A Review of Digital Twin Integration in Circular Manufacturing for Sustainable Industry Transition

Abstract

The integration of digital twin (DT) technology into circular economy (CE) frameworks has emerged as a critical pathway for achieving sustainable and intelligent manufacturing under the Industry 4.0 paradigm. This study addresses the lack of structured guidance for DT adoption in CE strategies by proposing two interrelated frameworks: the Sustainable Digital Twin Maturity Path (SDT-MP) and the Digital Twin Nexus. The SDT-MP outlines progressive stages of DT deployment—from data acquisition and real-time monitoring to AI-enabled decision-making—aligned with CE principles and Industry 4.0 capabilities. The DT Nexus complements this maturity model by structuring the integration of enabling technologies such as AI, IoT, and edge/cloud computing to support closed-loop control, resource optimization, and predictive analytics. Through a mixed-methods approach combining literature analysis and real-world case validation, this research demonstrates how DTs can facilitate lifecycle intelligence, enhance operational efficiency, and drive sustainable transformation in manufacturing. The proposed frameworks offer a scalable roadmap for intelligent circular systems, addressing implementation challenges while supporting Sustainable Development Goal 9 (Industry, Innovation, and Infrastructure) by promoting digital infrastructure, innovation-driven manufacturing, and environmentally responsible industrial growth. This study contributes to the advancement of digital infrastructure and sustainable circular supply chains in the context of smart, connected industrial ecosystems.

1. Introduction

The manufacturing industry is undergoing a significant transformation driven by the urgent need for sustainability and resource efficiency. In response to growing concerns about climate change, resource depletion, and environmental degradation, industries are shifting from traditional linear production models toward circular economy (CE) principles. This transition demands innovative digital solutions to optimize resource use, minimize waste, and ensure sustainability across product lifecycles. Among these technologies, digital twins (DTs) have gained considerable attention for their ability to create real-time virtual representations of physical assets, systems, and processes. By enabling continuous monitoring, predictive analytics, and AI-driven decision-making, DTs provide a data-rich foundation for optimizing operations, extending product lifespans, and achieving sustainability goals.

Despite their potential, the integration of DTs into CE frameworks remains underdeveloped and lacks sufficient empirical validation in real-world industrial settings. Existing research primarily highlights applications in predictive maintenance, production optimization, and supply chain visibility. However, few studies have addressed their direct contributions to CE strategies such as closed-loop production, waste reduction, and resource recirculation. Furthermore, large-scale DT adoption faces several barriers, including high implementation costs, data security concerns, and interoperability challenges. These limitations highlight the need for a structured approach to guide industries in effectively deploying DTs for sustainable, circular manufacturing. Unlike traditional lifecycle assessment (LCA) methods that offer static evaluations, DT-enabled lifecycle intelligence supports dynamic, real-time feedback that enhances circularity and responsiveness across the manufacturing value chain.

Meng et al. (2023) [1] focus on the construction sector and identify DT applications in design and demolition phases but do not present a lifecycle-wide framework or cross-sector applicability. Preut et al. (2021) [2] emphasize stakeholder-specific information flows in circular supply chains but stop short of proposing structured adoption pathways. Mügge et al. (2024) [3] conduct a systematic review with TRL-based lifecycle classifications but lack an implementation roadmap. Chi et al. (2023) [4] review multiple digital technologies in CE but offer limited depth on DT-specific frameworks. Rocca et al. (2020) [5] present a lab-scale DT-VR use case without broader generalization. Similarly, Pehlken et al. (2024) [6] focus on DTs for eco-design in the automotive sector, while Anwar et al. (2024) [7] and Ali et al. (2025) [8] explore sector-specific DT applications without offering a structured, cross-sectoral maturity model. In contrast, this study introduces the Sustainable Digital Twin Maturity Pathway (SDT-MP) and the DT Nexus Framework, which together support the scalable deployment of DTs across the full lifecycle—from design to monitoring, analytics, and decision-making. The SDT-MP maturity stages map closely to Industry 4.0 advancement levels, from data acquisition via IoT devices to autonomous decision-making supported by AI-driven analytics. This approach incorporates AI-enabled feedback and optimization mechanisms, thereby operationalizing DTs within CE strategies in a way that prior studies have not addressed.

To address this gap, this study proposes the Sustainable Digital Twin Maturity Path (SDT-MP), a systematic roadmap for DT adoption and implementation in sustainable manufacturing ecosystems. Specifically, the study aims to (1) identify the key phases and technological enablers for integrating DTs into CE strategies, (2) develop a structured maturity model to support industries at various stages of DT deployment, (3) evaluate the practical impact of DT-enabled CE strategies on resource efficiency, waste reduction, and lifecycle management.

In this review, the term “digital twin (DT)” refers to both real-time data-driven models and hybrid approaches that integrate physical simulation with sensor feedback. We consider DT applications that support feedback loops, monitoring, prediction, or optimization across the product lifecycle. The review excludes purely theoretical models or digital mockups lacking interaction with physical systems. The term “circular manufacturing” refers to manufacturing systems that aim to maximize resource efficiency and sustainability through circular economy (CE) strategies such as reuse, remanufacturing, recycling, and closed-loop production systems. This review focuses on industrial and manufacturing domains, including sectors such as electronics, automotives, process industries, and precision component manufacturing.

This review adopts a narrative approach to synthesize recent developments in digital twin applications for circular manufacturing. Literature was collected from Scopus, Web of Science, and Google Scholar using combinations of keywords such as “digital twin,” “circular manufacturing,” “Industry 4.0,” “lifecycle,” and “sustainability,” focusing on peer-reviewed publications. Studies were selected based on their relevance to DT integration across lifecycle stages (design, production, end-of-life) and their contribution to sustainability metrics or circular economy strategies.

This study employs a multi-methods approach, combining a comprehensive literature review with industrial case study analysis. This dual-validation strategy bridges theory and practice, offering both conceptual insights and empirical evidence to support the integration of DT technology into sustainable manufacturing systems. The contributions of this research are twofold. First, it introduces a structured DT adoption model that addresses the technological, operational, and economic challenges encountered in transitioning to DT-enabled CE frameworks. Second, it extends the discourse on Industry 4.0 and sustainability by positioning DTs as key enablers of responsible technological transformation. By offering a practical roadmap for digital adoption, this study supports both academia and industry in advancing sustainable, intelligent, and circular production systems. This work directly contributes to Sustainable Development Goal 9 (Industry, Innovation, and Infrastructure) by promoting digital infrastructure and sustainable industrialization. Through the structured integration of AI-driven DT technologies and CE principles, the proposed frameworks foster innovation, technological upgrading, and environmentally responsible industrial transformation.

To address the fragmented nature of DT integration in circular manufacturing, this review proposes two complementary frameworks: the Sustainable Digital Twin Maturity Path (SDT-MP), which outlines progressive stages of DT adoption aligned with circular economy outcomes, and the Digital Twin Nexus, which synthesizes the multi-dimensional nature of DT deployment across lifecycle phases. These frameworks serve as analytical structures to classify and evaluate prior work while guiding future DT-CE alignment. The proposed frameworks—SDT-MP and DT Nexus—not only guide digital twin adoption for circular economy goals but also serve as practical enablers for the Industry 4.0 transition by aligning DT implementation with real-time, intelligent manufacturing paradigms.

The remainder of this paper is structured as follows: Section 2 examines the evolution of manufacturing. It discusses the industrial revolutions and transformations in manufacturing paradigms to highlight the shift toward sustainability and digitalization. Section 3 discusses responsible technological transformation for sustainable and circular manufacturing, covering principles and metrics for sustainable manufacturing, CE strategies in manufacturing, and hyper-personalization, before concluding with an in-depth exploration of responsible technological transformation. Section 4 explores the future of sustainable and intelligent manufacturing, introducing the concepts of connected micro and urban smart factories, followed by the DT Nexus, which examines how DTs facilitate CE integration and sustainable manufacturing practices. Section 5 presents the real-world implementation of DT strategies in sustainable industries, providing case studies and empirical insights into their applications. Section 6 introduces an SDT-MP and discusses real-world implementations of DT to guide industries through structured adoption pathways. Finally, the paper concludes with key insights, practical recommendations, and directions for future research, thereby offering a roadmap for advancing DT-enabled sustainability in modern manufacturing.

2. Evolution of Manufacturing

Manufacturing transformations have been shaped by technological revolutions that increased the efficiency, productivity, and adaptability of global production systems. Each industrial revolution has brought about new paradigms in manufacturing, from mechanized production in the 18th century to the current intelligent, data-driven ecosystems. However, as manufacturing systems evolve, their social, environmental, and ethical implications have also expanded.

Past revolutions primarily emphasized efficiency and scale; however, current digital transformation demands a more responsible approach to technological innovation. I4.0, which introduces unprecedented opportunities for autonomous decision-making and hyperefficient production, also raises pressing concerns about energy consumption, ethical AI deployment, and equitable access to digital manufacturing technologies. AI-driven automation, DTs, and blockchain are redefining manufacturing by optimizing operations, but their unchecked deployment can lead to sustainability challenges, such as increased energy consumption, resource depletion, and ethical concerns in human–AI collaboration. As industries transition toward circular and sustainable manufacturing, a pressing need to balance innovation with responsibility exists, ensuring that technology-driven transformations align with environmental goals, ethical AI practices, and resilient supply chains. The following section examines how manufacturing paradigms have evolved to incorporate sustainability, hyper-personalization, and CE principles.

2.1. The Industrial Revolutions: Transitions in Manufacturing

Manufacturing transformation has been shaped by four major industrial revolutions driven by technological advancements, economic demands, and evolving societal expectations [9]. These revolutions have successively improved efficiency, productivity, and adaptability, leading to the development of intelligent, interconnected, and autonomous manufacturing ecosystems [10].

Table 1. Key economic, societal, and technological drivers of industrial revolutions in manufacturing [9,10,11,12].

Industrial Revolution	Economic Factors	Societal and Environmental Needs	Enabling Technologies
1st Industrial Revolution	Demand for mechanized production, increased labor costs	Rural-to-urban migration, need for higher production efficiency	Steam engines, water power
2nd Industrial Revolution	Expansion of global markets, cost reduction, economies of scale	Mass urbanization, growing demand for consumer goods	Electricity, assembly lines, mechanized tools
3rd Industrial Revolution	Shift towards automation, globalization, demand for high efficiency	Workforce reduction, need for cost-efficient production	IT, robotics, CNC machines, enterprise computing
4th Industrial Revolution (I4.0)	Smart factories, real-time analytics, sustainability concerns, AI-driven efficiency	Sustainability, customization, digital transformation, supply chain resilience	AI, DTs, IoT, cyber–physical systems (CPSs), edge computing

lists the key economic, societal, and technological drivers of each industrial revolution, highlighting their progressive impact on manufacturing systems and the enabling technologies that facilitated these transformations.

The first industrial revolution marked a significant shift from handcrafted goods to mechanized production. The emergence of steam engines and water-powered machinery revolutionized production processes, particularly in the textile industry, leading to higher efficiency and mass-production capabilities [11]. However, this period also introduced labor-intensive factory systems, requiring a new workforce dynamic and creating urban migration trends as populations moved toward industrial centers [10]. Although this revolution improved manufacturing output, it was constrained by manual labor requirements and limited automation [12].

The second industrial revolution [late 19th to early 20th centuries] was characterized by electrification, standardization, and the expansion of global supply chains. The widespread adoption of electric power, assembly line production, and mechanized tools enabled manufacturers to achieve higher production speeds and lower costs through economies of scale [11]. Innovations, such as Henry Ford’s assembly line, revolutionized industrial operations by introducing process standardization and modular production techniques [9]. This period also saw the rise of mass consumer markets, prompting industries to optimize logistics, distribution networks, and global trade strategies [10].

The third industrial revolution [mid-to-late 20th century] introduced automation, computer-integrated manufacturing, and robotics, transforming manufacturing into a highly efficient, programmable, and scalable system [12]. Adopting computer numerical control (CNC) machines, industrial robotics, and enterprise computing systems enabled higher precision, repeatability, and efficiency, thereby reducing dependence on human labor [9]. This era marked the beginning of digital transformation in manufacturing, as companies integrated information technology, cybernetics, and early AI to improve production planning, quality control, and supply chain management [11]. However, automation increased efficiency, and it also led to workforce displacement, necessitating the development of new skills and labor policies [10].

The fourth industrial revolution (I4.0) represents the convergence of advanced digital technologies, including artificial intelligence, the Internet of Things (IoT), DTs, cyber–physical systems, and edge computing [12]. These technologies enable autonomous decision-making, real-time analytics, and interconnected smart factories, allowing for unprecedented flexibility and adaptability [10]. The economic drivers of this revolution include the need for hyperefficient data-driven manufacturing systems, the rise of customized and on-demand production, and growing concerns over sustainability and resource optimization [11]. Integrating AI-driven automation, edge computing, and DTs facilitates predictive maintenance, self-optimization of production lines, and decentralization of control systems, thereby enhancing productivity and resilience [9]. However, these advancements also introduce challenges, including cybersecurity threats, workforce reskilling demands, and the high cost of technological infrastructure [12]. I4.0 is increasingly being leveraged to address global manufacturing challenges, including supply chain disruptions, carbon footprint reduction, and energy efficiency, positioning smart factories as key enablers of sustainable industrial growth [10].

As discussed, industrial revolutions have reshaped manufacturing processes, organizational structures, and workforce dynamics [9]. Although each revolution introduced new technological enablers, their successful adoption required corresponding shifts in manufacturing paradigms, aligning strategic, operational, and digital frameworks with evolving industrial needs [11]. The transition from handcrafted goods to automated data-driven production highlights the continuous efforts required to balance efficiency, scalability, and customization in response to market, technological, and environmental pressures [9]. Furthermore, understanding this evolution requires examining the manufacturing paradigms that define how production systems have adapted to new technological landscapes. The following section explores the shift from craft-based production to intelligent digital manufacturing by analyzing key enablers, industry trends, and long-term implications in the I4.0 era [12].

2.2. Transformations in Manufacturing Paradigms

Progress in manufacturing paradigms is driven by economic imperatives, technological advancements, and evolving societal and environmental needs. Each paradigm has shaped product development, business models, and production processes, reflecting the pursuit of efficiency, customization, and sustainability by the industries [13].

Table 2. Key economic, societal, and technological drivers of manufacturing paradigms [13,14,15,16,17,18].

Paradigm	Economic Factors	Societal and Environmental Needs	Enabling Technologies	Product Development	Business Models
Craft Production	Localized production, limited scalability	High customization, artisan expertise, generational skill transfer	Manual labor, artisan skills, hand tools	Unique, handcrafted, small-batch production	Custom, made-to-order production
Mass Production	Cost reduction through economies of scale, increased consumer demand	Urbanization, standardization, global market expansion	Assembly lines, electricity, mechanized tools	Standardized products, bulk production	Large-scale, economy-of-scale approach
Lean Manufacturing	Competitive markets, cost optimization, process efficiency	Waste reduction, improved resource utilization, sustainability	Automation, just-in-time (JIT), quality control systems	Minimizing defects, maximizing efficiency	Continuous improvement (Kaizen), JIT manufacturing
Mass Customization	Balancing efficiency and variety, customer-driven production	Demand for product variety, modular and flexible production	IoT, robotics, AI, digital platforms	Adaptable products with customer-driven features	Flexible production, cost-effective customization
Personalized Manufacturing	AI-driven efficiency, real-time adaptability, demand-responsive production	Sustainability-driven production, digital transformation, consumer co-creation	AI, IoT, DTs, open-architecture systems, CPS	Hyper-customization, real-time customer-driven production	Personalization, pay-per-use models, on-demand production

lists an overview of the key economic, societal, and technological drivers of each manufacturing paradigm, highlighting their impacts on product development, business models, and manufacturing processes.

