Deep dive: Digital twins for infrastructure & industry — what's working, what's not, and what's next

The global digital twin market surged to approximately $17.7 billion in 2024 and is projected to reach $24.5 billion in 2025, representing a staggering 40.1% compound annual growth rate according to Fortune Business Insights. For infrastructure and industrial applications specifically, this technology is reshaping how organisations design, operate, and optimise physical assets—with profound implications for sustainability outcomes. Yet beneath the hype lies a complex landscape of genuine breakthroughs, persistent challenges, and critical decisions that will determine whether digital twins deliver on their decarbonisation promise or become another chapter in measurement theatre.

Why It Matters

Infrastructure accounts for approximately 79% of global greenhouse gas emissions when considering both direct operations and embodied carbon in construction materials. The built environment alone represents nearly 40% of global CO2 emissions, with operational energy use and building materials contributing roughly equal shares. Digital twins offer a pathway to address both dimensions by enabling simulation-based optimisation before physical construction and real-time performance monitoring during operations.

The business case has become compelling. AVEVA's 2024 analysis of power plant implementations demonstrated that digital twins with integrated AI can achieve a 28% reduction in fuel costs, 20% reduction in maintenance expenditures, and 19.5% reduction in operations costs. These efficiency gains translate directly to emissions reductions when applied across energy-intensive infrastructure portfolios.

From a regulatory perspective, the UK's Building Safety Act 2022, the EU's Energy Performance of Buildings Directive, and emerging Scope 3 disclosure requirements under the Corporate Sustainability Reporting Directive (CSRD) are creating compliance drivers that make digital twin adoption increasingly necessary rather than optional. Organisations that invest now in robust digital twin capabilities will find themselves better positioned to meet tightening carbon accounting and verification requirements.

Key Concepts

Understanding digital twins for infrastructure requires distinguishing between several interconnected concepts that are often conflated in marketing materials.

Static vs. Dynamic Twins: A static digital twin is essentially a 3D model enriched with asset data—useful for design and planning but limited in operational value. A dynamic twin integrates real-time sensor data, enabling continuous synchronisation between physical and digital states. True sustainability value emerges primarily from dynamic implementations that can model actual performance against design specifications.

Descriptive, Diagnostic, Predictive, and Prescriptive Capabilities: Digital twins exist on a maturity spectrum. Descriptive twins visualise current conditions. Diagnostic twins identify why deviations occur. Predictive twins forecast future states using machine learning. Prescriptive twins recommend or automate interventions. Most infrastructure implementations today remain at descriptive or diagnostic levels, with predictive and prescriptive capabilities emerging primarily in controlled industrial settings.

Interoperability Standards: The lack of universal data standards remains a fundamental barrier. Key frameworks include Industry Foundation Classes (IFC) for building information modelling, the Open Geospatial Consortium's CityGML for urban-scale applications, and emerging Universal Scene Description (USD) formats pioneered by Pixar and adopted by NVIDIA Omniverse for real-time visualisation. Selecting standards with broad ecosystem support is critical for avoiding vendor lock-in.

Edge vs. Cloud Architecture: Infrastructure twins must balance data residency requirements, latency constraints, and computational needs. Edge computing enables real-time response for safety-critical applications but limits analytical capabilities. Cloud architectures support sophisticated AI processing but introduce connectivity dependencies and data governance challenges.

What's Working and What Isn't

What's Working

Predictive Maintenance in Controlled Environments: Digital twins have demonstrated clear ROI in environments with well-defined operating parameters and comprehensive sensor coverage. Siemens reports that their industrial customers using digital twins for predictive maintenance typically achieve 10-30% reductions in unplanned downtime. The key success factor is sufficient historical failure data to train reliable predictive models—typically requiring 2-3 years of operational data before meaningful predictions emerge.

