Connected Operations Research 2026

Manufacturing connectivity has progressed through several distinct generations. The first generation was largely point solutions — individual machines or production lines connected to proprietary monitoring systems that delivered data to operations teams in siloed, vendor-specific formats. The second generation introduced manufacturing execution systems (MES) and historian platforms that aggregated data across a facility, but the integration costs were high and the architectures were brittle. The current generation is characterized by a move toward unified data fabric architectures, where edge computing, industrial protocols, and cloud platforms converge to create a more flexible and scalable connectivity foundation. The transition between these generations is not clean — most manufacturing environments are operating across all three simultaneously, which is precisely what makes the brownfield challenge so persistent.

The industrial IoT market encompasses an extraordinarily diverse range of deployment contexts. A pharmaceutical manufacturer operating highly regulated batch processes has fundamentally different connectivity requirements than a heavy equipment fabricator managing high-mix, low-volume production. An automotive tier-one supplier with multiple global facilities has different governance requirements than a regional food processor with a single production site. This diversity is a consistent source of confusion in connected operations strategy discussions, where technology vendors naturally present their solutions as broadly applicable while the operational reality is highly context-specific. Effective connected operations strategy begins with a clear-eyed assessment of the specific operational environment, not with technology selection.

Technology maturity across the connected operations stack varies considerably. Industrial sensors and data acquisition hardware are mature, commoditized categories where the primary challenge is integration rather than capability. Edge computing platforms have matured significantly, with established vendors offering purpose-built industrial hardware and software stacks that are substantially more reliable and manageable than early edge deployments. Industrial protocol standards — particularly OPC-UA as a vendor-neutral information model and communication standard — have achieved broad enough adoption to meaningfully simplify integration work compared to the proprietary protocol landscape of five years ago. Cloud-based manufacturing analytics platforms have moved from experimental to production-viable. The least mature elements of the stack are organizational: the data governance practices, cross-functional operating models, and OT security programs that determine whether connectivity infrastructure actually delivers business value.

Enterprise adoption context matters for calibrating expectations. Leading manufacturers — organizations that have invested strategically and consistently in connected operations over multiple years — are operating at a qualitatively different level than the median. They have resolved the foundational data architecture questions, established operational technology governance, and are focused on applying connected operations data to increasingly sophisticated use cases: predictive quality, supply chain synchronization, energy optimization. The majority of manufacturers are in an earlier stage, working through the connectivity infrastructure, data standardization, and organizational change challenges that precede advanced analytics. This bifurcation means that benchmarking against leading-edge deployments can create misleading expectations — the practical question for most organizations is how to move efficiently through the foundational stages, not how to replicate the most advanced use cases.

The most consequential architectural trend in industrial connectivity is the emergence of the unified namespace (UNS) as an organizing principle for manufacturing data architecture. Rather than the point-to-point integration patterns that characterize legacy manufacturing IT environments — where each system has direct connections to each other system it needs to exchange data with — the UNS approach establishes a central data broker through which all systems publish and subscribe. The MQTT protocol is the most common implementation vehicle, with platforms like HiveMQ, Sparkplug B, and EMQX providing the infrastructure layer. The architectural benefit is significant: new systems can be connected without rewiring the existing integration fabric, and data becomes a shared resource rather than a siloed asset. The organizational challenge is equally significant: the UNS approach requires agreement on data models, naming conventions, and governance structures that cut across engineering, operations, and IT domains.

Edge computing architecture in manufacturing is converging around a tiered model that reflects the different latency, bandwidth, and reliability requirements at different points in the production environment. At the machine level, embedded computing handles real-time control and local data buffering. At the cell or line level, edge servers aggregate data from multiple machines, perform initial processing and filtering, run time-sensitive analytics, and manage protocol translation. At the facility level, more capable computing handles complex analytics, integration with enterprise systems, and data persistence. This tiered architecture is not a single vendor decision — it involves hardware from industrial computing vendors, operating system and containerization platforms, edge management software, and connectivity middleware. Managing this stack across a large facility or multiple facilities requires meaningful operational capability that organizations often underestimate in initial deployment planning.

The industrial wireless landscape is being reshaped by private 5G deployments, but the transition is more nuanced than technology vendor narratives suggest. Wi-Fi 6 and 6E continue to be the practical choice for most manufacturing wireless applications, offering adequate performance at lower deployment cost and organizational complexity. Private 5G is compelling for specific high-value use cases: deterministic low-latency applications like closed-loop robotics control, high-mobility applications like AGV fleets covering large facility footprints, and high-bandwidth applications like 4K visual inspection cameras. The organizations most aggressively deploying private 5G are large facilities where the use case density justifies the infrastructure investment and the in-house radio frequency engineering expertise. For smaller facilities or organizations without that expertise, managed private LTE is often a more practical near-term path.