Craft production, which was the earliest form of manufacturing, relied on artisanal expertise and localized production. This system was constrained by limited scalability and high production costs, as skilled artisans handcrafted products in small batches [14]. The business model predominantly followed a made-to-order approach in which each product was uniquely tailored to customer needs [15]. However, the absence of standardization and mass-production techniques limited craft production in niche markets [16]. The introduction of mass production in the early 20th century transformed manufacturing by enabling high-volume and cost-effective production [17]. Economic factors, such as growing consumer demand and cost reduction through economies of scale, were key drivers [13]. The assembly line standardized production workflows, reduced variability, and increased efficiency [14]. Despite its advantages, mass production prioritized efficiency over customization, leading to a lack of product variety [15]. Lean manufacturing emerged in response to competitive pressures and waste reduction needs, particularly in the automotive and electronics industries [17]. The adoption of Kaizen principles, just-in-time (JIT) methodologies, and continuous process optimization allowed manufacturers to eliminate waste, streamline operations, and improve resource efficiency [13]. Societal concerns over environmental impacts and sustainability further reinforced lean principles, making them a standard in modern industrial operations [14]. With advancements in flexible manufacturing systems, digital platforms, and modular design, mass customization has become the dominant paradigm [18]. Economic factors, including the need to balance efficiency with product variety, drove the transition from standardized production to consumer-driven manufacturing models [15]. Integrating digital tools, robotics, and AI-driven automation enabled real-time adaptation to customer preferences while maintaining production efficiency [14]. The latest paradigm, personalized manufacturing, is characterized by AI-driven decision-making, cyber–physical systems, and DTs [17]. This model emphasizes real-time adaptability, hyper-customization, and sustainability-driven production, leveraging IoT and advanced analytics to create demand-responsive systems [16]. In this era, consumers are becoming increasingly involved in co-creation and product personalization, shifting from passive to active participation in the manufacturing process [18]. The co-evolution of industrial revolutions and manufacturing paradigms underscores how technological advancements reshape production systems [13]. Integrating autonomous systems, decentralized production, and CE principles defines the next phase of industrial transformation [17]. However, the challenge lies in harmonizing these advancements with sustainability goals, ethical considerations, and long-term resilience to ensure that future manufacturing paradigms benefit businesses and society [14].

As manufacturing paradigms continue to evolve in response to economic pressures, technological advancements, and societal expectations, the focus has increasingly shifted toward sustainability, customization, and CE principles. Although I4.0 has introduced intelligent, data-driven, and decentralized manufacturing systems, emerging challenges necessitate a more responsible approach to technological transformation. The next phase of industrial evolution will be shaped by the ability to integrate sustainable production models, hyper-personalization, and responsible innovation to ensure resilience and efficiency in global manufacturing. These drivers, along with the challenges they introduce, are explored in Section 3, which defines the key concerns, values, parameters, and KPIs for sustainable and circular manufacturing and discusses the role of hyper-personalization in future production systems.

The evolution of industrial revolutions and manufacturing paradigms highlights the continuous drive for efficiency, adaptability, and customization. However, past transitions have often neglected environmental sustainability, ethical labor practices, and the long-term impact of industrial growth. As the manufacturing industry embraces AI, IoT, blockchain, and DTs, it must also address responsible technological transformation, ensuring that digital advancements do not exacerbate environmental or ethical challenges while enhancing sustainability. CE principles, energy-efficient AI algorithms, and blockchain-enabled transparency in supply chains are reshaping global manufacturing by minimizing waste, optimizing resource usage, and reducing the carbon footprint.

Section 3 explores the role of responsible technological innovation in fostering sustainable circular manufacturing models. As manufacturing enters a new phase defined by AI, hyper-personalization, and closed-loop production, the challenge lies in ensuring that these advancements serve economic and environmental objectives rather than merely accelerating production at the cost of sustainability.

3. Responsible Technological Transformation for Sustainable and Circular Manufacturing

The rapid advancement of I4.0 technology has brought manufacturing to a critical crossroads. Although AI, IoT, blockchain, and DTs have the potential to redefine efficiency, scalability, and automation, their unregulated adoption has accelerated environmental and ethical challenges. The prevailing linear production model, in which resources are extracted, used, and discarded, has severe ecological consequences, including resource depletion, excessive carbon emissions, and unsustainable waste generation. Without a shift toward responsible technological integration, these innovations risk deepening sustainability crises rather than mitigating them.

To reverse these unsustainable trends, the manufacturing sector must adopt circular and responsible production models prioritizing waste elimination, resource regeneration, and extended product lifecycles. Circular manufacturing moves beyond merely reducing waste and emphasizes remanufacturing, recycling, and material recovery to create closed-loop, sustainable production ecosystems. By integrating AI-driven automation, IoT-enabled resource tracking, and blockchain-based transparency, industries can optimize their resource efficiency, reduce emissions, and transition toward regenerative production models. However, achieving truly responsible technological transformation requires more than just technological adoption. It demands strong ethical governance, international regulatory alignment, and lifecycle-oriented design principles that ensure sustainability at every production stage.

A successful transition toward circular and responsible manufacturing requires a structured framework integrating sustainability, economic viability, and social responsibility. The triple bottom line (TBL) framework provides this foundation by ensuring that manufacturing strategies balance profitability, environmental stewardship, and equitable labor practices.

Additionally, the rise of hyper-personalization driven by AI, DTs, and I4.0 presents opportunities and challenges for sustainable production. Unlike traditional mass manufacturing, hyper-personalized production allows on-demand customer-specific product customization, which reduces overproduction, minimizes material waste, and enhances resource efficiency. However, if not responsibly implemented, hyper-personalization can increase energy demand and supply chain complexity, potentially undermining its sustainability benefits. This section explores how TBL principles, CE strategies, and hyper-personalization can be integrated to drive sustainable technology-enabled manufacturing transformation.

3.1. Principles and Metrics for Sustainable Manufacturing

Sustainability in manufacturing is a TBL-centered multi-dimensional concept. It integrates economic, environmental, and social considerations into industrial operations. The TBL framework ensures that manufacturing processes optimize resource efficiency, minimize environmental impacts, and maintain long-term economic viability while prioritizing social well-being [19]. Driven by regulatory mandates, increasing consumer awareness, and the necessity of resource optimization, sustainable manufacturing has become a core strategic objective for industries [20]. This aligns with global sustainability goals and encourages the transition from traditional manufacturing toward cleaner, circular, and technology-driven sustainable practices [21].

The existing literature identifies two primary approaches to achieving sustainability in manufacturing: the cumulative and trade-off approaches. The cumulative approach posits that economic, environmental, and social goals can be pursued simultaneously, reinforcing one another, whereas the trade-off approach suggests that prioritizing one dimension may come at the expense of another [19]. The cumulative approach is often associated with I4.0 technologies, such as AI, IoT, and DTs, which facilitate a balance between efficiency and sustainability. By contrast, firms adopting the trade-off approach may prioritize short-term financial gains, often at the expense of long-term environmental or social sustainability.

Based on the TBL framework, sustainable manufacturing is classified into three key disciplines: economic, environmental, and social. Economic sustainability ensures business profitability by minimizing waste, reducing production costs, and optimizing energy consumption. It emphasizes cost efficiency, resource optimization, and long-term profitability, enabling firms to remain competitive while reducing operational waste [22,23]. Environmental sustainability focuses on improving energy efficiency and conserving resources. It prioritizes reducing carbon footprint, enhancing resource utilization, and implementing closed-loop production systems to mitigate environmental impact [21,24]. Social sustainability promotes workforce safety, fair labor practices, and ethical supply chain management, ensuring long-term societal benefits. It encompasses workplace safety, ethical sourcing, and corporate social responsibility and contributes to employee well-being and sustainable business practices [25,26].

Recent studies have emphasized the need for integrated sustainability frameworks that combine sustainability objectives with process-oriented models to systematically measure and improve performance [27]. Sustainability-oriented innovation has played a pivotal role in this transition, with firms increasingly adopting green technologies, eco-design strategies, and digital solutions to embed sustainability into manufacturing [28]. Effective implementation requires data-driven decision-making and leveraging of bibliometric and content analysis methodologies to track progress and identify gaps in current sustainability practices [29].

A structured approach to sustainable manufacturing integrates economic, environmental, and social considerations into key performance metrics. Each discipline presents distinct challenges, values, and measurement parameters that collectively define sustainability assessments in industrial settings. The role of open-architecture products is gaining attention as modular design extends product lifecycles and reduces material waste [16]. Additionally, barrier analysis methods, such as the best–worst method (BWM), have been employed to systematically identify key obstacles to sustainability adoption and to provide structured solutions [30].

Table 3. Key disciplines, concerns, and performance indicators in sustainable manufacturing [20,21,22,23,24,25,26,27,28,29,30].

Disciplines	Major Concerns	Parameters	KPIs
Economic (Profit)	Cost efficiency, productivity, profitability	Production cost, ROI, energy cost savings	Profit margins, ROI, cost savings percentage
Environmental (Planet)	Carbon footprint reduction, waste management, resource efficiency	CO2 emissions, material consumption, energy usage	CO2 reduction, energy efficiency improvements
Social (People)	Workforce safety, ethical sourcing, social responsibility	Workplace safety incidents, fair wages, working hours	Injury rates, employee satisfaction, fair wage compliance

lists the core TBL disciplines of sustainable manufacturing, highlighting key concerns, values, parameters, and associated KPIs.

Adopting I4.0 technologies—including AI-driven analytics, DT-enabled monitoring and optimization, and automation—transforms sustainability measurement, allowing firms to track and enhance sustainability metrics in real time [31,32,33]. Organizations aiming to improve their sustainability performance must integrate these parameters into their strategic decision-making, investment planning, and operational frameworks [27]. To effectively incorporate sustainability into manufacturing, companies must analyze the drivers, challenges, and opportunities associated with the economic, environmental, and social dimensions. A well-defined sustainability framework enhances organizational resilience, ensures compliance with global sustainability standards, and strengthens long-term competitiveness [34]. This necessitates comprehensive sustainability assessments that leverage digital tools, circular economic strategies, and sustainable business models [28].

Table 4. Strategic considerations for sustainable manufacturing [20,21,22,23,24,25,26,27,28,29,30].

Dimension	Key Drivers	Challenges	Opportunities
Economic (Profit)	Market competitiveness, financial performance	High costs of sustainability adoption	Increased profitability, cost savings from efficiency
Environmental (Planet)	Regulatory compliance, climate change impact	Resource scarcity, waste management issues	Adoption of renewable energy, CE integration
Social (People)	Ethical labor practices, community well-being	Workforce resistance, skill gaps	Enhanced brand reputation, better employee retention

lists key sustainability dimensions and their associated strategic considerations.

Addressing these dimensions requires cross-functional collaboration, standardization across industries, and adopting next-generation sustainability frameworks. Companies implementing sustainability strategies often experience enhanced operational efficiency, stronger stakeholder relationships, and long-term competitive advantages [35]. However, digitalizing sustainability initiatives presents opportunities and challenges, requiring firms to overcome technological, organizational, and regulatory barriers [36,37]. Integrating emerging technologies, such as blockchain, AI, and DTs, enables real-time sustainability tracking and predictive insights. These advancements have helped manufacturers proactively address supply chain risks, carbon emissions, and resource efficiency, thereby strengthening their sustainability strategies.

Furthermore, systematic reviews of sustainability frameworks indicate that incorporating open-product architectures, modular design, and CE principles is essential to ensuring that manufacturing systems remain flexible, resilient, and environmentally responsible over time [38]. As industries move toward sustainable digital transformation, aligning industrial strategies with global sustainability objectives is critical for maintaining economic viability and environmental stewardship [39]. By embedding sustainable development principles into industrial strategies and leveraging intelligent data-driven decision-making, manufacturers can establish resilient, adaptive, and responsible production models that support business growth and environmental sustainability [40]. These sustainability principles are closely aligned with the objectives of Sustainable Development Goal 9, which emphasizes inclusive industrial development, sustainable infrastructure, and innovation-driven economic growth. The adoption of digital twins and AI technologies contributes to upgrading industrial capabilities and fostering resilient, ecoefficient production systems.

3.2. CE Strategies in Manufacturing

The CE has emerged as a strategic alternative to the conventional linear economy, which follows a “take–make–dispose” model. The CE fosters a regenerative industrial system that reduces resource consumption and waste generation while maximizing the utility of materials throughout their lifecycles. It integrates sustainability principles, resource efficiency, and closed-loop material flows to generate long-term economic and environmental benefits [41]. In manufacturing, CE signifies a paradigm shift, emphasizing product and process designs that extend material longevity, optimize resource utilization, and reduce reliance on virgin raw materials [42].

The 10R strategy is a foundational framework within the CE, outlining a hierarchical approach to resource efficiency: refusing, relinking, reducing, reusing, repairing, refurbishing, repurposing, recycling, and recovery. Each strategy plays a distinct role in minimizing waste and maximizing material value. For example, “Refuse” and “Rethink” promote alternative consumption models, such as product-as-a-service and modular design, whereas “Reduce” focuses on minimizing material use at the design stage [43]. “Reuse” and “Repair” extend product life spans, whereas “Refurbish” and “Remanufacture” restore used products to near-original conditions, decreasing the need for new production [44]. At lower levels of the hierarchy, “Repurpose” assigns new functions to used components, whereas “Recycle” and “Recover” extract valuable materials and energy from end-of-life products [45]. Implementing the 10R principles in manufacturing requires advanced production strategies and innovative business models. I4.0, including smart sensors, additive manufacturing, and AI, enhances real-time material tracking and improves remanufacturing efficiency [46]. Additionally, modular product design and DTs support predictive maintenance and lifecycle assessment, enabling manufacturers to extend product durability and optimize resource utilization [47]. Transitioning to CE-driven manufacturing requires supply chain collaboration, particularly in reverse logistics systems, which streamline the collection, sorting, and reintegration of used products into production cycles [48]. Although the 10R principles offer significant advantages, their implementation in manufacturing faces several challenges. Economic feasibility, regulatory barriers, and technological limitations often hinder their full-scale adoption. Companies must balance cost efficiency and sustainability, necessitating targeted incentives and policy support for CE practices [49]. Furthermore, standardized assessment metrics and CE-specific KPIs are essential for tracking progress and ensuring effective integration into industrial operations [50]. Overcoming these challenges requires continuous innovation and cross-sectoral collaboration to establish a fully circular production system.

