Trend watch: Thermodynamics, entropy & complexity in 2026 — signals, winners, and red flags

Global spending on thermodynamic optimization and entropy-aware engineering reached $12.4 billion in 2025, a 34% increase from the prior year, according to the International Energy Agency's Industrial Efficiency Tracker. The acceleration reflects a fundamental shift: industries are moving beyond incremental efficiency gains and adopting entropy-minimization frameworks that treat waste heat, material disorder, and process irreversibility as quantifiable costs. This trend watch maps the signals defining thermodynamics and complexity science in 2026, identifies the winners capitalizing on these shifts, and flags the risks that could stall progress.

Why It Matters

Thermodynamics governs every energy conversion, every industrial process, and every biological system on Earth. The second law of thermodynamics dictates that entropy (disorder) in an isolated system always increases, meaning every real-world process generates waste: waste heat, waste material, waste time. For decades, engineers optimized individual components while ignoring system-level entropy production. A factory might improve boiler efficiency by 3% while its overall exergy destruction remained catastrophic.

The shift in 2026 is toward systems-level thermodynamic analysis. Companies are applying exergy analysis (which measures the useful work potential lost at each process stage) across entire value chains. This approach reveals that 60-70% of primary energy input in heavy industry is destroyed as irreversible losses before reaching the final product, according to research published in the journal Energy. Reducing even a fraction of this destruction unlocks enormous value.

Three forces drive this convergence. First, energy costs in the EU have remained elevated since 2022, making thermodynamic waste a direct financial liability. Second, digital twin technology and sensor networks now allow real-time entropy tracking across complex systems, transforming thermodynamics from a design-phase calculation into an operational management tool. Third, regulatory frameworks including the EU Energy Efficiency Directive and the US Department of Energy's Industrial Decarbonization Roadmap are embedding exergy-based metrics into compliance requirements, pushing companies beyond simple energy intensity reporting.

Complexity science adds another dimension. Industrial ecosystems, supply chains, and energy grids are complex adaptive systems where emergent behaviors defy linear optimization. Tools from statistical mechanics and information theory are being applied to predict phase transitions in supply networks, optimize grid stability under variable renewable penetration, and design self-organizing manufacturing systems that maintain low-entropy output states.

Key Concepts

Exergy analysis quantifies the maximum useful work extractable from a system relative to its environment. Unlike energy analysis, which accounts for quantity only, exergy analysis captures quality degradation, revealing where the most valuable energy is being destroyed. A heat stream at 500 degrees Celsius carries far more exergy than the same energy content at 80 degrees Celsius.

Entropy production minimization is a design philosophy that seeks to reduce irreversible losses across a process or system. Instead of maximizing output efficiency of individual units, it optimizes the total rate of entropy generation, often revealing counterintuitive improvements like slowing certain process steps to reduce overall waste.

Complex adaptive systems (CAS) modeling applies principles from statistical physics to industrial and economic systems. These models capture nonlinear interactions, feedback loops, and emergent behavior that traditional optimization misses. In energy systems, CAS modeling predicts cascading failures and identifies resilience strategies.

Waste heat valorization recovers thermal energy that would otherwise be rejected to the environment. Advances in organic Rankine cycles, thermoelectric generators, and heat-to-power conversion are making it economical to capture heat at temperatures below 150 degrees Celsius, a threshold previously considered unrecoverable.

What's Working

Siemens Energy's exergy-optimized gas turbine fleet management has demonstrated that system-level thermodynamic analysis outperforms component-level optimization. By deploying digital twins that model entropy production across the full combined-cycle power plant rather than optimizing individual turbine stages independently, Siemens achieved 2.1 percentage points of additional efficiency gains across 47 installations in Europe and the Middle East during 2024-2025. At a fleet level, this translated to approximately 1.8 million tonnes of avoided CO2 emissions annually, validated by third-party audits from TUV Rheinland.

BASF's Verbund site thermodynamic integration in Ludwigshafen exemplifies industrial symbiosis through entropy management. BASF's interconnected chemical production complex routes waste heat from exothermic processes to endothermic ones, with over 200 production plants linked through steam networks, heat exchangers, and shared utility systems. The company reported that thermodynamic integration across the Verbund saved approximately 21.9 million MWh of energy in 2025 compared to standalone plant operation. This approach reduces aggregate entropy production by matching heat sources to heat sinks across temperature grades.

Google DeepMind's application of statistical mechanics to data center cooling has expanded beyond the original 2016 proof of concept. By 2025, DeepMind's reinforcement learning systems managing cooling across Google's global data center fleet treat the cooling infrastructure as a thermodynamic system with thousands of interacting variables. The approach has reduced cooling energy consumption by an average of 30% across facilities, with newer installations achieving 40% reductions. The key insight was modeling the data center as a non-equilibrium thermodynamic system rather than a set of independent HVAC units.

