'Lunch with a view': behind the scenes at Datum’s MCR2 project

As well as MCR2’s anticipated capacity and impressive resilience standards (reassuring for our clients/potential clients who depend on robust and efficient data centre solutions), key talking points included Datum’s sustainability initiatives, which underscore all aspects of the MCR2 construction process. Of particular relevance here were the use of PV panels and the incorporated heat reuse capability, which will harness waste heat for conditioning, reuse and delivery to local projects in the Wythenshawe area of Manchester.

Another key topic of discussion was AI in the data centre space. While AI promises remarkable enhancements in efficiency, scalability, and operational accuracy, our conversations showed that its current impact on data centre operations is nuanced. AI systems, particularly those implementing advanced machine learning or deep learning models, are exceptionally resource intensive, and rely heavily on GPU to handle massive amounts of data at high speeds. This reliance makes AI hardware and operations costly to deploy and sustain, with substantial power requirements adding to the total expenditure. We are seeing that the enterprise market has so far seen minimal immediate upheaval in terms of power capacity requirements or hardware usage in relation to AI, and that the up to 30kW per cabinet we provide is more than enough to accommodate the power demands of our clients, even those carrying out early-stage AI experimentation. As innovation in AI hardware continues, its adoption is likely to proliferate and, like all new technologies, it will progressively become more energy-efficient and cost-effective.