What does it take to build world-class data centers in harsh, unforgiving climates? We speak to a man who knows the success recipe better than most.
Words by Charudatt Chindarkar
When you think about it, like really think about it, infratructure and people are not all that different. Both operate on a set of principles, both have to be resilient (sometimes in the face of daunting odds), and both have central philosophies that define their future paths. We catch up with Himmath Mohammed, Chief Information Officer, Gulf Data Hub, to understand his playbook for growth, his personal leadership rules, and more.
You have been operating mission-critical infrastructure in 50°C ambient temperatures long before ‘climate risk’ became a buzzword. As India faces similar heatwaves, what is the single biggest architectural lesson (be it thermal walls or airflow design) that Indian operators need to steal from your playbook in the Gulf?
Having designed and operated mission-critical infrastructure in ambient temperatures touching 50°C for over a decade in the Gulf, the biggest architectural lesson I would urge Indian operators to adopt is designing for airflow discipline before chasing mechanical efficiency.
In extreme heat environments, cooling failures rarely start with chillers—they start with poor airflow management. Thermal walls, hot-aisle containment, and high-performance insulation are important, but unless airflow paths are predictable, isolated, and protected from mixing, no cooling system will perform as designed. In the Gulf, we learned early that air is the first fuel of a data center.
Indian data centers, particularly in legacy retrofits, still tolerate leakage—bypasses under racks, poorly sealed cable cutouts, mixed hot and cold zones. These are manageable in moderate climates, but during sustained heatwaves, they become failure multipliers. Our playbook has been to design buildings as thermodynamic systems, not just concrete shells: strict pressure differentials, fully segregated air streams, and envelope designs that treat heat ingress as an adversary to be defeated at the perimeter, not inside the data hall.
If there is one thing to “steal” from the Gulf, it is this: get the airflow right, and everything else becomes easier, cheaper, and more resilient.
We often talk about PUE (Power Usage Effectiveness), but in water-scarce regions like the UAE and Saudi Arabia, WUE (Water Usage Effectiveness) is just as critical. With AI demands skyrocketing, are you betting on Immersion Cooling as the only way to break the link between high compute and high-water consumption?
In water-stressed regions like the UAE and Saudi Arabia, WUE is not a sustainability metric; it is a license to operate. Long before AI workloads entered the mainstream, we were forced to decouple growth from water consumption.
Immersion cooling is a powerful tool, especially for high-density AI workloads, but I would caution against viewing it as the only answer. The real shift is not a single technology. Instead, it is a cooling philosophy.
At Gulf Data Hub, we prioritize water-neutral or near-zero-water designs: advanced air-based cooling, hybrid systems, high-temperature operating envelopes, and intelligent controls that dynamically adapt to workload and ambient conditions. For AI clusters exceeding conventional density thresholds, immersion absolutely has a role—but it must be deployed selectively, with a clear lifecycle and supply-chain strategy for fluids, maintenance, and heat reuse.
The future is not air versus immersion. It is right cooling for the right workload, with water treated as a scarce resource, not a default input. Breaking the link between compute growth and water consumption is achievable, but it requires discipline, not just innovation.
Both the UAE and India are aggressive about Data Sovereignty. You sit at the intersection of this trend. Do you see a future where the ‘Data Embassy’ model becomes standard—where a Gulf Data Hub facility in a different country effectively operates under the diplomatic data laws of its home nation?
Data sovereignty is no longer a regulatory checkbox; it is becoming a strategic pillar of national resilience. Operating across jurisdictions, we see governments increasingly asking not just where data is hosted, but under whose legal and operational control it truly sits.
The “Data Embassy” concept—where a facility operates under the data laws of its home nation even when physically located elsewhere—is intellectually compelling, and in some cases inevitable. For highly sensitive workloads (government, defense, financial systems), I do see a future where bilateral or multilateral frameworks enable sovereign data zones, enforced through legal treaties, operational controls, and cryptographic assurance.
However, this will not become a blanket standard overnight. Trust, verification, and enforcement mechanisms must mature first. What will become standard is operator credibility: governments will partner only with data center platforms that have proven neutrality, governance rigor, and operational excellence across borders.
In that sense, the data center operator becomes as important as the facility itself.
In a business that is defined by ‘zero downtime’ and high stress, what is the one personal leadership ritual that keeps you grounded when the alarms go off?
In a business where alarms are real, consequences are immediate, and downtime is not an option, my grounding ritual is simple: slow down before speeding up.
When an incident occurs, my first instinct is not to react, but to stabilize the room. I insist on clarity: what do we know, what do we not know, and who owns each decision?This pause (often just a few seconds long) prevents panic from masquerading as urgency.
Beyond that, I maintain one non-negotiable habit: post-incident reflection without blame. Every event, whether large or small, is an opportunity to strengthen systems and people. That mindset keeps stress from becoming personal and turns pressure into progress.
In mission-critical infrastructure, calm is not a personality trait; it is a leadership responsibility.