The CE framework provides a structured approach for minimizing resource consumption, optimizing waste management, and promoting sustainable industrial practices. The shift from a linear to a circular model is guided by key principles, including resource recovery, sustainable supply chains, and waste reduction [51]. Each 10R strategy is critical in enhancing material efficiency and mitigating the environmental impact of manufacturing operations [52]. Industries have integrated the 10R framework to enhance production sustainability using diverse strategies. In the electronics sector, a modular design enables easier repair and upgrades, thereby reducing product obsolescence and waste generation [49]. Similarly, the automotive industry has embraced remanufacturing, restoring used components to their original specifications and decreasing the reliance on virgin raw materials [51]. Additionally, industrial symbiosis, in which waste outputs from one sector serve as inputs for another, has gained momentum, fostering cross-industry resource optimization [52]. CE strategies fundamentally reshape supply chain management by emphasizing closed-loop production systems and reverse logistics. A well-designed circular supply chain ensures the efficient collection, sorting, and reintegration of end-of-life products into manufacturing cycles. Companies adopting circular principles benefit from reduced material consumption and improved resilience to resource scarcity [51]. However, full implementation faces challenges, including the depletion of raw materials, which necessitates greater reliance on secondary resources [49]. Additionally, regulatory compliance imposes hurdles as industries must navigate complex policies related to waste management and sustainability reporting [53]. Overcoming these barriers requires robust regulatory frameworks and cross-sector collaboration [51]. The effectiveness of CE practices is assessed using KPIs, such as recycling rates, waste reduction percentages, and secondary raw material utilization [54]. These metrics provide valuable insights into sustainability performance and enable continuous improvements to circular production models [51].

Table 5. CE strategies (0R to 9R), concerns, and KPIs.

Discipline	Major Concerns	Values	Parameters	KPIs
R0 Refuse	Eliminating unnecessary materials	Minimalism	Zero-waste strategies	Percentage material reduction in design
	Avoiding harmful chemicals	Safe material alternatives	Sustainable material sourcing	Percentage reduction in toxic materials
R1 Rethink	New business models	Servitization	PaaS models	Revenue from circular services
	Product lifecycle extension	Design innovation	Modular product design	Increase in average product lifespan
R2 Reduce	Minimizing resource use	Resource efficiency	Lean manufacturing	Material usage per product
	Reducing energy demand	Energy optimization	AI-driven energy management	Energy savings per operation
R3 Reuse	Maximizing component reuse	Product life extension	Design for reuse	Percentage of components reused in new products
	Refurbished part usage	Circular design	Closed-loop supply chains	Percentage of reused/refurbished components
R4 Repair	Right-to-repair policies	Consumer empowerment	Easy-repair product design	Repairability score
	Extended service contracts	Sustainable service models	Maintenance programs	Percentage of customers opting for repair
R5 Refurbish	Restoring products for resale	Circular refurbishment	Industrial remanufacturing	Percentage increase in refurbished sales
	Extending product usability	Secondary markets	B2B and B2C refurbishment	Resale value retention rate
R6 Remanufacture	Component-level reuse	Advanced disassembly	AI-assisted remanufacturing	Percentage recovery of valuable materials
	Large-scale CE	Remanufacturing policies	Government incentives	Market share of remanufactured goods
R7 Repurpose	Alternative product applications	Functional repurposing	Adaptive reuse strategies	Percentage of products repurposed
	Industrial by-product reuse	Cross-industry integration	Industrial symbiosis programs	Percentage of by-products utilized in other industries
R8 Recycle	Material recovery efficiency	High-quality recycling	AI-based sorting and processing	Percentage material recovery rate
	End-of-life recycling	Circular raw materials	Closed-loop recycling systems	Percentage of recycled materials in production
R9 Recover	Energy recovery from waste	Waste-to-energy	Biomass and thermal conversion	Percentage energy reclaimed from waste
	Alternative recovery methods	Innovative recovery	Chemical/material recovery processes	Recovery efficiency percentage

lists a structured overview of the 10R strategies, outlining the major concerns, core values, key parameters, and KPIs. This illustrates how each strategy enhances resource efficiency, minimizes waste, and promotes sustainability in manufacturing and supply chain operations. By implementing these strategies, addressing concerns, and monitoring KPIs, industries can effectively transition toward the CE, balancing economic viability with environmental sustainability.

Upgradable product design is a key enabler of a CE that extends product lifecycles, reduces material consumption, and improves resource efficiency. By facilitating modular upgrades rather than full product replacements, manufacturers can lower production costs while minimizing their environmental impact [55]. This approach is particularly beneficial in industries, such as consumer electronics and machinery, where rapid technological advancements often lead to obsolescence. Companies incorporating upgradability into their design strategies enhance sustainability and generate new revenue streams through after-sales services and component sales [55]. Circular manufacturing requires a flexible contingency-based approach that aligns production processes with resource efficiency and waste reduction principles. Successful startups implementing circular manufacturing have adopted flexible production lines, modular design principles, and advanced digital tracking systems to optimize material flows [56]. These adaptive strategies enhance supply chain resilience by minimizing reliance on virgin raw materials and maximizing the use of secondary resources [56]. Additionally, circular manufacturing fosters industrial symbiosis, wherein waste from one process becomes an input for another, promoting cross-industry collaboration and resource efficiency [56]. Economic incentives are instrumental in accelerating the adoption of CE practices. Dynamic pricing models and targeted investment strategies help businesses balance their cost-efficiency and sustainability goals. Companies that integrate recycling investments into their business models enhance their financial viability by leveraging material recovery and reprocessing [57]. Additionally, government subsidies and tax incentives are pivotal in promoting circular initiatives by offsetting the initial costs of transitioning to sustainable production [57].

Integrating upgradability, adaptive manufacturing, and economic incentives into CE strategies offers industries a pathway to sustainability while maintaining competitiveness. By prioritizing design innovation and financial models that support circular principles, businesses can establish resilient and resource-efficient operations that align with long-term environmental and economic objectives.

I4.0 technologies serve as key enablers of CE by enhancing resource efficiency, sustainable production, and supply chain optimization. Integrating AI, big data analytics, blockchain, and IoT improves transparency, traceability, and operational efficiency in circular supply chains [58]. These advancements have enabled real-time material tracking, predictive maintenance, and optimized resource flows, thereby reducing waste generation and extending product lifespan [59]. I4.0 facilitates sustainable production through smart manufacturing systems that leverage CPSs, automated decision-making, and DTs. These technologies enhance the precision of material usage, reduce energy consumption, and optimize product remanufacturing processes [60]. For instance, additive manufacturing (3D printing) supports on-demand production with minimal waste, whereas blockchain ensures the secure and transparent tracking of recycled materials and reused components [58]. I4.0 technologies drive the adoption of circular business models, including servitization, PaaS models, and reverse logistics. Smart sensors and IoT devices enable the real-time monitoring of product conditions, facilitate predictive maintenance, and extend product lifespans [59]. Additionally, digital connectivity strengthens collaboration among suppliers, manufacturers, and consumers, ensuring seamless reintegration of recovered and repurposed materials within circular supply chains [60]. Despite its advantages, integrating I4.0 with the CE presents challenges, including high implementation costs, data security risks, and the need for specialized digital infrastructure [59]. Additionally, although automation and smart technologies improve sustainability, effective adoption requires workforce upskilling and expertise in digital solutions [58].

I4.0 technologies are pivotal for advancing CE strategies, enabling sustainable production, resource optimization, and closed-loop supply chains. By leveraging digital innovation, businesses can enhance environmental sustainability while maintaining economic viability. However, to fully realize the potential of I4.0 in CE, addressing implementation challenges through policy support, targeted investments, and workforce upskilling is essential.

3.3. Hyper-Personalization

Hyper-personalization, powered by AI-driven predictive analytics and real-time data insights, has emerged as a transformative force in customer engagement, enabling businesses to create individualized experiences tailored to specific consumer behaviors and preferences [61]. Unlike traditional personalization approaches that rely on broad demographic segmentation, hyper-personalization employs advanced machine learning algorithms to analyze consumer interactions across multiple touchpoints, delivering context-aware and need-based recommendations [62]. This level of customization fosters greater customer satisfaction and loyalty while aligning with broader sustainability goals by reducing excess production and minimizing resource wastage [63]. As industries transition toward circular business models, hyper-personalization is crucial in demand-driven manufacturing, shifting from mass production to dynamic on-demand solutions that optimize resource use while enhancing consumer experience [64].

Hyper-personalization is crucial in sustainable and circular manufacturing by optimizing resource efficiency and minimizing waste using tailored production and consumption strategies [65]. By leveraging AI-driven demand forecasting, businesses can reduce excess inventory and overproduction, which are the primary sources of material and energy waste in traditional manufacturing models [66]. Furthermore, hyper-personalization enables localized production, which lowers the carbon footprint associated with global supply chains by shifting towards on-demand customer-centric manufacturing processes [63]. This shift aligns with the CE principles by promoting shorter supply chains, reducing transportation emissions, and enhancing supply chain resilience [66]. Additionally, modular product designs and upgradable components enhance sustainability by extending product lifecycles, reducing the need for frequent replacements, and enabling users to customize or repair their products rather than discard them [65]. This approach benefits the environment and supports the transition from traditional ownership-based models to service-oriented business models, such as PaaS, which encourage resource efficiency and circularity by ensuring that products remain in use for longer periods [66].

The convergence of hyper-personalization with Industry 5.0 principles facilitates the development of sustainable manufacturing ecosystems that prioritize human–machine collaboration, flexibility, and circular resource utilization [67]. Industry 5.0, emphasizing mass customization and consumer-centric production, leverages real-time data analytics and AI-driven decision-making to enhance product personalization while minimizing environmental impact [68]. By integrating hyper-personalization with advanced manufacturing technologies, such as additive manufacturing and DTs, companies can streamline production processes, reduce waste, and align product design with evolving consumer demands while incorporating the 4Cs of mass customization—co-creation, configuration, customer involvement, and choice navigation—to enhance product adaptability and user engagement [69]. This dynamic approach allows businesses to create highly individualized products without compromising sustainability, reinforcing CE principles by fostering resource-efficient production and extending product lifespans [67].

Integrating hyper-personalization within CE frameworks fosters a paradigm shift from conventional production models to adaptive demand-driven approaches. By leveraging DT technology, AI-driven analytics, and real-time data insights, businesses can enhance material efficiency and reduce waste through predictive maintenance and performance optimization [63]. Unlike traditional customization strategies, which rely on predefined product modules, hyper-personalization integrates real-time consumer data and AI-driven insights to dynamically tailor products and services to unique customer preferences [70]. This approach enhances resource efficiency by aligning production output more precisely with actual demand, minimizing excess inventory, and supporting sustainable consumption patterns [71]. Furthermore, DT technology facilitates hyper-personalization in circular manufacturing by simulating product performance and enabling the predictive optimization of materials, ensuring that products meet customer expectations while maximizing resource efficiency [72].

Hyper-personalization in circular business models is further strengthened by integrating mass customization principles with resource-sharing and remanufacturing strategies. The convergence of mass customization and CE principles enables businesses to design products with modularity and extended lifecycles, supporting refurbishment, remanufacturing, and repurposing initiatives [73]. By embedding digital intelligence into production systems, manufacturers can adjust product features and configurations in real time based on consumer preferences, ensuring efficient material utilization and reducing production waste [74]. Additionally, smart manufacturing environments enable companies to create flexible and responsive production lines that adapt to evolving customer needs without increasing environmental burden [75]. This approach fosters a closed-loop manufacturing system in which products are designed for reusability and repair, extending their lifespan and reducing material depletion [73].

Despite its potential to enhance sustainability, hyper-personalization within CE frameworks presents several challenges, particularly concerning data privacy, computational demand, and supply chain integration. The extensive collection and processing of consumer data required for hyper-personalized production raises significant concerns regarding privacy regulations, such as GDPR and CCPA, necessitating transparent consent management and robust data security measures [61]. Additionally, the real-time processing of vast datasets for personalized manufacturing relies on an AI-driven infrastructure that demands high computational power, which may increase energy consumption and offset sustainability benefits if not effectively managed [65]. Integrating hyper-personalization into circular supply chains also requires seamless coordination among manufacturers, suppliers, and logistics providers to optimize production efficiency and minimize resource waste and transportation emissions [63]. Emerging technologies, such as blockchain and decentralized digital ledgers, are being explored to enhance transparency in product lifecycle tracking, ensuring ethical sourcing, material recovery, and waste minimization in CE applications. Overcoming these challenges will require cross-industry collaboration; adaptive regulatory frameworks that balance innovation with ethical considerations; and continuous advancements in AI, smart manufacturing, and sustainability-driven digital transformation strategies.

Sustainability and circularity have become the defining pillars of modern manufacturing, and the role of advanced technologies in enabling these strategies has become increasingly significant. Section 3.4 explores the key enablers of future manufacturing, such as microconnected factories and USFs, and their synergistic integration with DT technologies to enhance sustainability, efficiency, and flexibility.

3.4. Responsible Technological Transformation

The rapid evolution of I4.0 and the emerging Industry 5.0 have catalyzed unprecedented advancements in manufacturing through AI, DTs, CPSs, and IoT. Although these technologies contribute significantly to operational efficiency, product customization, and resource optimization, their implementation also raises concerns regarding ethical responsibility, environmental sustainability, and socioeconomic impacts. Previous studies have established the critical role of these technologies in fostering sustainable manufacturing practices. However, achieving a truly responsible technological transformation requires a strategic balance between efficiency and ethical considerations. This section explores the key challenges and strategies to ensuring that responsible AI integration, secure DT deployment, sustainable CE strategies, and ethical hyper-personalization models are at the forefront of the next phase of industrial transformation.

As previously highlighted, AI-driven automation has enabled demand-driven production systems, material waste reduction, and supply chain optimization. However, unregulated AI adoption may lead to unintended consequences, such as workforce displacement, algorithmic bias, and safety risks in human–machine collaborations. Hyper-automated production lines risk replacing human workers without adequate reskilling initiatives. AI models trained on historical data may unintentionally reinforce biased decision-making, affecting product personalization and quality control. Additionally, if improperly integrated, AI-powered robotics can introduce workplace hazards and ergonomic issues.

To address these challenges, manufacturers must implement human–AI collaboration models that enhance worker efficiency without eliminating human roles. Developing explainable AI frameworks ensures accountability in smart manufacturing decisions, whereas workforce reskilling initiatives focused on AI governance, robotics maintenance, and human-in-the-loop decision-making can facilitate ethical workforce transitions.

DTs are a cornerstone of circular and sustainable manufacturing as they enable real-time tracking of material flows, predictive maintenance, and lifecycle optimization. In addition to improving the resource efficiency in smart factories, DTs facilitate remanufacturing, waste reduction, and closed-loop production by providing data-driven insights into product longevity, reusability, and recycling pathways. However, as DTs become increasingly integrated with cloud-based infrastructures and decentralized networks, their deployment introduces cybersecurity and data privacy challenges. DTs store highly sensitive operational data, making them prime targets for cyberattacks, industrial espionage, and unauthorized access. Additionally, unclear data governance policies complicate how manufacturing firms share, store, and utilize DT insights.

To ensure responsible technological transformation, manufacturers must implement secure-by-design architectures that leverage end-to-end encryption, zero-trust security frameworks, and blockchain-based transparency to protect data integrity and prevent tampering. Furthermore, compliance with global data sovereignty laws ensures that DT applications adhere to privacy, security, and ethical standards. By combining sustainability-driven innovations with robust cybersecurity measures, manufacturers can fully realize the potential of DTs in creating a resilient, transparent, and environmentally responsible production ecosystem. However, digital transformation extends beyond cybersecurity and operational efficiency and introduces new challenges in sustainability. CE strategies emphasize the urgent need to transition from linear production models to closed-loop regenerative manufacturing systems.

To ensure responsible technological transformation, manufacturers must proactively address these challenges by leveraging DT technology for sustainability, efficiency, and CE integration.