Workforce mobility technology is advancing on multiple fronts simultaneously. AR-assisted maintenance guidance has moved from proof-of-concept to production deployment in a number of industries, with the most mature use cases in complex equipment maintenance where the cost of errors is high and the knowledge transfer problem is acute. Digital work instructions delivered through tablet or wearable interfaces are seeing broad adoption as a replacement for paper-based processes, with benefits extending beyond efficiency to include compliance documentation, version control, and the ability to capture as-built versus as-designed deviations. Voice-directed workflows, which have been mature in warehouse operations for years, are finding new application in manufacturing environments where hands-free operation is a safety or ergonomic requirement. The integration of these workforce mobility tools with connected equipment data — surfacing relevant machine state information to the worker at the point of task execution — is where significant unrealized value remains.

“We spent the first two years of our IIoT program solving the wrong problem. We were focused on getting data out of machines when the real challenge was agreeing on what the data meant and who was responsible for acting on it. The technical connectivity work was six months; the data governance and organizational change work is still ongoing.”

The business case for connected operations infrastructure is most credibly framed around three categories of operational improvement. The first is asset performance: connected equipment generates the data required for condition-based and predictive maintenance approaches that reduce unplanned downtime, extend asset life, and allow maintenance resources to be applied where they are needed rather than on fixed schedules. The second is quality management: real-time process data enables faster detection and containment of quality deviations, reducing scrap and rework costs and supporting the root cause analysis that prevents recurrence. The third is operational efficiency: visibility into actual production throughput, cycle times, and constraint behavior enables operations teams to make better scheduling and prioritization decisions. Each of these categories has a different investment return profile and time-to-value horizon, which matters for investment sequencing decisions.

Asset performance improvement from connected operations is perhaps the most straightforward to quantify but the most frequently overstated in vendor-produced business cases. The realistic expectation depends heavily on the current maintenance maturity baseline, the criticality and failure modes of the equipment in scope, and the quality of the analytics applied to the connected data. Organizations that are transitioning from purely reactive maintenance — where equipment is repaired after failure — to condition-based approaches can achieve substantial reductions in unplanned downtime. Organizations already operating mature preventive maintenance programs will see more modest incremental gains. The implementation path matters as much as the technology: connected operations data only delivers maintenance value when it is integrated into the maintenance workflow, which requires changes to CMMS systems, maintenance planning processes, and technician work practices.

The workforce amplification dimension of connected operations business impact is receiving increasing strategic attention as manufacturers face skilled workforce constraints. Connected operations technologies — AR-guided maintenance, digital work instructions, remote expert assistance — enable less experienced workers to perform tasks that previously required years of specialized knowledge. This is not a replacement for workforce development; it is a means of accelerating capability deployment and protecting organizational knowledge from the attrition that comes with workforce demographic transitions. Based on Halkwinds' work across manufacturing organizations, the most effective workforce mobility programs are designed with direct input from frontline workers and operations supervisors — technology solutions designed without that input frequently fail to achieve adoption targets regardless of technical sophistication.

Multi-site operational visibility creates a category of business impact that is not available to organizations operating each facility in isolation: cross-facility performance benchmarking and best-practice transfer. When the same operational metrics are being captured consistently across facilities using standardized data models, it becomes possible to identify performance differentials, investigate their root causes, and systematically transfer effective practices. This capability is particularly valuable for manufacturers operating global facility networks where performance variation is significant but the mechanisms of that variation are poorly understood. The organizational challenge is that benchmarking creates internal accountability dynamics that require careful change management — facilities that perform poorly on benchmarked metrics need a supportive improvement pathway, not just exposure to the performance gap.

• Asset performance ROI depends critically on the current maintenance maturity baseline — organizations transitioning from reactive to condition-based maintenance will see different impact than those already operating mature preventive maintenance programs.
• Quality management use cases often deliver faster time-to-value than maintenance applications because the data-to-action pathway is shorter and the operational workflow changes are more contained.
• Workforce mobility solutions deliver maximum value when designed with direct input from frontline workers — technology-first approaches without operational co-design consistently underperform adoption expectations.
• Multi-site benchmarking requires standardized data models and consistent metric definitions across facilities before comparative analysis becomes meaningful — this data harmonization work is often underestimated.
• The business case for connected operations infrastructure should be structured as a portfolio of use cases with different time horizons, not a single monolithic ROI calculation.
• Connected operations data has significant strategic value for supply chain coordination and customer commitments — organizations that treat this data as an internal operational resource rather than a strategic asset leave value unrealized.
• Energy optimization, enabled by granular connected operations data, is emerging as a significant value driver as energy costs and carbon management commitments become more prominent in manufacturing cost structures.

