs and SCADA systems are two foundational technologies in industrial operations – and increasingly, two of the most important targets for modernization.
A PLC (Programmable Logic Controller) is a rugged industrial computer designed to control machinery and operational processes in environments such as factories, utilities, and infrastructure facilities. A SCADA (Supervisory Control and Data Acquisition) system sits above that layer, providing centralized visibility across an industrial environment: collecting telemetry, displaying operational status, triggering alerts, and allowing operators to monitor or intervene across distributed systems.
Historically, both were highly proprietary – hardware and software sold together, deeply coupled to specific vendor ecosystems, and difficult to update or extend without significant infrastructure investment.
Modern edge orchestration platforms are beginning to change that, virtualizing and modernizing these environments through software-defined approaches that decouple operational logic from the hardware beneath it.
What Is NANDA?
NANDA – the Networked Agents and Decentralized Architecture framework, developed at the MIT Media Lab – is an emerging standard for how AI agents discover, communicate with, and coordinate across distributed infrastructure.
Where most AI frameworks focus on what a single model can do, NANDA focuses on what networks of agents can do together. Agents in a NANDA-aligned system are designed to be interoperable, capable of finding other agents, negotiating tasks, delegating work, and operating across mixed infrastructure without requiring a centralized controller to manage every interaction.
The framework is sometimes described conceptually as an “Internet of AI Agents” – a topology where agents function less like applications and more like addressable services that can be composed, coordinated, and supervised at scale.
For organizations thinking about edge infrastructure and distributed operations, NANDA is worth watching. The same architectural problems it addresses in AI – heterogeneous environments, decentralized coordination, autonomous execution under governance constraints – are the problems industrial edge platforms have been working on for years. The two bodies of work are converging on the same solution space.
