Operations | Monitoring | ITSM | DevOps | Cloud

UK data centre occupancy hit 91% in 2024, according to Arizton market data, and new capacity is not arriving fast enough to close the gap. Grid connection wait times for new projects now run between five and 15 years, reports Data Center Dynamics, and Savills has attributed the 11% year-on-year drop in new capacity delivery to power constraints rather than a lack of demand or investment. Rising wholesale energy costs are addingpressure to an already tight market.

Ask an operations team where the week went, and you'll usually get a list of things that shouldn't have needed a person. Access requests provisioned by hand. A disk cleared for the ninth time this quarter. Certificates rotated one at a time because the renewal script was scoped, estimated, and never finished. None of it is difficult, and all of it is necessary. And at the end of the quarter there's nothing to point at, because the work left no trace beyond the absence of an outage.

The last few years have seen AI conversations dominated by the need for investment in hyperscale infrastructure as firms race to build ever larger training models. But as those conversations evolve, the emphasis is shifting to the next phase of AI adoption, focusing on the scaling of use cases and real-world value.

When on-premises capacity runs short, public cloud tends to be the first option infrastructure teams reach for. It is quick to provision, removes the hardware procurement problem, and sidesteps the question of what to do with an ageing estate. What it does not settle is whether migrated workloads will perform as the business requires once they are live in production, or whether the recovery design has kept pace with where services now sit.