Design Optimisation and Clash Detection: The use of digital twins during design and construction phases has become standard practice for major infrastructure projects. Bentley Systems reports that projects using their iTwin platform for design coordination typically identify 60-80% of potential clashes before construction begins, substantially reducing costly on-site modifications. This application succeeds because it addresses a well-understood problem with clear metrics.

Energy Performance Simulation: Building energy modelling using platforms like IES Virtual Environment has matured significantly. When calibrated against actual performance data, these tools can predict building energy consumption within 10-15% accuracy—sufficient for meaningful optimisation. The City of Salem, Oregon used AVEVA's digital twin platform to predict toxic algae blooms affecting water quality, demonstrating how simulation capabilities can address sustainability challenges beyond energy.

Urban-Scale Carbon Tracking: Cityzenith's SmartWorldOS platform has been deployed to track real-time carbon emissions across building portfolios, enabling cities to identify highest-impact intervention opportunities. Their Clean Cities initiative, which donates the platform to municipalities working toward carbon neutrality, has demonstrated that urban digital twins can synthesise disparate data sources into actionable carbon intelligence.

What Isn't Working

Scope 3 Data Integration: While digital twins excel at modelling Scope 1 and 2 emissions from direct operations and purchased energy, integrating Scope 3 supply chain emissions remains deeply problematic. Material carbon intensity data is inconsistent, supplier-specific information is often unavailable, and the upstream/downstream emission boundaries are inherently fuzzy. Organisations claiming comprehensive carbon modelling through digital twins are frequently engaging in measurement theatre—producing precise-looking numbers from imprecise inputs.

Retrofit Applications: Creating accurate digital twins of existing infrastructure presents fundamental challenges. Reality capture technologies like Matterport's 3D scanning provide geometric accuracy but cannot reliably identify concealed systems, material properties, or degradation states. The cost of achieving sufficient fidelity for older assets often exceeds the value of the resulting insights.

Cross-Platform Interoperability: Despite industry rhetoric about open standards, most major digital twin platforms remain effectively siloed. Data exported from one platform rarely transfers cleanly to another without significant transformation effort. This limits the ability to create integrated twins spanning multiple asset types from different vendors—a critical barrier for infrastructure portfolios with heterogeneous technology stacks.

Small and Medium Enterprise Adoption: While Fortune 500 companies have broadly adopted digital twin technologies, penetration among smaller organisations remains limited. The combination of upfront implementation costs, ongoing data management requirements, and specialised skill needs creates barriers that current SaaS pricing models have not fully addressed.

Key Players

Established Leaders

Siemens Digital Industries combines automation expertise with digital twin capabilities through its Xcelerator platform. Their strength lies in integrating operational technology (sensors, controllers, industrial systems) with information technology, enabling twins that span from component-level monitoring to facility-wide optimisation. The 70-hectare Siemensstadt Square development in Berlin serves as a showcase for urban-scale digital twin implementation.

Bentley Systems focuses specifically on infrastructure engineering through its iTwin Platform. With deep roots in civil engineering software, Bentley excels at transportation, water, and utility infrastructure applications. Their partnership with Siemens (announced 2016) creates complementary capabilities spanning design through operations.

AVEVA specialises in process industries including power generation, water treatment, and chemical manufacturing. Their CONNECT Industrial Intelligence Platform emphasises asset information management and AI-powered analytics. The Gwinnett County Water Resources implementation in Georgia demonstrated 20% efficiency improvements in water delivery through their digital twin approach.

Autodesk offers building and construction-focused digital twin capabilities through Tandem and its broader Construction Cloud platform. Their October 2025 partnership with Eaton introduced cloud-based energy twins specifically targeting buildings and data centres.

Emerging Startups

Cityzenith (Chicago, Series A, $11M raised) has differentiated through urban-scale digital twins focused specifically on carbon emissions tracking. Their SmartWorldOS platform includes a marketplace of sustainability-focused applications and has been deployed for India's new capital city of Amravati and Orlando's Sports District.