The brownfield connectivity challenge deserves direct analytical treatment because it defines the implementation context for the vast majority of manufacturing connected operations programs. Brownfield environments contain equipment from multiple generations, multiple vendors, and multiple protocol families. A single production line may include controllers communicating via PROFIBUS, sensors with 4-20mA analog outputs, newer equipment exposing OPC-UA interfaces, and legacy systems with proprietary serial communication. The practical approaches to connecting this heterogeneous environment fall into three categories: native protocol integration where modern connectivity tools support direct connection to legacy protocols; hardware retrofitting using edge devices or intelligent sensors that extract data from legacy equipment and translate it to modern protocols; and asset replacement where the cost of connectivity retrofitting exceeds the cost of equipment replacement with modern connected alternatives. The choice among these approaches requires honest assessment of equipment age, remaining useful life, and the operational disruption tolerance of the production environment.

Edge computing architecture decisions have long-term implications that are not always apparent at the start of a connected operations program. Organizations that deploy edge infrastructure without a clear architecture vision frequently arrive at an unmanageable diversity of edge hardware, operating systems, and software stacks that creates significant operational overhead. The trend toward containerization — running edge applications in Docker or Kubernetes-managed containers on standardized hardware — provides a more manageable path, but requires investment in edge infrastructure management capability. The selection of an edge platform should be evaluated against criteria including: operational management tooling, support for the industrial protocols relevant to the environment, integration with the target cloud or enterprise platform, security management capabilities, and the vendor's credibility for long-term support in industrial environments. Low-cost edge hardware that lacks a credible long-term support path creates significant operational risk in production environments.

Data governance for connected operations is the organizational capability most consistently underinvested relative to its impact on program outcomes. Connected equipment generates data at volumes and velocities that overwhelm organizations without clear policies for data retention, quality management, access control, and lifecycle management. The more fundamental governance challenge is semantic: ensuring that data generated across different systems, facilities, and equipment types means the same thing. A 'downtime event' recorded in a machine monitoring system may be defined differently than the 'downtime' recorded in an MES or production reporting system. Without explicit data model governance — defining exactly what each metric means, how it is calculated, and who is responsible for its accuracy — connected operations data becomes a source of organizational conflict rather than operational clarity.

Security architecture for connected operations must be designed from the beginning, not retrofitted after deployment. The OT/IT integration that enables connected operations creates network pathways between systems with very different security postures. Enterprise IT networks are designed with the assumption of connectivity and are managed with mature security practices. OT networks historically operated in isolation and contain equipment that cannot accept software patches, cannot participate in modern authentication protocols, and may exhibit undefined behavior if subjected to network security scanning. The practical architecture response involves network segmentation using industrial demilitarized zone (IDMZ) patterns that control traffic flow between IT and OT environments, unidirectional security gateways where appropriate, and application-layer data diodes for the highest-criticality OT systems. These architectural measures do not eliminate risk but they contain the blast radius of security incidents and provide the visibility required for effective security monitoring.

• Brownfield integration strategy must be determined equipment by equipment — there is no universal approach, and the right choice depends on equipment age, remaining useful life, protocol support, and production disruption tolerance.
• Edge platform selection should weight operational manageability and long-term vendor support credibility as heavily as technical capability — unmanageable edge infrastructure is a common source of program failure.
• Data governance investment should be front-loaded, not deferred — the cost of retroactively standardizing data definitions across a connected operations deployment is substantially higher than establishing governance at the outset.
• Industrial DMZ architecture and network segmentation are non-negotiable prerequisites for OT/IT integration — flat network architectures that connect OT and IT without segmentation create unacceptable risk.
• Industrial protocol expertise — OPC-UA information modeling, MQTT Sparkplug B configuration, legacy protocol translation — is a specialized capability that many organizations need to acquire through hiring, training, or partnership rather than assuming their existing IT staff can absorb.
• Program governance should include explicit ownership of the connectivity infrastructure layer, which often falls into an ambiguous space between IT, engineering, and operations — unowned infrastructure is poorly maintained infrastructure.