• Optimizing resource efficiency with digital twins enables real-time simulations and predictive analytics, allowing manufacturers to optimize production processes, reduce energy waste, and enhance material efficiency. By integrating edge computing and AI-powered DTs, companies can minimize the carbon footprint of manufacturing operations while ensuring energy-efficient production.
• Unlike traditional IoT-based tracking systems, enhancing closed-loop material tracking with DTs provides a dynamic virtual representation of the entire supply chain, ensuring accurate resource utilization, predictive remanufacturing, and lifecycle tracking. By monitoring the material flow and product durability in real time, DTs support remanufacturing strategies, material recovery, and waste reduction, which are key enablers of CE.
• Extending product lifecycles with modular design DTs to facilitate virtual prototyping and lifecycle assessment, enabling manufacturers to design modular upgradable components that reduce premature obsolescence. By continuously analyzing product performance and degradation, DTs can assist in adaptive maintenance planning, material reuse, and extending the functional life of industrial assets.

By embedding DT technology at the core of circular manufacturing strategies, industries can achieve higher operational resilience, lower waste generation, and enhanced sustainability while ensuring a transparent, secure, and responsible transformation toward Industry 5.0.

4. The Future of Sustainable and Intelligent Manufacturing: DTs and Decentralized Smart Factories

4.1. Connected Micro Smart Factory

Modular manufacturing systems have emerged as a transformative production approach, enabling manufacturers to build highly flexible and reconfigurable production environments. These systems rely on standardized, interoperable modules that can be dynamically reconfigured to accommodate varying product designs and requirements, thereby facilitating agile and scalable manufacturing processes [76]. At the core of modular manufacturing lies the Micro Smart Factory (MSF), which is a compact, intelligent production system that integrates advanced automation, robotics, and CPSs to enable high-efficiency, small-scale manufacturing [77]. The MSF concept aligns with I4.0 and Industry 5.0 paradigms by emphasizing decentralized production, real-time data analytics, and human–machine collaboration to enhance customization and operational efficiency [78]. Unlike conventional large-scale factories, MSFs enable localized demand-driven production, reducing logistical costs and environmental impact while improving responsiveness to consumer preferences [79].

A Connected MSF (CMSF) extends the MSF concept by integrating the Industrial IoT (IIoT), cloud computing, AI, big data analytics, and DTs to create an interconnected and highly autonomous manufacturing ecosystem [80]. This approach enables real-time monitoring, predictive maintenance, and autonomous decision-making, significantly improving production efficiency, sustainability, and adaptability [81]. A highly significant advancement in CMSFs is the integration of factory-as-a-service (FaaS), a paradigm in which production capabilities are offered as a service rather than as fixed, capital-intensive assets. FaaS leverages cloud-based platforms and DTs to enable businesses to achieve on-demand manufacturing, allowing small and medium enterprises to access high-end production capabilities without heavy upfront investments [82]. This enhances resource utilization and CE integration and supports mass customization and personalized production, driving greater sustainability and efficiency [83].

The CMSF model contributes significantly to sustainable manufacturing by reducing waste, optimizing resource utilization, and facilitating closed-loop production systems. By leveraging DT technology, manufacturers can simulate production processes, optimize energy consumption, and predict system failures, thereby minimizing downtime and reducing material waste [84]. Additionally, IoT-enabled real-time monitoring ensures continuous performance optimization, further enhancing operational sustainability [85]. From a CE perspective, CMSFs enable remanufacturing, component reuse, and material recovery through data-driven decision-making and predictive analysis. The modular and decentralized nature of CMSFs supports local production hubs that minimize transportation emissions while ensuring efficient material circulation within production cycles [78]. The FaaS model further enhances circularity by sharing manufacturing resources among multiple users, reducing idle capacity, and promoting a more sustainable production ecosystem [82]. Additionally, CMSFs enhance personalization by integrating AI-driven design automation with real-time consumer feedback mechanisms. This allows manufacturers to deliver hyper-personalized products while maintaining cost efficiency and sustainability. By leveraging DTs and cloud-based design platforms, manufacturers can dynamically adjust product specifications based on customer preferences, thereby reducing the need for excessive prototyping and minimizing raw material consumption [83].

The realization of connected micro smart factories heavily depends on technological advancements in IoT, AI, cloud computing, big data analytics, and DTs.

IoT and cloud computing: IoT sensors collect real-time data from machines, materials, and production lines, thereby enabling predictive analytics and performance optimization. Cloud computing ensures seamless data exchange between factory components and enhances flexibility and scalability [85].

AI and big data analytics: AI-driven algorithms analyze vast datasets to identify production inefficiencies, forecast demand, and automate decision-making, leading to smarter resource allocation and reduced production waste [79].

DTs: DTs provide a virtual representation of the physical factory environment, enabling real-time monitoring, simulation, and predictive optimization of manufacturing processes. This improves maintenance efficiency, energy management, and adaptive production strategies, ensuring sustainability and cost-effectiveness [80].

Reinforcement learning and CPSs: AI-powered reinforcement learning enables adaptive and resilient production control, allowing CMSFs to autonomously adjust operations in response to changing market demands and environmental conditions [83].

A CMSF represents a paradigm shift in modern manufacturing, bridging modular production, real-time data integration, and hyper-personalization with sustainability and CE principles. By leveraging FAAS, IoT, AI, and DTs, CMSFs can enhance manufacturing agility, reduce waste, support personalized production, and align industrial processes with next-generation, sustainable, and customer-centric manufacturing models. Future advancements in AI-driven automation, edge computing, and decentralized manufacturing networks will further strengthen CMSFs as critical enablers of the next industrial revolution, paving the way for self-optimizing, hyper-flexible, and resilient production ecosystems [83].

4.2. Urban Smart Factory

The urban smart factory (USF) represents an innovative manufacturing paradigm that integrates I4.0 technologies within urban environments to enhance sustainability, resilience, and human-centric production. Unlike traditional factories that are often located in peripheral industrial zones, USFs leverage urban infrastructure to provide localized, flexible, and highly automated production capabilities [86]. The rise in urbanization, mass personalization, and resource scarcity has driven the need for smart manufacturing solutions that optimize production processes while minimizing environmental and social impacts [87].

USFs are crucial for achieving Sustainable Development Goals (SDGs) by reducing transportation emissions, optimizing energy consumption, and integrating CE principles into production [88]. By utilizing local supply chains and minimizing waste generation, these factories contribute to resource-efficient production systems that align with modern sustainability imperatives [89]. Additionally, USFs facilitate local employment and enhance the resilience of urban economies by decentralizing production networks [90].

The successful implementation of USFs relies on various I4.0 technologies, including the IoT, DTs, big data analytics, AI, and cloud computing [91]. IoT-enabled devices facilitate real-time monitoring and control of manufacturing operations, ensuring seamless integration with urban logistics and smart city infrastructures [92]. DTs create virtual representations of production processes, enabling predictive maintenance, process optimization, and enhanced decision-making capabilities [93]. Furthermore, AI-driven analytics support adaptive production strategies that respond dynamically to market demand and urban conditions [94].

USFs employ a transformative manufacturing approach integrating sustainability, advanced technology, and urban infrastructure to create efficient and resilient production ecosystems. As these factories continue to evolve, the synergy between DTs, AI, and CPSs plays a pivotal role in enhancing their capabilities. Section 4.3 explores how these advanced technologies, particularly DTs, contribute to shaping the future of manufacturing through enhanced interoperability, process optimization, and real-time decision-making.

4.3. The DT Nexus: Enabling CE and Sustainable Manufacturing

The increasing complexity of global supply chains and the growing demand for sustainable production practices have driven the integration of DT technologies into CE frameworks. DTs offer real-time data exchange, predictive analytics, and system optimization, which are essential for achieving sustainability goals in manufacturing. The role of DTs in facilitating CE is expanding beyond process monitoring to support material reuse, resource efficiency, and lifecycle assessment [5].

The DT Nexus Framework (DTNF) illustrated in Figure 1 provides a structured approach for integrating DT technology into sustainable manufacturing and CE strategies. The framework comprises three interdependent dimensions, each of which is crucial in enhancing lifecycle efficiency, optimizing resource utilization, and increasing supply chain transparency.

The circular product lifecycle intelligence dimension focuses on enhancing lifecycle tracking and supporting sustainable product design using DTs. As shown in the figure, real-time monitoring allows manufacturers to track product conditions dynamically using sensor data and IoT-enabled DTs, thereby ensuring continuous data-driven insights. Predictive analytics leverage AI-driven insights to extend product lifespan, enabling proactive maintenance and failure prevention before significant breakdowns occur. Additionally, circular insights (repair and reuse) assist businesses in evaluating product wear, repair feasibility, and component reuse, reducing unnecessary production and supporting remanufacturing efforts. Finally, lifecycle extension strategies contribute to durability and sustainability by optimizing material usage and extending the product lifespan, which are crucial for long-term resource efficiency.

As illustrated in Figure 1, the CLMS dimension highlights the role of DTs in optimizing waste recovery, remanufacturing, and recycling loops, ensuring that manufacturing processes are aligned with sustainability goals. AI-driven process optimization enhances material efficiency by integrating real-time analytics into production systems and reduces material waste through data-driven insights. Remanufacturing and reuse allow for component-level reuse, adaptive disassembly, and product refurbishment, ensuring that products remain in circulation for longer periods. CPSs enable real-time adaptive production, improve process optimization, and facilitate automated reconfiguration of manufacturing networks. Another key component, zero-waste production, eliminates material waste through precision manufacturing and circular design principles, thereby ensuring the efficient use of resources. Additionally, standardization and interoperability play vital roles in ensuring cross-industry compatibility for DT-based CE applications by adopting standardized protocols that enable seamless integration. Finally, zero-waste product design and sustainable sourcing incorporate eco-design principles to minimize waste during the design phase while ensuring responsible material sourcing, thus contributing to the development of sustainable products.

The sustainable supply chain visibility dimension emphasizes digital twin-enabled supply chain tracking, logistics optimization, and sustainable sourcing. Material tracking ensures transparency and traceability by recording material flows across the supply chain, thereby allowing manufacturers to track the impact of a product’s lifecycle. Smart logistics and emission controls facilitate real-time logistics tracking, AI-driven route optimization, and carbon footprint reduction strategies, ensuring efficient and sustainable supply chain operations. Reverse logistics optimization enhances reuse, refurbishment, and recycling logistics, ensuring that products and materials are reintegrated into the economy rather than being disposed of prematurely. Furthermore, the supplier sustainability assessment evaluates supply chain partners for CE compliance using lifecycle analysis, emissions tracking, and environmental, social, and governance (ESG) metrics, ensuring that companies adhere to sustainable procurement practices. Additionally, circular business models and value chains promote PaaS, leasing, and closed-loop supply chains that maximize product utilization while minimizing waste. Industrial by-product repurposing redirects manufacturing by-products into secondary markets or alternative industries, enabling cross-industry material reuse and reducing landfill waste. Finally, waste-to-energy and material recovery utilize biomass, thermal recovery, and chemical recycling to convert non-recyclable waste into energy, maximizing material recovery efficiency while ensuring that waste is processed responsibly.

Through its three-dimensional structure, the DTNF creates a synergistic approach to sustainability, bridging the gap between technology-driven efficiency and environmental responsibility.

Table 6. The DT Nexus Framework: Three core dimensions [5,95,96,97,98,99].

Dimension	Key Role	Technology Enablers	DT Applications	Implementation Challenges	Ethical and Regulatory Considerations
Circular Product Lifecycle Intelligence	Enables real-time simulation and optimization of product lifecycles for circularity, extending material longevity and minimizing waste	AI-powered DTs + predictive analytics for intelligent decision-making in lifecycle management	Self-learning DTs for smart material tracking, AI-optimized remanufacturing, and blockchain-based circular design validation	High computational cost, energy-intensive AI models, and interoperability barriers between DT platforms	AI-driven lifecycle assessments require bias mitigation, and blockchain data transparency must balance privacy concerns
Closed-Loop Manufacturing Systems	Ensures autonomous process control in remanufacturing, recycling, and zero-waste production through adaptive DT intelligence	DTs + IoT + edge computing enable decentralized, responsive waste recovery systems	AI-driven reverse logistics, waste reduction models, and real-time optimization of decentralized smart factories	Complex coordination across supply chain nodes, standardization of CE metrics, and real-time AI model validation	Cross-border regulations on material reuse, AI compliance in automated remanufacturing, and sustainability monitoring gaps
Sustainable Supply Chain Visibility	Enhances transparency, traceability, and circular supply chain accountability through real-time data sharing and predictive optimization	DTs + AI-driven supply chain analytics + cloud-based circular networks for seamless, traceable logistics	Digital product passports, carbon footprint tracking, and AI-powered intelligent supply chain logistics	Lack of standardized global policies for DT traceability, data-sharing constraints, and integration complexity	AI-based risk modeling needs ethical alignment; environmental impact regulations must evolve with DT energy demands

lists a structured view of how DTs contribute to CE and sustainable manufacturing. It categorizes key functional areas, associated technologies, practical applications, and challenges industries face in implementing these solutions. Additionally, ethical and regulatory concerns are highlighted to ensure that DT deployment aligns with responsible technological transformation principles. This framework helps bridge the gap between digital innovation and sustainable practices, thereby promoting long-term economic and environmental benefits.

The three dimensions of the Digital Twin Nexus—circular product lifecycle intelligence, closed-loop manufacturing systems, and sustainable supply chain visibility—form an integrated framework that enables intelligent, closed-loop control and resource optimization across the manufacturing ecosystem. Circular product lifecycle intelligence delivers real-time insights into product condition, usage, and end-of-life potential, facilitating proactive decisions regarding reuse, repair, and re-design. These insights inform closed-loop manufacturing systems, which employ adaptive DT intelligence to autonomously manage remanufacturing, recycling, and zero-waste operations. In parallel, sustainable supply chain visibility ensures transparency, traceability, and predictive optimization across logistics networks. By interlinking suppliers, production systems, and recovery pathways through AI-driven analytics and digital product passports, it supports efficient resource circulation and waste minimization. Together, these dimensions reinforce one another: lifecycle intelligence drives manufacturing actions; manufacturing outputs influence supply chain flows; and supply chain feedback enhances lifecycle management. This synergistic interaction enables a resilient, self-optimizing, and circular industrial ecosystem.

Figure 2 visualizes the three core dimensions of the Digital Twin Nexus as an interconnected system supporting real-time data exchange, predictive analytics, and autonomous control. Collectively, they enable end-to-end optimization across the product lifecycle—from design to recovery—while dynamically aligning manufacturing processes with circular economy principles.

DTs facilitate the monitoring and tracking of product lifecycles from inception to end-of-life (EoL). DTs support decision-making in product design, refurbishment, and recycling by capturing data related to material composition, product usage, and maintenance history [95]. Integrating behavioral modeling techniques within DTs enables manufacturers to predict material degradation and determine optimal reuse strategies. For example, predictive analytics embedded within DTs can inform remanufacturing decisions by assessing the structural integrity of components before reuse [96].

A key CE principle is transitioning from a linear “take–make–dispose” model to a regenerative system where materials and components are continuously repurposed. DTs enable this transformation by creating virtual representations of physical assets, allowing for real-time monitoring and simulation of closed-loop production processes [97]. These systems facilitate automated reverse logistics, waste minimization models, and decentralized remanufacturing networks. AI-driven DTs optimize the collection, sorting, and reintegration of post-consumer materials. DTs identify inefficiencies in manufacturing workflows and propose real-time adjustments to reduce waste generation [5]. Moreover, DTs enable decentralized remanufacturing networks, which allow the coordination of manufacturing operations across multiple facilities to improve resource allocation and minimize transportation emissions [98].