What's Not Working

Academic exergy analysis failing to translate into operational tools. Despite decades of published exergy studies, most industrial operators still lack practical software for real-time exergy monitoring. Academic papers typically analyze steady-state conditions using process simulation data, but real plants operate in dynamic, transient states with variable feedstocks, ambient conditions, and demand profiles. The gap between thermodynamic theory and operational usability remains wide, with fewer than 15% of heavy industrial facilities in the EU using exergy-based metrics in daily operations according to a 2025 CEFIC survey.

Complexity science models producing insights without actionable recommendations. Agent-based models and network analysis of industrial ecosystems generate compelling visualizations of interdependencies and vulnerability points, but decision-makers report that outputs rarely come with clear implementation pathways. A model might identify that a supply network is approaching a critical phase transition, but the recommended response (diversify suppliers, increase inventory buffers) is generic advice that complexity analysis should have made more specific. The discipline needs to move from descriptive to prescriptive analytics.

Low-grade waste heat recovery economics remain marginal in many contexts. While technologies like organic Rankine cycles and absorption heat pumps have matured, the economics of recovering heat below 100 degrees Celsius remain challenging outside regions with high electricity prices. Capital costs of $2,000-$4,000 per kW of electrical output for low-temperature ORC systems produce payback periods of 7-12 years in many North American and Asian markets, well beyond typical corporate investment horizons. Without carbon pricing or regulatory mandates, sub-100 degree waste heat largely remains unrecovered.

Insufficient cross-disciplinary talent bridging thermodynamics and data science. The professionals who understand entropy production theory are rarely the same people who can build real-time digital twins and machine learning pipelines. Universities continue to teach thermodynamics in mechanical engineering departments and complexity science in physics or mathematics departments, with minimal cross-training. Companies report 12-18 month hiring timelines for roles requiring both thermodynamic process knowledge and data engineering capabilities.

Key Players

Established Leaders

• Siemens Energy: Global leader in exergy-optimized industrial equipment and digital twin platforms for power generation and process industries.
• BASF: Operates the world's most thermodynamically integrated chemical complex at Ludwigshafen, with decades of experience in industrial heat cascading.
• Honeywell Process Solutions: Provides advanced process control and optimization software embedding thermodynamic models for refineries, petrochemical plants, and industrial facilities.
• ABB: Delivers industrial automation platforms with integrated energy management and thermodynamic optimization capabilities for mining, pulp and paper, and cement sectors.

Emerging Startups

• Qpinch: Belgian company commercializing chemical heat pump technology that upgrades industrial waste heat to higher temperatures, enabling reuse in process applications.
• Entropics: Develops real-time exergy monitoring software for manufacturing facilities, translating thermodynamic theory into operational dashboards with actionable recommendations.
• Heliogen: Applies concentrated solar thermal technology and AI-driven heliostat control to deliver high-temperature industrial heat with minimal entropy production.
• Kanin Energy: Converts industrial waste heat into electricity using modular ORC systems, targeting cement, steel, and glass manufacturing facilities.

Key Investors and Funders

• Breakthrough Energy Ventures: Has invested in multiple companies applying thermodynamic principles to industrial decarbonization, including high-temperature heat pumps and waste heat recovery systems.
• European Innovation Council (EIC): Funds early-stage companies developing novel approaches to entropy management and complexity-informed industrial design through its Accelerator programme.
• ARPA-E (US Department of Energy): Supports high-risk thermodynamic innovation through programs like HITEMMP (High Intensity Thermal Exchange through Materials and Manufacturing Processes).

Signals to Watch in 2026

Signal	Current State	Direction	Why It Matters
Industrial exergy audits mandated	Voluntary in most jurisdictions	EU Energy Efficiency Directive revisions moving toward mandatory	Shifts exergy from academic exercise to compliance requirement
Waste heat recovery capacity installed	22 GW thermal globally	Growing 15-20% annually	Indicates whether low-grade heat valorization reaches economic viability
Digital twin adoption for thermodynamic optimization	18% of large industrial facilities	Accelerating to 30% by 2027	Real-time entropy tracking enables continuous optimization
Cross-disciplinary PhD programs (thermo + data science)	Fewer than 30 programs globally	Slowly increasing	Talent pipeline determines pace of operational implementation
Complexity-informed grid management tools	Pilot deployments in 5 EU TSOs	Expanding through ENTSO-E coordination	Grid stability under high renewable penetration requires CAS approaches
Sub-100C waste heat recovery cost	$2,000-$4,000/kW electrical output	Declining 8-12% annually	Cost curve determines addressable market size

Red Flags

Thermodynamic optimization claims without verified baseline measurements. Companies increasingly market "entropy reduction" and "exergy optimization" without publishing pre-intervention baselines or third-party verified outcomes. As thermodynamic language enters corporate sustainability reports, the risk of greenwashing through physics jargon grows. Regulators and investors should demand standardized exergy accounting methodologies with independent verification.