Neara (Australia/US, Series B) creates digital twins specifically for electrical utility infrastructure, enabling power companies to model how their networks respond to extreme weather events. Given increasing climate-driven grid stress, their predictive capabilities address an urgent resilience gap.

Akselos (Switzerland) applies reduced-order modelling techniques to create computationally efficient digital twins for large physical structures including offshore wind foundations and industrial equipment. Their approach enables real-time simulation of assets that would otherwise require prohibitive computational resources.

Key Investors & Funders

Government funding currently exceeds private venture investment in digital twin infrastructure. The US CHIPS Program allocated $285 million for digital manufacturing capabilities. The UK's Digital Twin Centre received £37.6 million in government backing. The EU's Destination Earth initiative is developing continental-scale environmental digital twins.

Venture capital activity has concentrated on Series B rounds, with four companies raising a combined $107 million in 2024-2025. Notable investors include Sequoia Capital (Cityzenith), and strategic investors including Schneider Electric and Siemens participating in enabling technology companies.

Sector-Specific KPI Benchmarks

Sector	KPI	Baseline Range	Digital Twin Target	Measurement Method
Commercial Buildings	Energy Use Intensity (kWh/m²/yr)	150-300	<100	Metered data + calibrated simulation
Water Utilities	Non-Revenue Water (%)	15-40%	<10%	Flow sensor network + hydraulic model
Manufacturing	Overall Equipment Effectiveness	60-75%	>85%	Real-time production monitoring
Transportation	Infrastructure Downtime (hrs/yr)	50-200	<25	Predictive maintenance model accuracy
Power Generation	Heat Rate Deviation (%)	2-5%	<1%	Continuous thermal model calibration

Examples

Gwinnett County Department of Water Resources (Georgia, USA): Serving over one million residents with 70 million gallons of treated water daily, Gwinnett County implemented AVEVA's digital twin platform across their production plants, distribution networks, and collection points. By creating a unified data environment that eliminated silos between engineering, operations, and IT systems, they achieved a documented 20% increase in clean-water delivery efficiency. The implementation used AVEVA's PI Data Infrastructure for real-time streaming combined with cloud-based analytics on Microsoft Azure. Critical to success was executive sponsorship that mandated cross-departmental data sharing and a phased rollout that allowed operational staff to build confidence before expanding capabilities.
Siemensstadt Square (Berlin, Germany): This 70-hectare mixed-use development demonstrates urban-scale digital twin implementation from planning through construction. Siemens Real Estate contracted with Bentley Systems to create a comprehensive twin that enables "building the district twice—first in the digital world and then in the real one." The web-based platform provides public access for community feedback while integrating engineering data, IoT sensor feeds measuring energy consumption and water usage, and enterprise resource planning systems. The explicit sustainability goal is a carbon-neutral district, with the twin enabling simulation and optimisation of all assets before physical construction begins.
Singapore Land Authority (National Digital Twin): Singapore's national digital twin initiative, developed in partnership with Bentley Systems, integrates 3D city models with real-time IoT data streams from thousands of sensors across the city-state. Applications include solar potential analysis for buildings, flood modelling, and pedestrian flow optimisation. The platform has been used to test urban planning scenarios, including analysing how building configurations affect wind patterns and urban heat island effects. Singapore's approach demonstrates how national-scale digital twins can support climate adaptation planning, though the resource intensity of their implementation exceeds what most municipalities can realistically achieve.