Sustainability in global manufacturing requires enhanced visibility and traceability across supply chains. DTs provide data-driven infrastructure that supports transparency in sourcing, production, and distribution [3]. The deployment of digital product passports ensures that every stakeholder in the supply chain has access to essential product lifecycle data, enabling informed decisions regarding material provenance, recyclability, and compliance with sustainability standards [96]. Additionally, DTs enhance carbon footprint tracking, allowing manufacturers to measure and mitigate greenhouse gas emissions across production networks. AI-powered models integrated into DTs can optimize supply chain logistics to reduce energy consumption and improve resource utilization [98].

Thus, the DTNF presents a transformative approach to CE integration by leveraging real-time data, predictive analytics, and intelligent automation. However, several challenges remain, including the need for standardized data-sharing protocols, interoperability between DT platforms, and alignment with regulatory policies [3]. Future research should explore the development of federated DT ecosystems that connect stakeholders across industries, thereby enabling collaborative sustainability efforts. Moreover, advancements in edge computing and AI further enhance the efficiency and scalability of DT applications in circular manufacturing [99]. By embedding DTs within the core CE principles, manufacturers can achieve a paradigm shift toward sustainable, data-driven, and resilient production systems. This transition is essential for environmental sustainability, enhancing economic viability, and fostering long-term competitive advantages in global manufacturing.

5. Real-World Implementation of DT Strategies for Sustainable Industry

The DTNF aims to integrate real-time data, AI, and sustainable manufacturing principles to enhance operational efficiency and CE adoption. To validate the applicability of DTNF in real-world industrial scenarios, this section presents case studies demonstrating its implementation across diverse domains, including supply chain optimization, smart manufacturing, sustainability assessment, human-centric DT applications, and AI-driven factory evolution. These case studies provide empirical insights into how DTs contribute to responsible and sustainable manufacturing by optimizing production processes, reducing waste, and improving decision-making.

5.1. Case Study 1: Digital Twins for Supply Chain Optimization and Circular Economy

Supply chains play a vital role in the sustainability of manufacturing operations and have significant implications for waste reduction, resource efficiency, and carbon emissions. Traditional supply chains often suffer from inefficiencies owing to a lack of real-time visibility, poor demand forecasting, and linear production–consumption models, which result in excessive waste. Adopting DTs in supply chain management offers a transformative approach to achieving sustainability and CE objectives by enabling predictive analytics, intelligent decision-making, and closed-loop logistics.

Kim et al. (2025) [100] investigated the role of DT-based prediction and optimization in enhancing the operational efficiency and resilience of dynamic supply chains. Their study introduced a supply chain DT (SCDT) methodology (Figure 3) that integrates real-time monitoring, predictive analysis, and optimization to mitigate logistics disruptions and improve resource allocation. By employing a DT system architecture, the study outlines key components, such as a real-time simulation model and an optimization module, enabling manufacturers to anticipate supply chain disruptions, streamline logistics operations, and reduce inefficiencies. This section presents a case study of the supply chain of an automobile body production company, demonstrating the effectiveness of DT-enabled predictive optimization. Traditional supply chains often face challenges, such as inventory shortages, delayed deliveries, and inefficient demand forecasting. By contrast, the SCDT system dynamically synchronizes supplier, manufacturer, and logistics information, ensuring real-time visibility across supply chain nodes. Through IoT-enabled data aggregation and AI-driven decision-making, the DT model continuously evaluates factory performance, transportation networks, and production schedules to minimize inventory fluctuations and enhance supply chain responsiveness.

A key feature of this study is its demonstration of metaheuristics-based supply chain optimization. Using a tabu search algorithm, the SCDT system determines optimal production locations and transportation routes based on real-time inventory and cost fluctuations. This approach reduces logistics costs, improves material tracking, and enhances coordination between supply chain entities. The simulation-based framework validated these findings by showing an 8.97%, 1.3%, and 8.82% reduction in inventory costs, logistics costs, and total supply chain expenses, respectively. In addition to economic efficiency, Kim et al. (2025) [100] highlighted the sustainability benefits of DT-enabled supply chain management. By dynamically adjusting production and delivery schedules, the system reduces overproduction and transportation emissions, thereby addressing key environmental concerns in industrial manufacturing. Additionally, by optimizing logistics networks in response to real-time disruptions, this study demonstrates how DTs facilitate energy-efficient, low-carbon supply chain operations.

These findings align with the DTNF by validating the logistics and sustainability layers. The DTNF emphasizes the integration of AI-driven DT technologies to create data-driven, sustainable manufacturing ecosystems. This case study provides empirical evidence that DTs enable responsible and intelligent supply chain transformation, reduce waste, improve efficiency, and foster circular economic initiatives.

5.2. Case Study 2: Real-Time AI and Edge Computing for Smart Manufacturing

Manufacturing operations require constant monitoring and optimization to ensure efficiency, minimize energy consumption, and reduce waste. Traditional manufacturing systems often rely on centralized data processing models, which can create bottlenecks in decision-making owing to latency and network dependency. Conversely, edge computing combined with DT technology allows for real-time data processing at the source, improving operational efficiency and sustainability. By integrating AI-driven DTs with edge-computing architectures, manufacturers can optimize energy usage, enhance predictive maintenance, and achieve higher levels of automation.

Kang et al. (2024) [101] proposed an edge computing-based DT (E-DT) framework designed to enhance real-time data processing capabilities in smart manufacturing. Their study addresses the limitations of conventional DT architectures, which rely heavily on cloud-based processing, resulting in high latency and data transmission bottlenecks. By integrating edge computing into DT operations, the proposed framework enables faster data analysis, reduced network congestion, and enhanced real-time decision-making. Their research is based on ISO 23247 [102], which is a reference architecture for DTs in manufacturing.

The E-DT framework is shown in Figure 4. It extends this standard by introducing a data fusion model and a multilayer architecture that optimizes data management. The framework consists of four key entities: observable manufacturing elements, data collection and device control entities (DCDCEs), core entities, and user entities. Unlike traditional cloud-dependent DT architectures, the DCDCE in E-DT is enhanced with distributed edge-processing capabilities, reducing the reliance on centralized computing. This minimizes bandwidth consumption and improves real-time data availability for manufacturing operations.

A case study on Wire + Arc additive manufacturing (WAAM) demonstrates the effectiveness of the proposed E-DT framework. WAAM, a metal 3D printing technology, generates vast amounts of real-time sensor and video data, posing challenges for real-time monitoring and quality control. The study implemented an edge AI-powered DT system that processed welding bead images locally using deep learning-based computer vision techniques. The MobileNetV2 deep learning model was deployed on edge devices to classify welding bead quality, ensuring early defect detection and process optimization. The results highlight significant performance improvements. The E-DT framework reduces data processing latency compared to cloud-based DT architectures, allowing manufacturers to detect welding defects in real time and make instant corrective adjustments. Additionally, the network load is reduced by preprocessing the data locally at the edge, leading to faster and more reliable data transmission. This study further validated that E-DT enables predictive maintenance, anomaly detection, and energy-efficient operations in digital manufacturing environments.

These findings support the DTNF by reinforcing its real-time decision support and optimization layer. The DTNF highlights the need for AI-powered adaptability in smart manufacturing and provides empirical evidence that edge-enabled DTs enhance responsiveness, improve energy efficiency, and optimize predictive maintenance workflows. By reducing the reliance on centralized cloud processing and enabling local, real-time AI-driven decision-making, the study highlights how manufacturers can transition toward more sustainable and circular production systems.

5.3. Case Study 3: Lifecycle Assessment of DT Sustainability

Manufacturing sustainability requires a holistic approach that extends beyond operational efficiency and energy savings. One of the most effective methods for assessing the environmental impact of manufacturing processes is lifecycle assessment (LCA), which evaluates the entire lifecycle of a product, starting from raw material extraction to production, usage, and end-of-life disposal. DT technology has the potential to revolutionize LCA methodologies by providing real-time data integration, predictive analytics, and simulation-based sustainability assessments. By embedding LCA within DT systems, manufacturers can dynamically adjust production processes to minimize environmental impacts, optimize resource utilization, and enhance CE strategies.

Piron et al. (2025) [103] introduced a soft-sensor-based DT framework that integrates LCA to enhance real-time environmental impact assessments in manufacturing. Their study addresses the challenges in LCA data acquisition by leveraging I4.0 and Industry 5.0 technologies, particularly soft sensors, to provide dynamic, real-time sustainability insights. The proposed framework aligns the ISO 23247 DT standards with the ISO 14040 LCA principles [104], offering a structured approach for monitoring energy consumption, emissions tracking, and material flows throughout the production process.

The study applied the LCA-DT integration model (Figure 5) to a case study of PVC extrusion, demonstrating how real-time environmental impact monitoring can inform sustainable manufacturing decisions. A soft sensor model was developed to estimate specific energy consumption of the extruder motor based on the screw speed, material hardness, and viscosity. Their obtained results indicated that the material properties significantly influence the energy demand, highlighting the importance of real-time process adjustments to improve energy efficiency and reduce environmental impact. The LCA-DT framework was structured into four key phases by integrating DT functionalities within the LCA methodology. The first phase, user domain and goal and scope definition, defines the objectives, system boundaries, and environmental impact categories to be assessed, ensuring alignment with sustainability goals. In the second phase, inventory analysis within the DT domain enables the DT to continuously acquire process data from manufacturing operations, supporting the formation of a dynamic lifecycle inventory. The third phase is impact assessment. In this phase, the LCA methodology evaluates the potential environmental impacts based on real-time process data, offering a more precise sustainability assessment than static LCA models. The final phase interprets, analyzes, and summarizes the findings, providing actionable insights for manufacturers to optimize processes and reduce environmental footprints. The key findings demonstrate the effectiveness of integrating DTs with LCA in sustainable manufacturing. The soft-sensor-based framework enables real-time tracking of energy consumption, leading to more efficient resource utilization. By continuously monitoring material flow dynamics, manufacturers can minimize process inefficiencies and improve sustainability. Reference [103] also emphasizes the importance of integrating AI-driven sustainability modeling in which machine learning algorithms help analyze the environmental impact factors and optimize production settings.

The study highlights how digital twin-driven LCA models facilitate the transition from retrospective sustainability assessments to real-time environmental impact monitoring. The findings validate the sustainability validation and optimization layer of the DTNF, demonstrating how AI-powered DTs serve as real-time sustainability intelligence systems. By embedding LCA within DT frameworks, manufacturers can achieve continuous environmental improvements, reinforcing the role of DTs in circular and eco-friendly manufacturing.

5.4. Case Study 4: Human-Centric DT Applications

The integration of DT technology into manufacturing primarily focuses on process optimization, predictive maintenance, and resource efficiency. However, as manufacturing systems become more intelligent and autonomous, a growing need exists to incorporate human-centric design principles into DT applications. A truly sustainable and technologically responsible manufacturing system must improve efficiency and enhance worker safety, ergonomics, and overall well-being. By leveraging AI-driven DTs combined with virtual reality (VR) and augmented reality (AR) systems, manufacturers can create human-centric smart factories where workers and digital systems collaborate seamlessly.

Contini et al. (2025) [105] investigated the role of DT technology in human-centric manufacturing by integrating AR/VR to enhance worker well-being, safety, and efficiency. Their study proposes an SDT (Figure 6) framework that integrates real-time sustainability monitoring with human-centered design principles to improve decision-making across product lifecycles.

Their research highlights how smart factories leverage digital tools, such as the IoT, AI, and big data analytics, to optimize resource allocation and improve sustainability practices. By combining DT with AR/VR, manufacturers can create immersive simulations that provide ergonomic analysis, real-time feedback, and virtual training. SDT continuously collects and processes key sustainability indicators, including energy consumption, emissions, and material use, offering dynamic insights for improving manufacturing processes. A core element of this approach is the development of interactive sustainability dashboards that visualize real-time performance metrics. These dashboards enable workers and designers to track KPIs, optimize operational parameters, and make informed sustainability decisions. SDTs also facilitate lifecycle assessment by simulating different production scenarios and assessing their environmental impacts before implementation.

This case study provides a strong validation of the DTNF by demonstrating the effectiveness of human-centric AI-driven DT applications. The DTNF emphasizes that responsible technological transformations must integrate worker-centric digital solutions to ensure sustainable manufacturing environments. This research aligns with the human–AI collaboration layer of the DTNF, proving that DTs can serve as intelligent, adaptive systems that enhance worker safety, ergonomic efficiency, and training effectiveness.

5.5. Case Study 5: AI-Driven DT Evolution

As manufacturing environments become increasingly complex, the role of DTs has evolved beyond simple process replication and monitoring. The next stage in DT development involves the integration of AI and generative AI, allowing these systems to adapt autonomously to changing production conditions, optimize workflows in real time, and enhance their resilience against supply chain disruptions. AI-driven DT evolution represents a shift from static digital replicas to self-learning and self-optimizing systems, making manufacturing more sustainable, flexible, and efficient.

Mata et al. (2025) [106] presented a comprehensive framework for integrating generative AI into DT design, addressing key challenges in adaptability and scalability within manufacturing systems. Their study highlights how the convergence of DTs and AI enhances real-time decision-making, optimizes resource allocation, and facilitates the rapid reconfiguration of production workflows. By employing a structured design approach that incorporates a morphological matrix, Fuzzy TOPSIS, and operator human knowledge, the framework (Figure 7) systematically improves the effectiveness of DTs. Their study focuses on implementing this framework in a reconfigurable micromachine-manufacturing environment, demonstrating its potential for optimizing production processes, reducing design complexity, and ensuring sustainability. The main contribution of this study is its structured methodology for DT development. The proposed framework integrates S4 features—smart, sustainable, sensing, and socially impactful elements—to enhance DT design. This study introduces a multistep decision-support approach that leverages Fuzzy TOPSIS to evaluate various DT configurations and determine optimal implementations. This AI-enhanced framework aligns DT development with industry requirements while improving flexibility in complex manufacturing environments. Furthermore, this research underscores the importance of human expertise in AI-assisted DT design, highlighting that while AI-driven recommendations improve efficiency, expert validation remains critical for ensuring optimal system performance.

This case study demonstrates how generative AI assists in adaptive DT design by continuously refining the selection of DT components, optimizing the system architecture, and automating performance assessments. Through iterative learning, the AI model improves its decision-making capabilities, enabling rapid reconfiguration of DT structures in response to evolving manufacturing requirements. Reference [97] also explored the integration of large natural language processing models within the DT framework, illustrating their role in automating design evaluations and enhancing collaborative decision-making.

This research provides strong validation for the DTNF, reinforcing its emphasis on AI-driven adaptability and real-time system optimization. By demonstrating the applicability of generative AI in DT evolution, [97] advances the potential of intelligent self-learning DTs capable of addressing dynamic industrial challenges. These findings contribute to the ongoing discourse on AI-enhanced manufacturing by positioning DTs as a pivotal technology in the transition toward adaptive, data-driven, and sustainable industrial ecosystems.