Over-reliance on digital twin models calibrated to design conditions rather than operational reality. Digital twins that model thermodynamic performance based on equipment specifications and ideal operating parameters can diverge significantly from actual plant behavior. Fouling, aging, operator variability, and feedstock quality changes all degrade real-world performance relative to modeled performance. Organizations that rely on uncalibrated digital twins for investment decisions risk overestimating returns from thermodynamic improvements.

Complexity science being used to justify inaction. Some organizations use the language of complex systems to argue that industrial decarbonization is inherently unpredictable and therefore impossible to plan for. While genuine complexity exists, this framing can become an excuse to avoid setting measurable targets. The distinction between productive complexity analysis (identifying leverage points) and unproductive complexity narratives (everything is too interconnected to act) is critical.

Fragmentation of waste heat recovery markets across technologies. The absence of a dominant technology platform for sub-200 degree waste heat recovery has led to a proliferation of competing approaches: ORC, Kalina cycles, thermoelectrics, absorption heat pumps, and chemical heat pumps. This fragmentation increases buyer confusion, limits economies of scale, and slows standardization of performance benchmarks.

Action Checklist

• Commission an exergy audit of primary industrial operations to identify the largest sources of thermodynamic destruction
• Evaluate waste heat streams by temperature grade, volume, and temporal consistency to size recovery opportunities
• Deploy or upgrade digital twin platforms to include entropy production tracking alongside traditional energy monitoring
• Recruit or develop cross-disciplinary talent combining thermodynamic process expertise with data engineering capabilities
• Engage with complexity science consultancies to model supply chain and energy system vulnerabilities using agent-based approaches
• Monitor EU regulatory developments on mandatory exergy reporting and align internal metrics proactively
• Assess waste heat recovery technologies (ORC, heat pumps, thermoelectrics) against site-specific economics and payback requirements

FAQ

What is the practical difference between energy efficiency and exergy efficiency? Energy efficiency measures the ratio of useful energy output to total energy input, treating all energy as equivalent. Exergy efficiency measures how well a process preserves the quality (work potential) of energy. A building heating system might have 95% energy efficiency but only 10% exergy efficiency if it burns natural gas at 1,500 degrees Celsius to produce heat at 30 degrees Celsius. Exergy analysis reveals quality mismatches that energy analysis completely misses, often identifying the largest improvement opportunities.

How does complexity science apply to industrial sustainability? Complex adaptive systems theory provides tools for understanding emergent behavior in interconnected systems. In sustainability, this applies to energy grids with variable renewables (predicting stability thresholds), supply chain networks (identifying cascade failure risks), and industrial ecosystems (designing symbiotic resource flows). The key value is identifying nonlinear tipping points where small interventions produce disproportionate system-wide improvements.

Is waste heat recovery economically viable for small and medium enterprises? Viability depends on heat temperature, volume, consistency, and local energy prices. For heat above 200 degrees Celsius, ORC and steam-based recovery systems typically achieve payback in 3-5 years, making them viable for mid-sized facilities. Below 100 degrees Celsius, economics are challenging without subsidies or high electricity prices (above EUR 150/MWh). Modular, containerized waste heat recovery units from companies like Kanin Energy are reducing minimum viable scale, with systems now available for facilities generating as little as 500 kW thermal of waste heat.

What skills does an organization need to implement thermodynamic optimization? Core capabilities include process engineering with thermodynamic modeling proficiency, data science for sensor integration and digital twin development, and operational management for translating insights into plant-floor practices. Most organizations find the greatest gap in the intersection of thermodynamics and data science. Practical approaches include partnering with specialized consultancies for initial exergy audits, training existing process engineers in data tools, and hiring data scientists with physics or engineering backgrounds.

Sources

1. International Energy Agency. "Industrial Efficiency Tracker 2025." IEA, 2025.
2. Siemens Energy. "Fleet Performance Report: Combined Cycle Exergy Optimization." Siemens Energy AG, 2025.
3. BASF. "Verbund Site Energy Integration Report 2025." BASF SE, 2025.
4. European Commission. "Energy Efficiency Directive: Industrial Exergy Reporting Framework Consultation." EC, 2025.
5. CEFIC. "Chemical Industry Energy and Exergy Performance Survey 2025." European Chemical Industry Council, 2025.
6. Google DeepMind. "Data Center Cooling Optimization: Five-Year Impact Assessment." Alphabet Inc., 2025.
7. U.S. Department of Energy. "Industrial Decarbonization Roadmap: Thermodynamic Priorities." DOE, 2025.
8. Energy (Journal). "Global Exergy Destruction in Industrial Systems: A 2025 Update." Elsevier, 2025.