Information and communications technology (ICT) has become both an enabler of global sustainability and a growing source of environmental strain. As digital networks expand and devices proliferate, the sector’s carbon footprint has risen sharply, challenging the idea that technology inevitably drives efficiency. Recent academic research—from optimization models inspired by swarm intelligence to analyses of device obsolescence and behavioral use patterns—reveals that ICT’s environmental impact depends as much on human and systemic factors as it does on technical innovation. Together, these findings suggest that the future of sustainable digital growth lies in intelligent coordination, longer product lifecycles, and behavioral adaptation rather than in hardware scaling alone.
The ICT industry now contributes between 2 and 4 percent of global CO₂ emissions—comparable to the aviation sector. Despite major gains in energy efficiency, the sheer scale of global data use, cloud computing, and connected devices has offset many of these improvements. A 2025 EurekAlert summary of a multi-decade statistical study found that ICT adoption in the United States correlates positively with rising CO₂ emissions. While newer data centers and processors consume less energy per computation, the total computational load continues to grow faster than efficiency gains. The authors concluded that “the rebound effect”—where efficiency improvements lead to higher overall consumption—has become a defining feature of the digital economy.
The structural forces driving this trend are well known. Cloud infrastructure, 5G connectivity, and edge computing have all expanded exponentially, each requiring power-intensive hardware and cooling systems. Even as chipmakers deploy more efficient transistors and servers adopt dynamic load balancing, the global appetite for data—streaming, AI computation, and real-time analytics—has grown faster than technological mitigation. According to research from the Journal of Cleaner Production, global data center demand has more than tripled since 2015, and despite improvements in energy efficiency, absolute emissions have climbed by over 60 percent.
In response to these challenges, some researchers are seeking algorithmic solutions. A recent paper titled Swarm Intelligence to Reduce ICT’s Carbon Footprint (arXiv, 2025) proposes using distributed swarm-based algorithms to optimize energy use across networks. By drawing inspiration from collective animal behavior—such as flocking birds or schooling fish—the model applies convex optimization to route digital workloads more efficiently, balancing energy use, latency, and carbon cost. The system dynamically schedules computing and data transmission based on real-time grid emissions, effectively allowing networks to “move” workloads toward cleaner energy sources as conditions change.
Such models represent a shift from static efficiency improvements to adaptive coordination, where the system continuously optimizes not only for performance but also for environmental impact. Simulation data from the arXiv study indicated that swarm-based routing could reduce energy consumption in large ICT infrastructures by 15–20 percent, and associated CO₂ emissions by up to 25 percent, without significant degradation in service quality.
This approach aligns with a broader movement toward “green AI” and intelligent network management. Instead of focusing solely on faster hardware, researchers are rethinking how distributed computing systems—data centers, edge nodes, and networks—collaborate. The emphasis on convex optimization also reflects a growing recognition that sustainability requires system-level planning rather than isolated hardware upgrades.
Yet optimization cannot address all facets of the problem. Another study, Environmental and Economic Impact of I/O Device Obsolescence (arXiv, 2025), explores a parallel issue: the ecological toll of rapid device turnover. In ICT systems, software updates, new standards, and hardware cycles often render perfectly functional equipment obsolete. Printers, routers, and peripherals may become incompatible with new protocols long before their physical components fail. The result is a growing tide of electronic waste (e-waste), much of which contains hazardous materials and is costly to recycle.
According to data from the United Nations University, global e-waste exceeded 60 million metric tons in 2024—roughly equivalent to 6,000 Eiffel Towers—and less than 20 percent was formally recycled. The arXiv study calculates that extending the average lifespan of I/O devices by just two years could cut global ICT-related e-waste by more than 10 percent annually. The authors argue that the economic model of “planned obsolescence” contradicts sustainability goals, incentivizing replacement over repair or software backward compatibility.
This issue reflects a deeper paradox in the digital economy: progress often demands replacement. As new software and hardware emerge, old systems lose compatibility and security support, creating a structural incentive for consumers and organizations to upgrade. The economic benefits of innovation—greater functionality, faster performance, new features—must now be balanced against their environmental costs. For ICT policymakers, this means embedding sustainability criteria into design and procurement standards, encouraging modular and interoperable systems, and rewarding companies that prioritize durability over disposability.
The behavioral dimension adds another layer of complexity. A systematic review published on SpringerLink argues that ICT’s net environmental effect depends heavily on how people use technology, not merely on what the technology is. For instance, telework, e-learning, and virtual collaboration can reduce travel emissions and office energy use, but if digital tools lead to longer working hours, additional screen time, or rebound consumption—such as increased streaming or home energy use—the benefits diminish.
The study found that behavioral and systemic factors often outweigh technological efficiency in determining ICT’s environmental footprint. In one cited example, remote work initially reduced commuter emissions by 20 percent in European cities, but subsequent increases in household energy use and leisure travel offset up to half those gains. Similarly, while digital services can reduce paper consumption and physical logistics, the growth of data-heavy entertainment, advertising, and AI applications creates new demand that often cancels the savings.
These findings align with economic models of rebound effects first described by William Stanley Jevons in the 19th century—the idea that increased efficiency lowers costs, which in turn stimulates more consumption. In the ICT sector, this manifests as an expanding digital ecosystem that is both cleaner per unit and larger in aggregate. As the Springer review emphasizes, “technological innovation must be matched by behavioral adaptation and systemic policy coordination” if ICT is to achieve genuine sustainability.
Case studies further illustrate the complexity of ICT’s carbon dynamics. In Japan, telecom operators implementing AI-based traffic management systems cut network energy use by 18 percent over two years, yet total emissions fell only 5 percent because data traffic grew 50 percent during the same period. In contrast, a European initiative under the Horizon 2020 program combined smart network routing with renewable energy sourcing, achieving a 30 percent net reduction in CO₂ emissions. The difference lay in system-level integration—coordinating technology deployment with energy grid management and consumption behavior.
The convergence of these studies signals a new phase in ICT sustainability thinking. The question is no longer simply how to make devices or networks more efficient but how to restructure the entire digital ecosystem—from hardware design and software standards to user practices and policy frameworks. Swarm intelligence may optimize energy routing; modular design may slow e-waste growth; behavioral coordination may unlock real emission reductions. But none of these measures can succeed in isolation.
To reconcile innovation with sustainability, the ICT sector will need a new kind of systems thinking—one that treats emissions not as a byproduct but as a parameter in every optimization problem. The challenge is not technological incapacity but alignment: aligning incentives, algorithms, and behavior toward a collectively sustainable outcome.
Key Takeaways
- ICT’s carbon footprint continues to grow despite efficiency gains due to rebound effects and exponential demand for data and computation.
- Swarm intelligence offers promising tools for optimizing energy consumption dynamically, potentially cutting ICT emissions by up to 25 percent.
- E-waste from device obsolescence remains a critical issue, driven by short hardware lifecycles and limited repairability.
- Behavioral and systemic factors—such as telework habits and digital consumption patterns—often determine whether ICT reduces or increases emissions.
- True sustainability in ICT requires integrating optimization, design, and behavioral change into a unified systems approach.
Sources
- arXiv — Swarm Intelligence to Reduce ICT’s Carbon Footprint — Link
- arXiv — Environmental and Economic Impact of I/O Device Obsolescence — Link
- EurekAlert — ICT Adoption in the U.S. Linked with Rising CO₂ Emissions — Link
- SpringerLink — Time-Use Effects Matter More Than Technology Itself — Link
- Journal of Cleaner Production — Data Centers and Global Energy Trends — Link
- United Nations University — Global E-Waste Monitor 2024 — Link
