Alois Reitbauer, Chief Strategist at Dynatrace, explains how AI can drive IT sustainability by exposing inefficiencies and enabling optimization.

As organizations pursue digital transformation while facing increasingly stringent sustainability regulations, from the UAE’s Net Zero 2050 Strategy to Saudi Arabia’s Green Initiative, AI is under scrutiny for its high resource usage and associated greenhouse gas emissions. The need for large amounts of computational power to train and subsequently run large models cannot be ignored, but the broader IT sustainability conversation is disproportionately focused on AI, even as other systemic factors in IT pose a greater environmental burden.

Across the GCC, particularly in the UAE and Saudi Arabia, national strategies like Net Zero 2050 and Vision 2030 are accelerating digital transformation while putting green IT practices firmly on the agenda, bringing the environmental impact of energy-intensive technologies like AI into sharper focus. The truth is that the broader IT ecosystem is still riddled with inefficiencies. Idle servers, short hardware lifecycles and a lack of visibility into infrastructure waste are persistent challenges hindering sustainability initiatives. AI may be in the spotlight as the most attention-grabbing technology, but rather than being the main culprit, it should be the catalyst that drives a wider reassessment of IT resource efficiency.

IT infrastructure hindering sustainability

AI has brought longstanding inefficiencies and hidden infrastructure costs into sharper focus, something organizations should treat as a catalyst for action, not a crisis. Governments in the region are already responding to this challenge at scale. In Saudi Arabia, new large-scale data centers are being designed for AI workloads, combining high-density computing with energy-efficient architectures, including one targeting 1.5 GW powered by renewables. In the UAE, the environmental cost of rapid digitalization is being addressed through aggressive action on electronic waste, backed by regulation and public–private partnerships to ensure hardware lifecycle management keeps pace with the nation’s sustainability goals. Globally, the real opportunity lies in optimizing what we are already operating.

One of the most overlooked issues in data center energy efficiency is the prevalence of idle or “zombie” servers, machines that remain powered on while performing little to no useful work. According to a 2024 report by Lawrence Berkeley National Laboratory, even when idle, conventional servers can consume between 27% and 36% of their maximum power. These servers may appear functional but deliver no meaningful computing output, representing a significant and ongoing source of wasted energy – a consideration that’s particularly relevant as new hyperscale data centers scale up across the Gulf. Identifying and decommissioning these underutilized servers is now a key priority in sustainable data center management.

This under-utilization leads to significant resource wastage and directly contributes to greenhouse gas emissions.  Yet, it remains invisible to most organizations due to siloed data and limited operational awareness. Simply put, these under-utilized servers approximately equate to every second server not needing to be produced and installed to begin with. Significant environmental gains as well as cost savings can be achieved by reducing the number of servers in production and utilizing older, cheaper hardware to maintain existing systems. As data centers continue to grow and usage increases, ignoring the problem is simply unsustainable.

Another long-standing concern is the amount of CO₂ emissions caused by hardware production and underutilization, rather than energy consumption. Studies show that embodied emissions from the production and disposal of IT hardware can account for more than half of total lifecycle emissions, making sustained use and reuse a critical factor in IT sustainability efforts, especially in regions reliant on imported hardware, where supply chain resilience is as important as emissions reduction. The hardware lifecycle, from extraction of raw materials to disposal, is rife with inefficiencies such as unoptimized systems, upgrading unnecessarily, and improper disposal methods. Prioritizing sustained use, cloud migration, and proper recycling procedures is necessary for a greener approach.

Sustainability in IT isn’t about perfection but about optimization and adaptability. Systems that can self-adjust to reduce waste and respond dynamically to changing demands are critical to tackling the structural inefficiencies that persist in the technology ecosystem.

Visibility into operations

If organizations are serious about prioritizing sustainability, they must move beyond ad hoc improvements and adopt systems that provide real-time insight and control. This is where observability becomes crucial.

Observability delivers the data and context organizations need to make informed and impactful operational decisions. With a clear view into how infrastructure is performing, teams can identify underutilized resources, reduce energy waste and maximize hardware utilization. All while extending the lifespan of their IT systems.

For instance, observability can determine where energy is being consumed unnecessarily, enabling teams to redistribute workloads dynamically and efficiently. These insights not only support sustainability goals aligned with national digital transformation strategies across the Middle East, but also drive cost savings and improved operational resilience.

By facilitating real-time awareness of performance, usage, and inefficiencies, observability makes sustainability efforts tangible. It transforms environmental goals into measurable outcomes organizations can act on.

An Industry-Wide Shift

Systemic factors hindering sustainability efforts is an industry-wide issue. But once they’re visible, organizations are equipped to take meaningful action. Or they knowingly choose not to. And that choice carries not only environmental costs, but also financial and reputational ones.

The environmental impact of AI is real, but the industry’s first step must be a fundamental mindset shift: one that prioritizes optimization, efficiency and the ability to respond dynamically to evolving systems.

Sustainability is ultimately a systems engineering challenge and observability is the tool that equips organizations to meet it head on. By delivering deep, actionable insights into where waste occurs, observability transforms sustainability from an abstract goal into a concrete operational priority, particularly when the scale of the potential cost-savings becomes clear to the business.

Once inefficiencies are visible, organizations can either act or knowingly ignore the consequences. We can’t fix what we can’t see. But once we see it, we’re accountable. With observability, sustainable IT becomes more than an ambition, it becomes a measurable, operational reality.By Alois Reitbauer, Chief Strategist, Dynatrace

Related post

View all