Action Checklist

Audit existing data infrastructure: Identify all systems generating relevant operational data, assess data quality and accessibility, and map current integration points before selecting a digital twin platform
Define specific use cases with measurable outcomes: Avoid boil-the-ocean approaches by selecting 2-3 initial applications where success can be objectively measured within 12-18 months
Evaluate interoperability requirements: Document all systems that must exchange data with the twin, assess vendor lock-in risks, and establish standards requirements (IFC, CityGML, USD) before procurement
Establish data governance protocols: Define ownership, access controls, and retention policies for twin data, particularly where sensitive operational information is involved
Build internal capability before scaling: Invest in training existing staff rather than relying entirely on vendor support—long-term success requires organisational understanding of both the technology and its limitations
Plan for model calibration and validation: Allocate ongoing resources for comparing twin predictions against actual performance and iteratively improving model accuracy

FAQ

Q: How long does it take to achieve meaningful ROI from infrastructure digital twins? A: Realistic timelines vary significantly by application. Design optimisation and clash detection deliver value immediately during project delivery. Predictive maintenance typically requires 18-36 months to accumulate sufficient historical data for reliable predictions. Energy optimisation can show results within 6-12 months if baseline metering is already in place. Organisations expecting faster returns should focus on well-defined use cases with clear success metrics rather than enterprise-wide implementations.

Q: What data quality threshold is necessary before implementing a digital twin? A: There is no universal threshold, but as a practical guideline: descriptive twins can function with 70-80% data completeness if gaps are documented; diagnostic twins typically require 85-90% completeness with consistent timestamping; predictive capabilities need 95%+ completeness plus sufficient historical depth (typically 2-3 years of operational data). Organisations often underestimate the data preparation effort required—budget for 30-50% of implementation time on data quality work.

Q: How should organisations approach Scope 3 emissions modelling given current data limitations? A: Transparency about uncertainty is essential. Use industry-average emission factors where supplier-specific data is unavailable, clearly document data sources and assumptions, and present results as ranges rather than false precision. The Science Based Targets initiative's guidance on Scope 3 provides reasonable methodologies. Avoid digital twin platforms that promise precise Scope 3 accounting without acknowledging inherent data limitations—such claims indicate either naivety or intentional obfuscation.

Q: What skills are required to maintain infrastructure digital twins effectively? A: Successful implementation requires a cross-functional team spanning: data engineering (ETL pipelines, data quality monitoring), domain expertise (understanding the physical systems being modelled), analytics capability (statistical analysis, potentially machine learning), and visualisation/UX skills for stakeholder engagement. Most organisations understaff ongoing maintenance relative to initial implementation, leading to model degradation over time.

Q: How do digital twins integrate with Building Information Modelling (BIM) requirements? A: BIM provides the geometric and asset data foundation that digital twins extend with operational information. The UK's BIM Level 2 mandate and ISO 19650 establish data structure expectations that digital twin platforms should consume. Ensure selected platforms support IFC import/export and can maintain synchronisation as design models evolve. The distinction is roughly: BIM captures "as-designed" and "as-built" states while digital twins add "as-operated" continuous monitoring.

Sources

Fortune Business Insights, "Digital Twin Market Size, Share & Growth Report 2025-2032," January 2025. Market size and growth projections for the digital twin sector.
AVEVA, "Reimagining Digital Insight for 2024: How AI and the Digital Twin Will Accelerate Industry 5.0," 2024. Cost reduction statistics from power plant implementations.
Bentley Systems Blog, "A Twin in Berlin: The Siemensstadt Square Digital Twin Project," 2024. Case study details on the Siemens Real Estate implementation.
International Energy Agency, "Buildings Sector Energy Consumption and Emissions Tracking," 2024. Statistics on built environment contribution to global emissions.
IMARC Group, "Digital Twin Market Size, Growth, Analysis & Forecast 2033," 2024. Alternative market sizing methodology and regional analysis.
Gwinnett County Department of Water Resources and AVEVA, "Unlocking Your Industrial Digital Twin," 2024. Water utility efficiency improvement case study.
Cityzenith, "Clean Cities Initiative: Donating Digital Twin Platform to Help Cities Cut Emissions," 2023. Urban carbon tracking implementation approach.
Hexagon, "2025 Digital Twin Statistics," 2025. Executive adoption rates and technology investment patterns.