5.6. Case Study 6: AI-Powered Digital Twin Applications for Sustainable I4.0

As industries transition toward I4.0, digital transformation strategies are increasingly integrated with sustainability objectives. Among these strategies, AI-powered DTs are emerging as pivotal technologies for energy efficiency, waste minimization, and resource optimization in manufacturing. The ability of DTs to simulate, analyze, and optimize production processes in real time makes them essential tools for industries seeking to align their operations with circular economic principles. However, implementing AI-driven DTs at scale requires a comprehensive framework considering technological advancements and sustainability.

Mustapha and Abdulquadri (2024) [107] examined the role of digitalization in sustainable manufacturing, emphasizing how I4.0, including AI, DTs, blockchain, and IoT, contributes to sustainability efforts. Their study explores how data-driven decision-making, real-time monitoring, and predictive analytics enable manufacturers to optimize resource utilization, reduce environmental impacts, and transition toward CE models.

A central theme of their research is the ability of DTs to enhance industrial sustainability by providing real-time tracking of resource flows and optimizing energy consumption. This study highlights how companies that integrate AI-driven DTs achieve higher levels of efficiency through predictive maintenance, process optimization, and intelligent supply chain management. Traditional linear production models, in which raw materials are extracted, used, and discarded, are being replaced by closed-loop manufacturing systems in which DTs facilitate material reuse, recycling, and waste reduction strategies. The paper presents several case studies across industries, including automotives, electronics, and aerospace manufacturing, demonstrating the effectiveness of DTs in reducing carbon footprints and improving sustainability performance. One case examined an electronics manufacturer that deployed AI-powered DTs to analyze historical power usage and real-time IoT sensor data. The system identified inefficiencies in energy consumption and optimized machine operation schedules, leading to significant reductions in overall electricity usage. Another case detailed how an automotive supplier leveraged real-time DT monitoring to track material flows, reduce waste generation, and increase the utilization of recycled components.

It also outlines a three-phase model for AI-powered DT deployment in sustainable manufacturing. The first phase, real-time data acquisition, involves collecting operational metrics, such as energy consumption, material utilization, and production efficiency, using IoT-enabled sensors. The second phase, AI-driven analysis, employs machine learning models to detect inefficiencies, predict maintenance requirements, and suggest sustainability-focused optimization. In the final phase, the DT system autonomously adjusts production schedules, machine configurations, and material flow management to ensure optimal resource utilization while minimizing waste. These key findings demonstrate the significant impact of AI-enhanced DTs on sustainability. Companies adopting these technologies report notable reductions in waste generation and energy consumption in addition to increased operational resilience, owing to their predictive maintenance capabilities. By dynamically optimizing manufacturing processes based on real-time environmental data, DTs serve as intelligent sustainability tools that enhance I4.0-driven CE models.

5.7. Case Study 7: Digital Twin for Composting Optimization and ROI

A real-world digital twin implementation was conducted by Vargas et al. (2025) [108] in a composting facility in Cajamarca, Colombia. The DT methodology was structured across three layers—physical, cloud, and virtual—and integrated low-cost sensors with a real-time dashboard The system enabled dynamic monitoring and control of composting processes, leading to a statistically significant increase in process efficiency. Specifically, the composting efficiency improved from 73.57% to 83.65% (+10.08%), and monthly compost output increased by 1200 kg. The intervention also yielded a remarkable economic outcome, with the estimated return on investment reaching 18,957.6%. Field validation confirmed data reliability (p < 0.001), and challenges related to sensor calibration and system connectivity were addressed.

5.8. Case Study 8: Energy DT for Heating Tunnel Optimization

An energy digital twin system was developed and experimentally validated by Billey and Wuest (2024) [109] for a heating tunnel in a laboratory-scale smart manufacturing environment. The DT architecture consisted of a physical subsystem, a virtual model implemented and a bidirectional communication layer using OPC UA. This setup enabled synchronized control and scenario-based optimization of the heating process in real time. During experimental validation, the optimized heating cycle achieved up to 40% energy savings compared to the baseline configuration, while maintaining consistent temperature performance. The implementation also revealed technical challenges, such as precise calibration of the virtual model and maintaining communication integrity for reliable feedback loops. This case demonstrates the interplay of key DT Nexus dimensions: circular product lifecycle intelligence and closed-loop manufacturing systems.

5.9. Case Study 9: Review of DT Applications for Energy Efficiency

Ba et al. (2025) [110] conducted a comprehensive review of 50 studies evaluating digital twin applications in energy efficiency across manufacturing, smart buildings, and industrial processes. The review concluded that DT implementations can yield energy savings of up to 30%, along with reductions in operational costs and enhanced predictive maintenance. However, the authors also highlight consistent challenges: high upfront investment, the complexity of integrating heterogeneous IoT systems, and cybersecurity concerns. This systemic analysis spans SDT-MP Levels 3–5—capturing transitions from modeling to optimization—and underscores the relevance of all three DT Nexus dimensions as DTs enable lifecycle-aware control, closed-loop operations, and improved supply chain traceability.

5.10. Case Study 10: Product Digital Twin for Smart Remanufacturing

Kerin et al. (2022) [111] developed and validated a generic asset model to enable the implementation of product digital twins (DTs) for high-value remanufacturing applications. The study focuses on addressing uncertainties in remanufacturing—such as variable core quality, disassembly challenges, and material matching—by translating them into 16 specific DT functional requirements. The proposed solution utilizes a three-level Unified Modeling Language (UML) class model that integrates real (physical asset), virtual (digital representation), and process (remanufacturing operations) layers. The generic asset model leverages CAD-based design data, in-process quality metrics, sensor outputs, and lifecycle information to support disassembly planning, process routing, and predictive simulations. An example implementation involving large off-road engines shows that while some product and process data exist in current systems (e.g., MES, QMS, ERP), integration and automation are limited. The DT architecture enables simulations to predict remaining useful life (RUL), failure modes, and remanufacturing opportunities based on real-time fleet data. Key DT features such as unique identification, as-built data, quality estimation, and eFuture state prediction support improved remanufacturing planning. The implementation aligns with SDT-MP Level 4 (Predicted) and highlights core DT Nexus dimensions—Circular Product Lifecycle Intelligence and Closed-Loop Manufacturing Systems. Despite current gaps (e.g., lack of unified identifiers, sensor limitations, and fragmented systems), the study provides a foundational DT model adaptable to various remanufacturing contexts.

The digital twin interface integrated CAD models and operational data to provide dynamic visualization, traceability, and fault detection capabilities. Results showed enhanced responsiveness in data linkage and reduced latency in information access across the disassembly, cleaning, reassembly, and testing stages. This implementation highlights the practical value of digital twin-enabled intelligence in managing the complexities of circular remanufacturing. The case aligns with SDT-MP Level 4 (Predicted) and emphasizes the DT Nexus dimensions of circular product lifecycle intelligence and closed-loop manufacturing systems.

Table 7. Comparative overview of digital twin case studies for sustainable manufacturing.

Author	Focus Area	DTNF Dimensions	Key Results
Ba et al. (2025) [110]	Energy Efficiency (Review)	All Three Dimensions	Up to 30% energy savings, enhanced predictive maintenance
Kim et al. (2025) [100]	Supply Chain Optimization	Sustainable Supply Chain Visibility, Circular Product Lifecycle Intelligence	8.97% inventory cost reduction, 1.3% logistics cost reduction
Vargas et al. (2025) [108]	Waste Management Composting	All Three Dimensions	10% increase in composting efficiency, ROI 18,957.6%
Kang et al. (2024) [101]	Edge AI for Additive Manufacturing	Circular Product Lifecycle Intelligence, Closed-Loop Manufacturing Systems	Real-time defect detection, energy-efficient operations
Piron et al. (2025) [103]	Lifecycle Assessment	Circular Product Lifecycle Intelligence, Sustainable Supply Chain Visibility	Real-time energy monitoring, soft-sensor impact assessment
Contini et al. (2025) [105]	Human-Centric DT and VR/AR	Human-Centric Layer (Extended DTNF)	Improved worker ergonomics, AR/VR-based training and safety
Mata et al. (2025) [106]	Generative AI in DT Design	Circular Product Lifecycle Intelligence, Closed-Loop Manufacturing Systems	Optimized DT design, reconfigurable systems, NLP integration
Mustapha & Abdulquadri (2024) [107]	AI-Powered DT in Sustainable I4.0	All Three Dimensions	Reduced energy/waste, predictive maintenance, smart logistics
Billey & Wuest (2024) [109]	Energy Digital Twin for Heating Tunnel	Circular Product Lifecycle Intelligence, Closed-Loop Manufacturing Systems	40% energy savings in heating tunnel, real-time bidirectional optimization
Kerin et al. (2022) [111]	Smart Remanufacturing with Product Digital Twins	Circular Product Lifecycle Intelligence, Closed-Loop Manufacturing Systems	Improved traceability, dynamic asset linkage, enhanced information responsiveness across remanufacturing stages

summarizes real-world case studies that demonstrate the implementation of digital twin technologies across diverse sectors. Each case is assessed based on its sectoral focus, application goals, achieved outcomes (e.g., waste reduction, energy savings, or efficiency improvements), alignment with the Digital Twin Nexus Framework (DTNF), and maturity level from the Sustainable Digital Twin Maturity Path (SDT-MP). These empirical studies collectively validate the DTNF’s applicability in driving sustainable transformation in manufacturing ecosystems.

The above research reinforces the importance of AI-enhanced digital twins as enablers of sustainable smart factories, aligning with the DT Nexus Framework, which emphasizes the role of real-time AI decision-making and predictive analytics in optimizing manufacturing systems. Additionally, it validates AI-powered DTs as critical components of sustainable Industry 4.0. As industries increasingly prioritize carbon neutrality, resource efficiency, and waste reduction, DT technologies provide a scalable pathway for organizations to meet environmental compliance goals while maintaining operational excellence.

A comparative analysis of the case studies presented in this section reveals several key trends in digital twin (DT) adoption for sustainable manufacturing. First, AI-driven real-time analytics significantly enhance the adaptability and intelligence of DT systems, enabling autonomous optimization of complex industrial operations. Second, the integration of DTs with circular economy (CE) principles supports waste minimization, closed-loop material flows, and sustainable resource management. Third, human-centric DT implementations improve not only productivity but also worker well-being, safety, and ergonomic performance. These insights confirm the practical relevance of the Digital Twin Nexus Framework (DTNF) in guiding responsible and intelligent transformation across manufacturing sectors.

The empirical validation of the DTNF and the Sustainable Digital Twin Maturity Path (SDT-MP) across these case studies provides compelling evidence of their utility in supporting sustainable and circular manufacturing transitions. Each case illustrates a specific facet of DT value—ranging from supply chain resilience and energy optimization to lifecycle-based sustainability assessment and AI-assisted system reconfiguration. Collectively, the findings demonstrate that DTs are key enablers of responsible industrial innovation, driving reductions in environmental impact, resource consumption, and operational inefficiencies. By offering a structured roadmap and maturity framework, the DTNF and SDT-MP enable the integration of DT technologies into future-ready, adaptive, and resilient manufacturing ecosystems.

However, the reviewed literature also highlights a persistent lack of standardized sustainability metrics across DT implementations, limiting cross-case comparability and comprehensive impact assessment. To address this gap, future DT–CE integrations should incorporate measurable sustainability indicators—such as carbon footprint reduction, the material circularity index, and energy return on investment—directly into DT architectures and decision feedback loops. Embedding such KPIs will be essential for validating long-term environmental outcomes and advancing evidence-based DT deployment in circular manufacturing.

5.11. Strategies to Overcome Barriers

Despite the promising outcomes demonstrated in these case studies, several systemic challenges remain. Key barriers include data security concerns, interoperability limitations across heterogeneous systems, and insufficient regulatory frameworks for sustainable DT deployment. Moreover, scaling DT implementations across global supply chains requires greater collaboration between manufacturers, digital solution providers, and policy stakeholders. Addressing these issues will be critical to achieving widespread, ethical, and transparent adoption of DT technologies.

To overcome these barriers and support responsible DT adoption, several actionable strategies are proposed:

1.

Policy Incentives and Regulatory Alignment

• Governments and industry bodies should establish policy incentives that promote the integration of DTs in line with CE and sustainability goals. This includes tax incentives for green digitalization, R&D funding for DT infrastructure, and alignment with international sustainability standards (e.g., ISO 50001 [112] for energy management, ISO 14040 for LCA). Harmonizing regulations across borders will also enable more scalable and compliant DT adoption across global supply chains.

2.

Development of Secure-by-Design Architectures

• Addressing data security concerns requires embedding cybersecurity measures into DT system design from the outset. Secure-by-design approaches incorporate encryption, access control, and data provenance tracking to safeguard sensitive operational and environmental data. Standardized frameworks such as NIST SP 800-207 (Zero Trust Architecture) can guide manufacturers in building resilient DT ecosystems.

3.

Interoperability Through Open Standards and Modular Platforms

• To overcome fragmentation among heterogeneous systems, DT solutions should adopt open interoperability standards (e.g., OPC UA, ISO 23247) and modular system architectures. These enable seamless data exchange across diverse platforms, supporting end-to-end lifecycle integration and facilitating third-party technology integration. Collaborative standardization efforts, such as those by ISO, IEC, and ETSI, play a key role in reducing vendor lock-in and improving system compatibility.

4.

Cross-Sector Collaboration and Capacity Building

• Bridging the gap between technological innovation and practical deployment requires coordinated action. Multi-stakeholder partnerships—spanning academia, industry, and government—can help accelerate knowledge transfer, codevelop implementation roadmaps, and ensure alignment with Sustainable Development Goals (SDGs). Workforce reskilling and digital literacy programs are also necessary to prepare human capital for the operationalization of intelligent DT systems.

Together, these strategies provide a foundation for navigating the complex technical, organizational, and ethical challenges associated with DT adoption. By embedding these considerations into both strategic planning and technical design, manufacturers can ensure that digital twin technologies evolve in a direction that supports sustainable, equitable, and resilient industrial ecosystems.

6. Sustainable DT Maturity Path

6.1. The Sustainable DT Maturity Path

As the global manufacturing landscape shifts toward sustainability and CE principles, the role of DTs has become increasingly significant. DTs, which provide virtual representations of physical systems, have been widely adopted to enhance operational efficiency, predictive maintenance, and supply chain optimization. However, their potential to drive sustainable manufacturing and CE integration remains underexplored. Although many companies have implemented DTs for process monitoring and energy optimization, a growing need exists for a structured approach to assess and progressively enhance their sustainability impacts.

The “Framework for Sustainable Industry 4.0 Maturity Path,” introduced by Khan et al. (2025) [113], provides a model for assessing I4.0-driven sustainability maturity. Building on this concept, the current study proposes an SDT-MP framework designed to assess the progressive adoption of AI-driven DTs in sustainable manufacturing. This framework is essential for several reasons. First, it provides a structured roadmap for companies to evaluate their DT maturity levels, thus ensuring a measurable transition from basic monitoring to fully autonomous circular manufacturing systems. Second, it bridges the gap between DT adoption and sustainability objectives by emphasizing how AI-enhanced DTs can optimize resource utilization, minimize waste, and facilitate industrial symbiosis. Third, it serves as a decision-making tool for organizations investing in DT technology, ensuring that their technological advancements align with CE strategies and regulatory sustainability standards. By integrating SDT-MP with the DTNF, industries can strategically enhance their AI-powered DT systems while ensuring a stepwise transition toward sustainable, self-optimizing, and regenerative industrial ecosystems.

Table 8. The proposed SDT-MP.

