The surge in AI data center power demand tied up with gas turbine shortages is creating a perfect storm. Because, at a time when data centers are looking to expand their power generation capacities to meet growing AI workloads energy consumption, multi-year delivery delays are expected for the equipment needed for expansion.
With conventional power expansion strategies being curtailed by this bottleneck, a new strategy emerges: using turbine cooling as an immediate response strategy. Instead of waiting years for new turbines, utilities can optimize existing assets immediately through TIAC, cooling technologies for gas turbines that act as a power augmentation strategy. The working principle is clear: by increasing turbine output, especially during high ambient temperatures, turbine cooling offers a practical, fast-track response to AI-driven grid stress, all while improving operational efficiency and utility resilience.
As such, turbine cooling strategies allow utilities to articulate their own strategic response against AI workloads energy consumption and turbine supply chain delays.
Building from ARANER’s expertise in expanding power augmentation and energy efficiency in data centers, here’s a look at current data center power demand issues, turbine supply chain delays and the role of turbine cooling as a generation asset optimization strategy.
Why is AI data center power demand growing so fast? A new era of electricity demand
The advancement of AI is inaugurating a new era in electricity demand: one marked by workloads which require massive computing power housed in GPU clusters and hyperscale AI infrastructure. And with this new era comes unprecedented pressure for power grids and data centers, which are seeing previous paradigms and expectations collapse.
Behind advanced computing tasks like machine learning and LLMs lies energy-intensive computational infrastructure. And the problem lies not only on the energy intensity required today, but also on the expectations as these new tasks develop.
A look at figures can provide a clear picture of how this energy intensity is expected to grow:
- While GPT 3 model’s training has been calculated to have consumed 1.29 GWh of electricity (Patterson et al., 2021), estimations point towards GPT4 having required around 50 times that figure.
- The International Energy Agency mentions how “a typical AI-focused data centre consumes as much electricity as 100 000 households, but the largest ones under construction today will consume 20 times as much.”
The analysis also highlights how “data centres account for around one-tenth of global electricity demand growth to 2030 (...). However, the significance of data centres in driving electricity demand differs by country.” For instance, “in the United States, data centres account for nearly half of electricity demand growth between now and 2030” according to their analysis.
AI workloads’ energy consumption is thus on the way to become a structural, long-term demand driver. A trend that intensifies public scrutiny and accountability around data center performance parameters and their energy footprint, while also placing considerable strain on utilities that are not prepared for the demand surge.
Utilities are facing a growing capacity problem
Following the surge in AI workloads energy consumption, utilities and power grids are facing unprecedented pressure to scale up. The formula is simple: either develop and work towards new strategies that effectively meet new demand standards, or collapse under their weight.
More specifically, data centers are being faced with the following questions that require urgent solutions, as described in Xin Chen et al. (2025):
- Peak demand challenges: AI data centers consume electricity in ways that are different from traditional patterns, as intensive model training or large-scale inference operations are characterized by sudden and unpredictable spikes in power demand. Fluctuations like these create uncertainty for grid operators, who see how the match between supply and demand in real time needed to operate becomes more complicated.
- Grid congestion risks: much of the current electrical infrastructure isn’t capable today of supporting the expanding demand required by large-scale data centers.
- Difficulty scaling generation capacity fast enough: the fast-paced AI industry is being met with a generation capacity that is struggling to scale quickly enough. On the one hand, a whitepaper published by Wunderlich-Malec mentions how “expanding transmission and substation capacity often requires 5 to 10 years of planning, driven by permitting, environmental reviews, and construction delays.” On the other hand, issues like gas turbine shortages are straining data centers’ ability to expand their generation capacity.
Why are utilities struggling to respond to AI center power demand growth? The role of gas turbine shortages
Industry-wide gas turbine shortages are added to the perfect storm created by the surge in AI data center power demand, directly affecting the capacity of utilities to respond quickly to data center energy consumption growth.
Turbine supply chain delays now mean there are multi-year wait times for new gas turbines to the point where there are 5-year wait times for large natural gas turbines, and 18-36 months for smaller ones, according to Utility Dive.
