DataBank CEO on the impact of data centre shortages on AI
In an interview with Technology Magazine, DataBank CEO Raul Martynek highlights how global shortages in data centre capacity are becoming a major constraint on the adoption of artificial intelligence. Martynek explains that while much media attention has focused on the scarcity of GPUs for AI workloads, the larger bottleneck may be the limited availability of data centre facilities with sufficient power, cooling and network infrastructure to host those processors. Before the AI boom, demand for data centres was already rising due to enterprise digitisation and cloud migration, leaving little excess capacity. Operators typically build on a just-in-time basis, which means they are only now scrambling to factor AI’s dramatic growth into future forecasts. The surge in energy demand — AI workloads can require up to five times the power of traditional computing — exacerbates these pressures, given grid constraints and more frequent heat waves. Martynek acknowledges that constructing new data centres is increasingly costly and slow due to inflation, power constraints, supply chain delays and land acquisition challenges. Despite these obstacles, he anticipates sustained high demand for AI-ready infrastructure, rising market prices for capacity, and a shift toward building closer to major metro edges to support low-latency AI applications.