The organizational barriers to connected operations programs are at least as consequential as the technical barriers, yet they receive less attention in planning and investment discussions. The most common organizational failure mode is the absence of sustained executive sponsorship combined with the presence of organizational incentives that work against cross-functional collaboration. Connected operations programs require sustained cooperation between IT, OT engineering, operations, and maintenance functions that often have different priorities, different reporting structures, and different definitions of success. Programs that launch with strong executive visibility but lose that sponsorship during the transition from pilot to production scale frequently stall at a stage where significant infrastructure investment has been made but operational value has not yet been systematically captured.

The cybersecurity risk profile of connected operations is distinctive and requires specific executive attention. Legacy industrial control systems were designed for reliability and safety, not security — many contain firmware that cannot be updated, communication protocols with no authentication mechanisms, and human-machine interfaces running end-of-life operating systems. Connecting these systems to enterprise networks without appropriate security architecture creates exposure to threat actors who have demonstrated both the capability and the motivation to target industrial environments. Ransomware attacks on manufacturing operations have resulted in production shutdowns lasting days to weeks, with financial impacts that substantially exceed the cost of the security investments that would have prevented or contained them. The organizational challenge is that OT security investment is often evaluated against a pre-incident risk posture rather than the post-connection risk posture that connected operations programs create.

Data quality is a persistent and underappreciated risk in connected operations programs. Connected equipment data is only as useful as it is accurate and complete, and the conditions that affect data quality — sensor drift, network packet loss, timestamp synchronization errors, protocol translation artifacts — are endemic in industrial environments. Organizations that build analytics applications and operational workflows on top of connected data without rigorous data quality monitoring create operational dependencies on unreliable foundations. The consequence is not typically a catastrophic failure but a gradual erosion of trust in connected operations data as operations teams observe discrepancies between sensor data and physical reality. Once trust in connected data is lost, it is difficult to rebuild — and the operational value of the program diminishes proportionally. Data quality investment, including automated anomaly detection and data lineage tracking, is not optional for production-scale connected operations programs.

Vendor concentration and platform lock-in represent strategic risks that manufacturing executives should explicitly evaluate in connected operations program architecture decisions. The industrial IoT platform market includes both established automation vendors with deep OT integration expertise and newer cloud-native platform vendors with more advanced data and analytics capabilities. Organizations that commit deeply to a single vendor's connected operations platform may find themselves with limited negotiating leverage, constrained to a technology roadmap not of their choosing, and facing significant migration costs if the vendor's strategic direction diverges from their operational needs. Architecture decisions that preserve data portability, favor open protocols, and maintain the ability to substitute components of the connectivity stack provide meaningful strategic protection against vendor concentration risk.

• Organizational alignment across IT, OT engineering, operations, and maintenance is the most common failure mode in connected operations programs — governance structures must be designed before technology deployment begins.
• OT cybersecurity risk increases at the moment of IT/OT network integration — the risk posture of connected operations is fundamentally different from isolated OT environments, requiring a security architecture review before connectivity is expanded.
• Data quality management is a production-critical capability, not an analytics optimization — connected operations programs that do not invest in data quality monitoring will experience trust erosion that undermines operational value.
• Vendor concentration risk in industrial IoT platforms is real and should be explicitly evaluated — open protocol adoption and data portability provisions in vendor contracts are meaningful risk mitigation measures.
• The transition from pilot to production scale is the most commonly underestimated phase of connected operations programs — operational support models, incident response procedures, and change management processes required for production systems are substantially more demanding than pilot environments.
• Skills gaps in industrial protocol expertise, OT security, and edge infrastructure management are structural constraints that program timelines must account for realistically.

The near-term priority for most manufacturing organizations should be establishing the foundational data infrastructure that makes advanced connected operations use cases achievable. This means resolving the brownfield connectivity challenge for priority equipment through a combination of protocol adapters, edge hardware retrofits, and targeted asset replacement. It means establishing the data architecture — whether unified namespace or another structural approach — that will make new use cases addable without architectural rework. It means investing in OT security architecture before expanding IT/OT network integration, not as a parallel track but as a precondition. And it means establishing the data governance structures — semantic data models, data quality standards, ownership definitions — that will prevent the data quality and trust erosion problems that plague programs built on unstable data foundations. This foundational work is less exciting than advanced analytics use cases, but it determines whether those use cases are achievable at all.

The medium-term roadmap for connected operations should be structured around use case portfolios that build capability sequentially. The most effective sequencing pattern observed across successful programs moves from equipment monitoring (connectivity and data capture) to condition-based maintenance (applying connected data to maintenance workflows) to cross-facility visibility (extending standardized data capture and analytics across multiple sites) to predictive and prescriptive applications (applying machine learning to connected data for proactive operational guidance). Each stage builds on the data infrastructure, organizational capability, and operational trust established in the previous stage. Programs that attempt to skip stages — deploying predictive analytics on top of inconsistent connectivity infrastructure, for example — consistently encounter the data quality and organizational change problems that the skipped stages were designed to address.