Maturity Level	Description	Integration with CE	Key DT Capabilities
Monitoring and Compliance	Basic DT usage for real-time monitoring of energy and material flow	- Tracks reduce and recover efforts - Focus on reporting compliance	- IoT sensors collect basic sustainability data - Limited AI-based decision-making
Process Optimization and Waste Reduction	DT supports reactive process adjustments based on sustainability indicators	- Supports reuse, repair, and refurbish - AI-driven waste reduction strategies	- AI-driven predictive maintenance reduces material waste - Energy efficiency optimization models
Circular Process Integration	DT enables proactive lifecycle analysis and sustainable design	- Enables remanufacture and repurpose strategies - Predicts optimal material usage	- AI enhances process reconfiguration for sustainability - Integration with product lifecycle analysis
Autonomous Closed-Loop Manufacturing	DT actively manages material cycles and product recovery	- Strengthens recycle and repurpose initiatives - Enables AI-driven product lifecycle extensions	- Closed-loop material flow optimization - Real-time AI decision-making for circular production
Regenerative and Self-Adaptive Manufacturing	DT achieves fully circular, zero-waste production, self-optimizing for minimal impact	- Implements all 10Rs, achieving full CE integration - Facilitates industrial symbiosis	- AI-driven DT ecosystem, automatically adjusting supply chains and production based on resource availability and demand fluctuations

outlines the Sustainable Digital Twin Maturity Path (SDT-MP), a five-level framework proposed to guide the progressive integration of digital twin technologies in sustainable and circular manufacturing. Each maturity level reflects a distinct stage in digital twin adoption—from basic real-time monitoring (Level 1) to fully regenerative and self-adaptive manufacturing systems (Level 5). The table highlights how DT capabilities evolve from compliance-focused applications to AI-driven, closed-loop ecosystems that align with circular economy (CE) principles. By linking each level to specific CE integration strategies and digital functionalities, the SDT-MP enables industries to assess their current digital twin maturity and identify actionable pathways toward more sustainable, autonomous, and resource-efficient operations.

Level 1: Monitoring and Compliance—Passive Sustainability Tracking

At the foundational stage, DTs function primarily as monitoring tools, collecting real-time data on energy consumption, material flows, and environmental performance. However, they are not actively influenced by sustainability decisions. The primary purpose of this stage is regulatory compliance and basic sustainability reporting to ensure that industries can track KPIs related to emissions, resource utilization, and waste generation. From a CE perspective, DTs at this level contribute to the reduce and recover principles by identifying areas of excessive energy consumption and material waste. However, because AI-driven optimizations have not yet been implemented, industries remain in reactive rather than proactive sustainability modes. The key technological capabilities at this level include basic IoT-enabled tracking systems, real-time dashboards for sustainability data visualization, and automated compliance reporting. The primary limitation is that DTs at this stage operate in a siloed manner and provide passive data collection without advanced analytics or predictive modeling.

Level 2: Process Optimization and Waste Reduction—Reactive Sustainability Enhancement

At this stage, DTs begin to assist in sustainability optimization, although improvements remain reactive and not predictive. AI models are introduced to analyze historical and real-time production data, helping to identify process inefficiencies and opportunities for energy and waste reduction. The role of DTs in the CE expands to include reuse, repair, and refurbish, as AI-driven predictive maintenance ensures that equipment lifespans are extended, unnecessary replacements are minimized, and material waste is reduced. Additionally, some level of resource efficiency tracking has been implemented, enabling industries to optimize production schedules to reduce excess energy consumption and raw material waste. Technologically, industries at this stage integrate AI-powered process-monitoring tools, real-time waste-tracking mechanisms, and early-stage predictive analytics to improve sustainability. However, decision-making still relies on human intervention, and optimization is limited to predefined rule-based algorithms rather than self-learning AI models.

Level 3: Circular Process Integration—AI-Driven Predictive Sustainability

This stage marks a transition from reactive to proactive sustainability management, as AI-powered DTs begin predicting and optimizing material flows, energy efficiency, and waste recovery strategies. At this level, closed-loop material management models are introduced to ensure that production waste is repurposed into secondary manufacturing processes rather than being disposed of. From a CE perspective, DTs at this level actively support remanufacture and repurpose operations. AI-driven lifecycle assessment models predict optimal pathways for material reuse, ensuring that waste is minimized at every stage of production. DTs also enable dynamic resource allocation and real-time adjustment of production variables to improve sustainability outcomes. The primary technological capabilities at this level include real-time AI-enhanced material tracking, predictive sustainability simulations, and closed-loop recycling modeling.

Level 4: Autonomous Closed-Loop Manufacturing—AI-Managed CE Ecosystems

At this level, DTs evolve from predictive tools to autonomous sustainability managers, actively controlling resource flows, material circularity, and energy consumption in real time. AI-powered DTs continuously optimize waste recovery, remanufacturing logistics, and supply chain sustainability, ensuring that the material loops are closed and industrial symbiosis is achieved. The CE application has expanded to include recycle and repurpose initiatives, with DTs autonomously managing the allocation of secondary materials, automating recycling workflows, and optimizing production schedules to minimize raw material dependency. Technological advancements at this stage include fully AI-driven material flow adjustments, real-time waste redistribution across multiple production units, and self-optimizing factory networks that dynamically adjust operations based on sustainability metrics. However, full implementation across industries remains constrained by interoperability challenges and the need for standardized sustainability data-sharing protocols.

Level 5: Regenerative and Self-Adaptive Manufacturing—Fully Autonomous DT CE

At the highest level of maturity, DTs function as self-learning AI ecosystems, continuously improving sustainability performance without human intervention. At this level, factories achieve near-zero waste production by leveraging industrial symbiosis models, in which multiple factories exchange resources and co-optimize CE strategies. At this stage, DTs fully implement all the 10Rs of the CE, creating self-sustaining AI-driven industrial ecosystems that dynamically balance material flows, energy efficiency, and waste recovery. Autonomous AI-powered DTs self-adjust production workflows based on real-time CE simulations to ensure sustainable product lifecycle management and inter-factory sustainability collaboration. The primary challenge at this stage is to scale AI-powered DT sustainability across multiple industries, which requires global standardization, ethical AI considerations, and regulatory frameworks to facilitate digitally driven CEs.

The SDT-MP provides a structured roadmap for industries to integrate AI-driven DTs into CE models. By aligning with the 10Rs of the CE, this model ensures a stepwise transition from basic monitoring to fully regenerative AI-powered manufacturing systems. When combined with the DTNF, the SDT-MP empowers manufacturers to assess their DT capabilities, plan AI-driven sustainability enhancements, and drive next-generation circular smart factories. This dual-framework approach ensures that DT technologies enable operational efficiency and catalysts for long-term industrial sustainability and environmental resilience.

To further guide practical adoption,

Table 9. SDT-MP maturity levels with representative milestones and transition criteria.

Maturity Level	Representative Milestones	Transition Criteria
Monitoring and Compliance	- IoT sensors installed and data collected - Basic dashboards operational - Sustainability data reports generated	Data from critical assets consistently collected and integrated into compliance-reporting workflows
Process Optimization and Waste Reduction	- Predictive maintenance implemented - Waste analytics models operational - Basic AI recommendations executed	Demonstrated ability to implement AI-informed process adjustments that reduce waste and improve energy use
Circular Process Integration	- Lifecycle analysis integrated - Remanufacturing and repurposing simulated - Circular KPIs monitored	Circularity insights used in product/process re-designs; lifecycle data actively informs engineering or design
Autonomous Closed-Loop Manufacturing	- AI-led material flow optimization running - Product lifecycle extensions autonomously triggered	DT independently guides remanufacturing, reuse, or recovery actions within production without manual override
Regenerative and Self-Adaptive Manufacturing	- All 10R strategies supported by DT - AI adapts to supply/demand in real time - Industrial symbiosis enabled	System self-adjusts to changing resource or demand conditions; ecosystem-oriented resource circulation fully embedded

presents representative milestones and transition criteria for each maturity level of the SDT-MP. These criteria help assess organizational readiness and progress, highlighting the growing role of AI, lifecycle intelligence, and autonomous systems in achieving circularity goals. Rather than prescribing fixed thresholds, the table emphasizes demonstrable capabilities—such as real-time monitoring, predictive optimization, and system self-adaptation—that characterize each maturity stage in advancing toward regenerative manufacturing.

Ethical Considerations and Maturity in Digital Twin Adoption

As digital twins (DTs) become central to sustainable industrial transformation, ethical maturity must evolve in parallel with technological advancement. Ethical maturity refers to the systemic integration of ethical principles—such as transparency, accountability, inclusivity, and human-centricity—throughout the DT lifecycle, from design to deployment and governance.

To ensure socially responsible DT adoption, ethical considerations must be embedded in both technical frameworks and organizational practices. This includes the following:

• Transparency and Traceability: Ensuring that AI-driven decision-making processes within DTs are explainable and auditable, particularly when outcomes influence human workers, environmental impacts, or resource allocation.
• Human-in-the-Loop Control: Maintaining human oversight in automated processes, especially in safety-critical or ethically sensitive operations, thereby upholding worker autonomy and preventing algorithmic overreach.
• Data Protection and Privacy: Safeguarding personal and operational data through robust cybersecurity and privacy-preserving mechanisms, particularly in contexts involving real-time monitoring of human behavior or worker performance.
• Stakeholder Alignment and Inclusivity: Engaging diverse stakeholders—including engineers, operators, regulators, and affected communities—throughout the development process to ensure that DT systems reflect shared values and avoid marginalizing vulnerable groups.
• Fair and Accountable AI Governance: Addressing algorithmic bias, ensuring fairness in model training, and establishing governance structures to audit, intervene in, or halt unethical system behaviors.

These principles ensure that human operators, managers, and decision-makers are not only empowered to interpret and act on DT insights, but are also central to the system’s design, oversight, and ethical use.

The Sustainable Digital Twin Maturity Path (SDT-MP) highlights that achieving high levels of DT maturity necessitates parallel development of ethical maturity. DT implementation cannot be considered truly sustainable or scalable unless it also adheres to ethical standards that foster trust, social legitimacy, and long-term stakeholder value.

6.2. Real-World Implementation of DT

As industries transition toward CE principles, integrating AI-powered DTs has emerged as a transformative force in sustainable manufacturing. However, the maturity of DT applications varies significantly and ranges from basic monitoring systems to fully autonomous regenerative ecosystems. This section presents five case studies, each corresponding to a different maturity level within the SDT-MP. These studies illustrate how DTs have evolved from passive sustainability tracking to self-adaptive circular manufacturing, demonstrating a gradual progression toward AI-driven CE integration.

6.2.1. Monitoring and Compliance (Level 1)

At this stage, DTs serve as structured tracking tools, collecting real-time operational data without actively influencing decision-making. Martinez et al. (2021) [114] examined a didactic manufacturing system in which a DT was integrated into an automated pyramid framework to enhance shop floor monitoring. The study demonstrated that the DT enabled the real-time tracking of operational metrics, including equipment status and energy consumption, ensuring compliance with manufacturing process standards. Although the automation pyramid approach facilitated hierarchical monitoring, the system lacked AI-driven decision-making and was limited to reporting and visualization functions. Automated process optimization and predictive analytics were not implemented. Instead, the DT provided real-time insights into system performance while maintaining a structured, multilevel monitoring framework. The study aligns with the stage where DTs focus on data collection and structured compliance tracking to ensure visibility and reporting without direct process optimization.

6.2.2. Process Optimization and Waste Reduction (Level 2)

During this stage, DTs move beyond passive monitoring and begin to actively enhance sustainability through AI-assisted optimization. Although improvements remain reactive rather than fully predictive, the DTs at this stage support real-time process adjustments, reducing waste, inefficiencies, and energy consumption. However, decision-making remains constrained, and full AI-driven autonomy has not yet been achieved.

Ullah & Younas (2024) [115] developed a DT for a flexible manufacturing system that monitored energy efficiency, process scheduling, and machine coordination, and implemented AI-driven optimizations to enhance system performance. Their study demonstrates how the DT replicated the system’s behavior while incorporating real-time AI models for energy reduction and process scheduling. The system actively adjusts scheduling and machine coordination, leading to measurable improvements in energy efficiency and production output. However, the system does not fully integrate predictive sustainability modeling or closed-loop resource allocation, which would enable a shift toward a more autonomous and proactive sustainability framework.

Arsecularatne et al. (2024) [116] conducted a systematic review examining the DT technology in a built environment to optimize energy consumption and reduce carbon emissions. This study highlights the integration of DTs with building information modeling to improve energy efficiency across the lifecycle of a building. However, data integration challenges and a lack of standardization continue to limit the full automation of energy management. Although AI-driven optimizations have been introduced, DTs remain reactive rather than fully autonomous in sustainability decision-making.

Zhang et al. (2024) [117] developed a DT-based multilevel synchronized control system to improve batch production and warehousing coordination in paint manufacturing facilities. By integrating AI-driven optimization models, including genetic algorithms and adaptive large-neighborhood searches, the system refines scheduling, minimizes warehouse congestion, and enhances resource efficiency. This case study demonstrates that DTs enable production logistics synchronization, leading to improved material utilization and reduced system inefficiencies. However, when AI-powered optimization is implemented, fully autonomous decision-making is constrained by synchronization limitations. The study highlights a shift from basic compliance tracking to an AI-assisted, optimization-driven manufacturing environment, where DTs support process improvements but do not yet achieve fully autonomous control.

Ma et al. (2022) [118] proposed a DT- and big data-driven framework for sustainable smart manufacturing in energy-intensive manufacturing industries by integrating real-time production monitoring, lifecycle data analytics, and energy management strategies. Their study examined two real-world applications in Southern and Northern China. Company A implemented an EMS to reduce unit energy consumption and costs, whereas Company B utilized big data-driven lifecycle analysis to enhance sustainability practices. The DT system analyzed energy consumption patterns and supported dynamic adjustments, contributing to waste reduction and operational efficiency improvements. However, although predictive analytics played a role in optimizing production parameters, AI-driven autonomous control was not fully implemented. This research aligns with a stage in which DTs facilitate data-driven sustainability improvements and reactive process optimization, enabling energy-efficient decision-making but not yet achieving full AI-powered circular process integration.

Park et al. (2020) [119] developed a cyber–physical energy system (CPES) to improve the energy efficiency in textile dyeing by leveraging IIoT devices and manufacturing-related big data. The system collects and structures real-time operational data using a product, process, resource, and energy data model. An artificial neural network-based process instruction module recommends optimal time and temperature settings for the dyeing steps, whereas a classification module predicts repeated dyeing risks across the rise–hold–decline stages of the temperature curve. The system achieved a 10.69% increase in energy efficiency and significantly reduced operator dependency. Although CPESs replace trial-and-error with data-driven decision support, they lack real-time closed-loop control and autonomous feedback execution. Functionally, the CPES shares characteristics with early-stage DTs in terms of real-time monitoring and AI-supported optimization, positioning it within a mature stage focused on reactive sustainability improvement and process efficiency.

6.2.3. Circular Process Integration (Level 3)

In this stage, DTs evolve from reactive process optimization to predictive sustainability modeling, enabling dynamic adjustments in resource recovery and material flows.

