Dave Philp, Chief Value Officer at Bentley Systems, discusses how AI, digital twins, clean energy and integrated infrastructure planning are shaping the next generation of data centres across the UAE and wider Middle East.

The UAE and Saudi Arabia are investing heavily in AI and digital infrastructure. From your perspective, what makes the Middle East such an exciting market for the next generation of data centers?

One of the most substantial transformations we are witnessing is the transition of the UAE to a programmatic mindset in terms of data centre development, which makes it a desirable market for the next generation of data centres. We usually consider countries such as UK in terms of building or developing a singular project, but it is impressive that the UAE has been prioritising data centre corridors and digital infrastructure instead of focusing on individual hyperscale data centres.

While speaking to colleagues across the UAE and the Middle East region, I found it interesting as the UAE is constantly investing in the infrastructure needed to achieve its ambitious AI goal. This further includes larger digital ecosystems, sovereign cloud capabilities, and energy infrastructure. I believe it is not just about creating more data centres; it is also about establishing an infrastructure that can help achieve the ambitious economic vision of the UAE. It is important to understand that establishing a robust digital infrastructure is key to power AI, smart city initiatives and industrial diversification. This strategy is part of a much larger national economic plan. By combining digital aspirations with investments in energy and physical infrastructure, the UAE is successfully laying the groundwork for an AI-powered economy. We can also witness that the UK is making efforts to move towards a future economic model, which can motivate other countries to follow suit.

• Power is a growing challenge for AI data centers. How can the Middle East balance rising AI demand with its clean energy goals?

I believe that the UAE has secured a unique position in this regard as any capital investment in AI infrastructure needs to be strategized according to clean energy goals. As the Minister of Energy and Infrastructure highlighted, AI-enabled optimisation can coordinate data-centre demand with grid capacity, renewable generation, storage and shared cooling infrastructure. It can also identify opportunities to recover and reuse waste heat where local conditions make that technically and commercially viable.

This, in my opinion, will define the UAE’s next generation data centres. Integrated planning, which takes into account how data centres interact with renewable energy, solar opportunities, district cooling, battery storage, and demand management, will be prioritised over isolated engineering decisions. These factors need to considered by the UAE, not just in terms of separate workflows.

As I mentioned, these will be common across the region, but we are also witnessing unique models in the UAE. We are seeing a shift from isolated projects towards phased campuses and data-centre corridors supported by shared power, water, cooling and connectivity infrastructure.

• As AI data centers require more cooling, how can operators build facilities that are both efficient and responsible with water use?

Yes, it is interesting. It is similar to not having a favourite child, like you do not want operators to choose between energy and water, but how to optimise both to create a resilient thermal management system. However, it is my belief that it should start the very beginning of the project, investment level, thinking about it as a water ecosystem. It is not just about water on site, but about the sourcing within it.

Additionally, I believe that there is a lot of innovation within data centres as well. Now, when it comes to water-stressed environments, usage of reclaimed water should be considered. On the other hand for high-density AI workloads, direct-to-chip and other liquid-cooling approaches can remove heat closer to the source. Closed-loop systems recirculate coolant rather than continually consuming it, although the overall water and energy performance still depends on how the facility rejects heat to the external environment. Also, we must understand how it will integrate within the local recycling infrastructure within there as well.

As a result, we can now model data centres, which will reduce pressure on potable supplies and improve operational resilience. Effective energy and water management are beneficial for businesses as well. They can monitor performance and optimise water usage, which benefits both the community and operators.

The most significant factor, in my opinion, is that we can move from reactive to predictive water and cooling management with digital twinning and AI. When connected to trustworthy operational data and engineering models, infrastructure digital twins can help operators move from static reporting towards predictive management, testing changes in workload, climate, water availability and equipment performance before those conditions affect operations.

And that can offer significant potential for resilience as well as sustainable data centres with robust governance. This requires a comprehensive and holistic approach to both energy and water, rather than individual components.

Ultimately, the objective is not simply to minimise water consumption in isolation. It is to optimise the complete thermal system, because some lower-water cooling configurations may use more electricity. The right solution depends on climate, workload density, water stress, grid carbon intensity and resilience requirements.

• As the Middle East invests in smart cities like NEOM, what role will data centers play in enabling these developments?

I believe that smart cities should communicate with each other as they are powered by AI. This, showcases proper planning works exceptionally when we consider master planning and data centre planning from urban systems that will be integrated into it.

Additionally, integrated value chains are necessary for communities within the UAE to move forward with smart cities. Now, when we discuss value chains, we mean the energy, water, cooling, mobility, and data services.

A significant portion of our work is conducted on city or municipality level, where we use digital twins to enable planners understand interdependencies and future scenarios that can optimise assets within the framework of smart cities. This is where Bentley Systems’ approach to infrastructure digital twins Infrastructure can provide planners with a shared environment in which to understand dependencies, test future scenarios and make more confident investment and operational decisions.

We frequently consider the data centre to be something we would prefer to keep hidden. But, in reality it is a strategic enabler of better, more resilient urban growth. If we do it correctly and consider, let us call it a digital built UAE, which is plausible, becoming an engine room for smart cities.

In fact, this achievement has to be celebrated. If we do it correctly, it becomes a positive contributor behind them. Therefore, I think that smart cities require smart infrastructure, and data centres are becoming a part of what we now refer to as civic backbone. It must be present, accountable, and advantageous to the communities it serves. Moreover,  it all comes down systems, smart dependencies, and data exchange between various departments. Because city-scale infrastructure spans many owners and systems, the digital-twin environment must be capable of federating trusted information through open standards, interoperable interfaces and appropriate governance.

• Beyond speed to market, what do you think will define the next generation of successful data centers in the Middle East?

Speed to market is still a very significant factor, but it must transcend that. I believe that if I was an investor considering data centres, I would clearly want to increase revenue at an expedited rate, but I would also want to ensure that it is resilient for a substantial time, which obviously includes water and energy.

 It can understand it is sustainable, but I also want certainty. There are certain longer-term concerns, such as whether there will be droughts in the future. My opinion is that it must evolve. Compute hardware and thermal requirements will continue to evolve over the life of the facility, often much faster than the supporting civil, power and utility infrastructure.

Additionally, I also believe it must be flexible and adaptable as cooling technology and workloads evolve. Therefore, we need to consider how we may apply digital twinning once again, not just for capital building, but also for operational excellence.

As AI campuses move into operation, investors will increasingly look beyond capital cost and installed megawatts towards the productivity of the infrastructure: how reliably and efficiently it converts energy and compute capacity into useful AI output. Measures such as cost per token and tokens per watt will sit alongside PUE, WUE, carbon intensity and availability, providing a fuller view of operational and commercial performance.