For organizations evaluating private 5G investments, the strategic recommendation is to ground the investment decision in specific, quantified use cases rather than general capability arguments. The most compelling near-term use cases — AGV fleet coordination in facilities with large footprints, high-bandwidth visual inspection for quality applications, mobile robotics in environments where cable management is a safety or flexibility constraint — share characteristics that should guide evaluation: deterministic latency requirements that Wi-Fi cannot reliably meet, high device density in localized areas, or mobility patterns that require seamless handoffs across large coverage areas. Organizations without use cases that clearly meet these criteria should continue with Wi-Fi 6 infrastructure and revisit the private 5G investment case as use cases mature.

The long-term strategic opportunity in connected operations extends beyond operational efficiency to market differentiation and new business model enablement. Manufacturers that have built robust connected operations infrastructure are increasingly able to offer product-as-a-service and outcome-based commercial models — enabled by the equipment performance data and remote monitoring capabilities that connected operations programs generate. This shift from transactional product sales to ongoing service relationships creates revenue streams with different characteristics than traditional manufacturing margins and can represent a meaningful competitive differentiator in markets where product commoditization is advanced. Organizations that scope connected operations programs purely as cost reduction initiatives may be underestimating the strategic optionality that connectivity infrastructure creates.

The trajectory of connected operations technology over the next several years will be shaped by several converging forces. AI inference at the edge — running machine learning models on edge hardware rather than sending data to cloud platforms for analysis — is advancing rapidly, driven by both the development of purpose-built AI inference hardware and the improvement of model compression techniques that make sophisticated models deployable on constrained devices. This development has significant implications for connected operations: latency-sensitive applications that currently require cloud roundtrips will become executable at the edge, enabling closed-loop control applications that are not feasible with current architectures. The practical constraint is model development and maintenance capability, which requires data science expertise that most manufacturing organizations are still building.

The convergence of physical and digital environments in manufacturing — often framed as digital twin technology — is maturing from visualization tools to operational systems that actively inform production decisions. The most advanced implementations maintain real-time synchronized models of production assets that are used for process simulation, predictive quality modeling, and maintenance planning. The foundational requirement for digital twin applications is precisely the connected operations infrastructure this report examines: high-quality, standardized, real-time data from production assets. Organizations that are investing in connected operations infrastructure today are building the data foundation on which digital twin applications will run. The organizations that lack that foundation will find the path to operational digital twins much longer than vendor presentations suggest.

The regulatory and customer expectation environment for manufacturing transparency is evolving in ways that create additional strategic urgency around connected operations capability. Supply chain transparency requirements — driven by customer expectations, regulatory requirements, and ESG reporting obligations — increasingly require manufacturers to demonstrate, not simply assert, operational performance claims. Quality certifications, environmental compliance, labor practice standards, and carbon accounting all benefit from or will increasingly require connected operations data as evidence. Organizations that have invested in connected operations infrastructure will find compliance with these evolving requirements substantially less burdensome than organizations that must build that infrastructure reactively. The regulatory direction is clear even where specific requirements remain unsettled — connected operations data will increasingly be required for competitive participation in demanding markets.

Halkwinds is an enterprise technology research and advisory firm focused on the intersection of operational technology, industrial digitalization, and enterprise software. Halkwinds Research produces analysis of the platforms, architectures, and implementation patterns that drive operational outcomes across manufacturing, logistics, and asset-intensive industries. Our advisory work spans connected operations strategy, OT/IT integration architecture, industrial data governance, and the organizational capabilities required to extract durable value from manufacturing technology investments. Halkwinds combines practitioner expertise with rigorous analytical methodology, producing research that is grounded in the operational realities of industrial environments rather than the technology narratives of platform vendors.

Halkwinds' work in connected operations draws on direct engagement with manufacturing organizations across automotive, aerospace, food and beverage, pharmaceuticals, and industrial equipment sectors. Our analysts have direct experience with the brownfield connectivity challenges, OT security program development, and operational change management that determine whether connected operations programs succeed or stall. The Halkwinds Research Hub publishes findings from this work to support manufacturing executives and technology leaders in making well-informed investment decisions in a market characterized by rapid technology evolution, aggressive vendor marketing, and a significant gap between technology capability claims and operational deployment realities.