Mangers et al. (2023) [120] developed a DT-based framework for EoL process chain optimization in PET bottle recycling. Their study utilizes circular value stream mapping and digital state flow modeling to enhance decision-making regarding material reuse and recycling efficiency. However, although state modeling enabled structured process improvements, full AI-driven sustainability control was not implemented, and decision-making remained rule-based rather than autonomous. The study aligns with the stage where DTs transition from optimizing individual processes to facilitating CE integration, ensuring sustainability-driven product design and material reintegration strategies.

6.2.4. Autonomous Closed-Loop Manufacturing (Level 4)

In this stage, DTs transition from predictive tools to self-optimizing and self-regulating systems, thereby enabling autonomous process control and adaptive manufacturing.

Jiang et al. (2024) [121] developed a dual closed-loop self-optimization framework for copper disk casting by integrating real-time simulations and manufacturing feedback to enhance precision and efficiency. By implementing a casting package motion curve optimization model, the DT autonomously adjusted the material flow based on dynamic feedback, ensuring consistent casting quality and reducing manual intervention. The closed-loop feedback mechanism enables adaptive process adjustments, thereby improving the stability and efficiency of the casting process. Although self-optimization was successfully implemented, fully autonomous AI decision-making was constrained by process synchronization limitations. This study aligns with a stage in which DTs transition from predictive tools to self-optimizing systems, demonstrating the increasing role of AI-driven closed-loop control in adaptive manufacturing.

An Energy Digital Twin system was developed and validated by Billey & Wuest (2024) [108] for a heating tunnel in a laboratory-scale smart manufacturing environment. The DT architecture included a physical subsystem, a virtual model implemented and a bidirectional communication layer via OPC UA, enabling real-time data exchange and scenario-based control of the heating process. Experimental results showed energy savings of up to 40% compared to baseline configurations, with consistent temperature control maintained. The system leveraged a linear optimization model to determine optimal turn-off temperatures, minimizing energy use while maintaining stability. While the system employed feedback-informed adjustments, it was not fully autonomous; optimization required simulation-based analysis and operator involvement. The study also noted challenges in model calibration and communication reliability. This implementation highlights the role of DTs in energy efficiency and aligns with DTNF dimensions related to Circular Product Lifecycle Intelligence and Closed-Loop Manufacturing Systems.

6.2.5. Regenerative and Self-Adaptive Manufacturing (Level 5)

At this stage, which is the highest stage of DT maturity, DTs evolve into self-learning, AI-driven ecosystems, dynamically optimizing manufacturing operations with minimal human intervention. These DTs exhibit fully autonomous and adaptive decision-making capabilities, continuous learning from data, and collaboration across multiple production environments to enhance resource efficiency, waste reduction, and sustainability. At this stage, DTs transition from self-regulation to full self-adaptation across industries, forming intelligent, interconnected, and resilient manufacturing ecosystems. Bolender et al. (2024) [122] developed a self-adaptive DT framework that autonomously configures cyber–physical production systems using case-based reasoning (CBR). The system captures human expertise and refines decision-making by learning from past cases, ensuring continuous process optimization and resource efficiency. By integrating CBR-driven adaptation, the DT identifies emerging challenges, retrieves relevant historical cases, and applies self-optimizing process adjustments. This modular DT architecture enables real-time monitoring and automated adaptation, thereby improving production stability and sustainability. However, although the system achieved self-adaptive capabilities, it still relied on predefined case knowledge rather than on fully unsupervised AI learning. The study aligns with the transition toward fully self-learning DTs, reinforcing the role of CBR-driven AI adaptation in achieving intelligent, self-optimizing manufacturing ecosystems. Lehmann et al. (2024) [123] introduced the Internet of DTs (IoDT) framework, which integrates Intelligent DTs (IDTs) with multi-agent system (MAS) principles to enable autonomous decision-making and anticipatory manufacturing behaviors. Within this framework, IDTs function as autonomous agents that continuously analyze environmental data and predefined optimization objectives to adapt and optimize resource allocation, minimize waste, and enhance sustainability. The IoDT architecture facilitates multi-factory collaboration, allowing IDTs to negotiate and coordinate across different facilities to self-optimize and self-regulate manufacturing processes. Although the study demonstrates progress toward AI-driven manufacturing ecosystems, full self-adaptation across industries remains an area for further research. This case study aligns with the transition toward fully autonomous DTs, reinforcing the role of MAS-driven AI in achieving interconnected, zero-waste, and adaptive industrial systems.

These case studies illustrate the gradual evolution of DT technologies in sustainable manufacturing. As industries progress through the SDT-MP levels, AI-powered DTs transform from passive monitoring tools into intelligent, self-learning sustainability agents. By linking these case studies with the DTNF, industries can strategically plan their DT adoption, ensuring a structured transition toward fully regenerative, self-sustaining industrial ecosystems. These findings reinforce the idea that DTs are not merely process optimization tools but are essential enablers of long-term CE innovation and resource resilience.

6.2.6. ESG-Driven DT Innovations: Industry and Policy Perspectives

As the global focus on ESG performance intensifies, more manufacturing leaders are integrating DT technologies for operational efficiency and to drive sustainability and CE impact. Recent industrial and policy developments have revealed that DTs play a growing role in decarbonization, resource traceability, waste reduction, and regenerative system design. These emerging use cases highlight how DTs advance beyond traditional manufacturing optimization and evolve into strategic enablers of corporate ESG objectives. To reflect this evolution, additional real-world case studies are presented below, each mapped to the SDT-MP, demonstrating how leading global organizations deploy DTs to accelerate sustainable innovation and circular transformation.

Siemens and Mercedes-Benz (2024) [124] developed a digital energy twin to optimize energy usage and support sustainable factory planning. This system was first implemented at the Mercedes-Benz ‘Factory 56’ in Sindelfingen, Germany. It simulates building behavior, technical equipment performance, and energy load profiles. This enables data-driven design decisions in the early planning stages, improves energy efficiency, and reduces emissions at greenfield and brownfield sites. This tool is central to Mercedes-Benz’s goal of achieving carbon-neutral global production by 2039. Although the digital energy twin supports predictive sustainability modeling, it operates at the design and planning levels rather than performing closed-loop operational control. This positions it at a mature stage, where DTs facilitate circular process integration and sustainability-centric infrastructure planning.

Siemens and Acciona (2024) [125] collaborated to develop a DT framework for optimizing the design, operation, and maintenance of water treatment plants. The system creates a virtual replica of the physical assets, enabling real-time process optimization, early fault detection, and virtual commissioning of the control software. Additionally, a DT serves as a training environment for plant operators, allowing them to simulate operational scenarios in a risk-free setting. Acciona implemented a system to reduce water loss, enhance energy efficiency, and streamline predictive maintenance in industrial water infrastructure. Although the system significantly advances autonomous operational capabilities and supports resource optimization in real time, it does not feature self-learning or adaptive AI behaviors. This positions the application at a maturity level where DTs enable closed-loop control and AI-assisted sustainability management in complex infrastructure systems.

Siemens (2024) [126] implemented a holistic DT at its logistics center in Nuremberg, enabling real-time transparency and closed-loop optimization of intralogistics operations. The system synchronizes virtual and physical warehouse environments, allowing users to simulate processes, visualize material flows, and make predictive adjustments. By integrating the Siemens Xcelerator and DT services, the platform provides end-to-end data access and planning execution alignment. The key functionalities include early-stage layout simulation, predictive congestion management, and dynamic system configuration. The solution enhances efficiency, throughput, and adaptability while contributing to long-term sustainability through reduced resource use and waste. Although the system exhibits high levels of autonomy and feedback control, it is not fully self-learning or regenerative. As such, it aligns with the maturity stage, in which DTs manage autonomous closed-loop production and logistics systems within a CE.

Siemens and JetZero (2024) [127] partnered to create a comprehensive DT ecosystem for the design, manufacture, and operation of next-generation sustainable aircrafts. Utilizing the Siemens Xcelerator platform, this collaboration supports the development of JetZero’s blended-wing-body aircraft, which aims to reduce fuel consumption and emissions by over 50% compared with conventional designs. The DT spans the entire product lifecycle from early-stage engineering to in-flight performance simulation, enabling real-time optimization, predictive maintenance, and regenerative system feedback. Siemens’ integrated AI capabilities further support adaptive design iterations and operational efficiency, contributing to decarbonization in the aerospace industry. The project exemplifies a maturity stage in which DTs are fully autonomous, AI-driven, and embedded within an industrial ecosystem that dynamically optimizes sustainability across design, production, and end use, defining the characteristics of regenerative and self-adaptive manufacturing.

LG Electronics (2024) [128] implemented a DT at its Changwon manufacturing plant in Korea by upgrading its visual simulation tool to a real-time, data-driven virtual model of the assembly line. This transition resulted in a 17% improvement in productivity, 70% enhancement in product quality, and 30% reduction in energy consumption. The system enables engineers to simulate multiple production configurations and analyze energy use, material flows, and potential inefficiencies before implementing changes to the shop floor. Although the DT integrates sustainability metrics and supports predictive optimization, the system has not yet been operated with autonomous AI control or closed-loop feedback. Thus, this case aligns with a maturity stage in which DTs enable predictive sustainability modeling and lifecycle-aware decision-making, defining the characteristics of circular process integration.

Although leading manufacturing companies, such as Siemens and LG Electronics, demonstrate how DTs are being operationalized to meet ESG goals on the factory floor, an equally important evolution is occurring at the global policy and systems level. International organizations and multilateral coalitions are increasingly exploring how DTs and AI technologies can be scaled to address environmental governance, energy transitions, and CE strategies at the national and planetary levels. The following case studies from the United Nations Environment Programme and the Coalition for Digital Environmental Sustainability (CODES) illustrate how DTs are envisioned not merely as operational tools but also as enablers of systemic, data-driven, and self-adaptive sustainability infrastructures, offering a broader perspective on the future maturity of DT applications.

The UNEP-CCC Climate Technology Progress Report (2024) [129] presents a systems-level overview of how digital technologies, particularly AI and DTs, enable a transition toward a low-carbon and renewable-energy-driven economy. This highlights multiple cases in which AI-powered platforms and predictive control systems are integrated into national energy strategies to improve grid stability, increase efficiency, and align energy use with sustainability goals. The report emphasizes how DT-enabled platforms support renewable integration through dynamic simulations, cross-sectoral coupling (e.g., energy–water–agriculture), and demand-side management. However, although these implementations reflect data-driven foresight and resource optimization, they remain dependent on policy frameworks, human intervention, and fragmented infrastructure. This report underscores the role of DTs in accelerating renewable energy deployment and CE strategies by modeling energy scenarios, enabling predictive interventions, and facilitating co-optimization across energy systems. Nonetheless, the absence of fully autonomous AI ecosystems or industrial symbiosis models limits these cases to the maturity stage of circular process integration, where DTs are advancing predictive sustainability modeling but have not yet achieved autonomous closed-loop operation.

The CODES Action Plan for a Sustainable Planet in the Digital Age (2022) [130] outlines a global framework to harness digital transformation, including DTs, for achieving environmental and social sustainability. One of the core innovations of the plans is the development of a Planetary DT, envisioned as an interoperable digital ecosystem capable of monitoring and modeling complex relationships between environmental, economic, and social systems. This initiative calls for the real-time ingestion of data across diverse platforms, integration of AI for sustainability analysis, and the establishment of safeguards for transparency, privacy, and data ethics. Additionally, the CODES framework includes digital product passports to enhance lifecycle traceability, circularity, and ESG compliance. These efforts emphasize real-time environmental governance, multi-stakeholder participation, and closed-loop feedback systems for adaptive and regenerative development. Although still in the design and policy mobilization phase, this action plan aligns with Level 5 maturity by advocating AI-powered, self-adaptive digital ecosystems that foster CE transitions and planetary-scale sustainability monitoring.

These extended ESG-aligned case studies illustrate how DTs evolve beyond operational efficiency tools into critical enablers of sustainability transformation. Across industrial and policy contexts, DTs are strategically leveraged to address energy efficiency, circular material flows, emissions reduction, and regenerative system management. Although most current applications remain at Levels 3 and 4 of the SDT-MP, focused on predictive modeling and closed-loop optimization, a clear trajectory toward fully autonomous, AI-driven, and ecosystem-integrated DTs is emerging. By linking corporate implementation with global governance strategies, the convergence of AI, sustainability, and DT technology is a pivotal step toward realizing self-adaptive CE ecosystems at industrial and planetary scales.

7. Conclusions

This study explores the integration of digital twins (DTs) into circular economy (CE) practices and addresses existing research gaps by proposing a structured maturity model—the Sustainable Digital Twin Maturity Path (SDT-MP)—and the Digital Twin Nexus Framework (DTNF)—for systematic adoption across industries. Unlike prior research that often addresses DTs in isolated applications (e.g., logistics, smart factories, or product design), this study offers a comprehensive and lifecycle-spanning perspective that captures the broader role of DTs in enabling CE transitions.

Theoretically, this study contributes to the field by conceptualizing DT implementation as a progression across distinct phases—design, development, implementation, operation, and utilization—mapped against sustainability goals and digital maturity levels. It bridges literature gaps by integrating enabling technologies such as AI, IoT, and cloud/edge computing into a unified roadmap for circular digital transformation. This framework supports academic discourse on how digitalization and circularity intersect, offering a foundation for further theory-building in sustainable manufacturing.

Practically, the proposed frameworks offer actionable guidance for industries aiming to embed DTs into CE strategies. The SDT-MP and DTNF enable organizations to assess their current digital twin maturity, plan targeted digital investments, and prioritize the adoption of technologies that enhance lifecycle intelligence, resource efficiency, and closed-loop control. By emphasizing autonomous decision-making and real-time responsiveness, this study supports smart production planning, predictive maintenance, and risk-resilient operations.

Despite these contributions, further empirical validation is essential. Future research should focus on applying the frameworks to diverse industrial contexts to evaluate their technological readiness, economic feasibility, and sectoral adaptability. Additional studies could explore interoperability challenges, regulatory alignment, and long-term impacts on circularity metrics and ESG performance. Longitudinal research and simulation-based modeling can also enrich our understanding of digital twin evolution over time.

This research underscores the pivotal role of digital twins in enabling sustainable, data-driven manufacturing ecosystems. By advancing structured frameworks for DT adoption and circular intelligence, it lays the groundwork for future developments in industrial digitalization. In addition to supporting Sustainable Development Goal 9 (Industry, Innovation, and Infrastructure), the proposed frameworks also contribute to Sustainable Development Goal 12 (Responsible Consumption and Production) by enabling lifecycle traceability, reducing waste, and promoting material reuse through digital twin integration. These insights promote scalable, intelligent, and sustainable production, reinforcing the long-term potential of DTs in shaping resilient and regenerative industries.

Looking ahead, emerging technologies such as edge computing, federated learning, generative AI, and decentralized data infrastructures present promising avenues to strengthen DT deployment in circular manufacturing. Edge computing can enhance responsiveness and reduce data latency in DT-enabled control systems, especially in resource-constrained or security-sensitive environments. Federated learning allows collaborative model training across distributed sites while preserving data privacy—crucial for multi-stakeholder industrial ecosystems. Generative AI offers new possibilities for design automation, synthetic data generation, and dynamic optimization within DT environments, accelerating innovation and reducing development costs. Decentralized architectures further support system resilience and secure data exchange across value chains. However, integrating these technologies also presents challenges, including implementation complexity, interoperability constraints, and the lack of standardized metrics. Future research should address these barriers and evaluate the long-term impact of emerging technologies on circularity outcomes, digital maturity, and industrial sustainability.