On top of this, rising equipment and project costs are further complicating the picture. Citing EPRI research, the source mentioned above explains average gas turbine prices have experienced a nearly 50% increase in the six months prior to March 2026. The source also cites “stress across the global manufacturing supply chain” as “the larger driver” of this increase, with demand growing both for large-scale turbines as well as smaller units.
According to the analysis, gas turbine shortages cannot be understood as a one-time, isolated market disruption. Instead, these are caused by structural issues intrinsic to the supply chain which have no immediate, quick-fix solution. Following the source, the manufacturing process for turbines lies at the heart of this issue, as “meeting today’s demand requires more castings, expanded factory space, and additional assembly capacity”, as well as more skilled workers capable of delivering quality products.
In this context, utilities cannot rely only on newly-built capacity to meet rising AI data center power demand. Those quick to understand these constraints and find alternatives to bypass them are set for a decisive strategic advantage.
How turbine cooling helps increase power output immediately
Turbine Inlet Air Cooling refers to the cooling technologies for gas turbines whose aim is to cool the intake air of a gas turbine in order to achieve power output augmentation.
Through TIAC, performance is improved by addressing the physical relationship between air temperature and turbine efficiency: because higher temperatures decrease air density (and thus mass flow), cooling the intake air allows the turbine to process more mass, resulting in higher power production.
Through this mechanism, TIAC technologies are capable of increasing a gas turbine’s power output by 10% to 30%, and in some cases by more than 30%. This increased output is particularly decisive during high-temperature periods, which have a negative effect on turbine performance.
The benefits of TIAC can be further amplified when combined with Thermal Energy Storage (TES) technologies: systems that can store chilled water during off-peak hours in order to then provide cooling during peak demand times.
The implementation of TIAC and TESTIAC systems stands out as a strategy for facing both gas turbine shortages and increased data center energy consumption, achieving important efficiency improvements without the need to replace equipment. Rather than waiting for turbine supply chain delays to resolve, utilities can enhance their already existing equipment’s efficiency. All in a matter of months, instead of various years.
More specifically, TIAC directly addresses the issues presented by AI workloads energy consumption described above in this article:
- Because it extracts more power from turbines that are already functioning, TIAC’s improved efficiency helps reduce pressure on the grid, without requiring new equipment.
- TIAC in combination with TES technologies provides direct relief for peak demand challenges faced by utilities. On the one hand, TIAC can be employed to maximize power output during demand peaks. On the other hand, TES allows data centers to store cooling capacity during off-peak periods and discharge it when demand increases.
- TIAC and TES provide generation asset optimization by moving towards flexible paradigms, so that power can be augmented when additional capacity is needed.
As such, TES and TIAC provide a path for improving utility resilience and generation capacity with existing assets at a time when the path forward seems harder and harder to find.
In a context where demand peaks are expected to be higher, faster and more unpredictable than ever, utilities that proactively address constraints now position themselves to be leaders in the not-so-distant future.
By embracing technologies like TIAC and TES, they ensure optimization today, bypassing the current constraints generated by infrastructural problems.
If integrated as part of a broader strategy towards energy efficiency (one that, for instance, also includes interventions around data center cooling), TIAC and TESTIAC provide important operational optimizations. At the same time, this improvement in terms of energy efficiency represents a major advancement considering the growing scrutiny surrounding data centers’ environmental footprint.
ARANER’s approach to Turbine Cooling and Power Augmentation
ARANER provides Engineering, Procurement, and Construction expertise for utilities looking to extend their power augmentation and improve their efficiency.
Our experience in high-performance thermal systems allows us to build the flexible and efficient energy infrastructure solutions that utilities need for the AI-era grid challenges ahead. This includes our interventions on Data Center Cooling, as well as our customized Turbine Inlet Air Cooling (TIAC) and Thermal Energy Storage (TES) solutions for power augmentation. As part of our customized solutions, we take into account the integration of these new systems with existing infrastructure.
Our focus is on fast implementation times and boosting operational efficiency, so that data centers can have access today to the solutions they need for managing the AI data center power demand surge.
Want to learn more about how we help optimize data center operations through TIAC, TES and other thermal engineering solutions?
Download our white paper Optimizing thermal performance in modern data centers or get in touch with us to find out how we can help.
LINKEDIN
YOUTUBE