From "Nice to Have" to Structural KPI
In 2026, sustainability in IT no longer fits within the realm of the "nice to have." It has migrated from a reputational goal to an operational, financial, and architectural criterion. Wavestone frames this clearly by treating Sustainable-by-Design IT as one of the year's central trends, arguing that sustainability must be integrated from the very conception of architecture, infrastructure, and software, rather than merely compensated for afterward.
This shift gained momentum because digital infrastructure has become significantly heavier. The race for AI, cloud computing, and data centers has elevated energy consumption and put pressure on water, materials, and network capacity.
Simultaneously, Deloitte describes 2026 as the year companies exit the pilot phase and begin rebuilding operations around AI, focusing on a few high-value use cases, redesigning workflows, and managing sensitive technological dependencies. This combination makes sustainability a structural issue: it is no longer enough to scale AI; organizations must decide how to scale, where to run it, when to process it, and at what environmental cost.
What Changes When Sustainability Becomes an IT KPI
When sustainability truly enters technology management, it ceases to be just an ESG report and begins to influence decisions regarding architecture, procurement, development, and operations. This includes selecting cloud regions, designing workloads, hardware efficiency, cooling, equipment reuse, energy observability, and prioritizing lightweight software.
A Shift in Mentality
In practice, this means the question is no longer "how much carbon did we emit at the end of the year?" but rather "how was the stack designed to emit less from the start?".
This is an important mental shift because it brings sustainability closer to performance and resilience. Microsoft, for example, has been treating data center design, cooling, and cloud management within a narrative of sustainable engineering, not just compliance.
[ CARBON_AWARE_SCHEDULING_SIMULATOR ]
Observe how spatial and temporal shifting optimizes energy consumption for flexible workloads.
Sustainable-by-Design: Architecture Born with Constraints
If carbon-aware scheduling acts on execution, the sustainable-by-design paradigm acts earlier: during the design phase. Wavestone treats this concept as an explicit trend for 2026, suggesting that systems must be designed to balance performance, cost, and environmental impact.
This involves multiple layers. In infrastructure, it includes region selection, cooling, computational density, renewables, water use, and equipment lifecycle. In software, it involves the choice of languages, execution patterns, caching, compression, time-shifting jobs, reducing unnecessary calls, and better hardware utilization.
Microsoft also highlights in its technical research that reducing cloud emissions involves both operational emissions and embodied emissions in the hardware. This shows that green computing is not just about energy in use, but also about the material cost of the stack. Sustainable architectures are not merely "powered by renewables"; they must also make better use of existing resources.
How the AI Boom Made Sustainability Operational
Sustainability became a central IT theme because AI changed the scale of the problem. Deloitte describes a scenario where the discussion shifts from "doing a little AI everywhere" to focusing on a few governed, high-value use cases while redesigning entire operations.
If a company adds AI to old processes, it amplifies computational costs without redesigning waste. Redesigning workflows end-to-end achieves more value with less redundant processing.
Optimization
Data Centers: Energy, Water, and Cooling as Engineering Agendas
Zero Water Cooling
Microsoft announced a new AI-optimized data center design that consumes "zero water for cooling", utilizing chip-level cooling to avoid evaporating millions of liters annually.
Geothermal Stability
Reuters reported Google's partnership to expand geothermal energy use, seeking stable power sources for digital growth without relying solely on intermittent solar and wind.
Cleaner Data Centers Are Not Enough if Software Wastes Resources
A common mistake is reducing green computing to infrastructure. The software plays a direct role. The Green Software Foundation argues that application behavior can be altered to emit less carbon through time-shifting, better capacity usage, and smarter execution decisions.
The software dictates demand. A poorly designed system with excessive API calls, redundant pipelines, unnecessary permanent jobs, and low data efficiency pressures the data center regardless of its electrical source. The software is part of computing's climate problem because it decides when, where, and how much hardware will be used.
Conclusion: Sustainable IT is Better Engineering
An important point to ensure the text doesn't sound moralistic is showing that green computing does not mean slowing down digital innovation. In fact, the sources suggest the opposite: the pressure for AI, cloud, and resilience is forcing technology to be better designed. Sustainable IT appears less as a sacrifice and more as a rigorous engineering discipline.
Sustainable IT in 2026 is not just "less polluting" IT. It is an IT that is more conscious of what it executes, where it executes, when it executes, and why it executes. This is the point where sustainability ceases to be a promise and becomes true engineering.
Frequently Asked Questions
It is an architectural approach where environmental impact is considered from the inception of the system, balancing performance, cost, and sustainability across infrastructure and software.
The practice of shifting flexible workloads temporally (to times when the grid is cleaner) or geographically (to regions with higher renewable energy availability) to reduce carbon emissions.
The massive compute requirements of AI have made sustainability a structural operational issue, forcing companies to rebuild workflows efficiently rather than just adding AI to wasteful legacy processes.
Software dictates hardware demand. Even a data center powered by 100% renewables is inefficient if poorly written software wastes compute cycles, memory, and network bandwidth.
>> Bibliographic_References.log
- [01] Wavestone. Technology trends 2026: 7 trends shaping the future of IT.
- [02] Deloitte. Tech Trends 2026.
- [03] Google Cloud Blog. Google’s approach to carbon-aware data center.
- [04] Google Blog. We now do more computing where there’s cleaner energy.
- [05] Microsoft. Sustainable-by-design: next-generation datacenters consume zero water for cooling.
- [06] Microsoft Research. Designing Cloud Servers for Lower Carbon.
- [07] Green Software Foundation. Carbon Aware SDK overview.
- [08] Reuters. Google partners with Nevada utility for geothermal to power data centers.