Artificial Intelligence
Data
Data Centres
Operations
'AI is the new oil—and data centres are the refineries'
With AI adoption reshaping global industries, Straightline Consulting’s Managing Director, Craig Eadie, shares his insights regarding how data centres are powering the GenAI revolution:
"The age of AI is here. Generative artificial intelligence (GenAI) is rewriting the rulebook when it comes to everything from software development and call centre productivity to copywriting — boosting efficiency and, depending on who you ask, on track to raise the GDP of industrialised nations by 10-15% over the next decade.
"The impact of AI will reshape the global economy over the coming years, consolidating value among the companies that successfully capitalise on this moment — and disrupting those that don’t. The 'arms race' to develop the next generation of AI technologies — like Google’s new Veo 3 video generation tool, released at the start of June, which is already making headlines for its ability to allow anyone willing to pay $249 per month to create hauntingly lifelike, realistic videos of everything from kittens playing to election fraud — is accelerating as well. AI has become the new oil: the global fuel for economic growth. Unlike oil, however, GenAI alone isn’t valuable. Rather, its power lies in the ability to apply GenAI models to data. That process, akin to refining crude into petroleum, happens in the data centre.
"Productivity is far from the only thing GenAI is turbocharging. This rush to build, train, and operate new GenAI models is also accelerating the race to build the digital infrastructure that houses them. Goldman Sachs predicts that global power demand from data centres will increase 50% by 2027 and by as much as 165% by the end of the decade, largely driven by GenAI adoption.
"As someone working in the data centre commissioning sector, it’s impossible to overstate the impact that GenAI is having, and will continue to have, on our industry. GenAI has exploded our predictions. It’s even bigger than anyone anticipated. The money, the scale, the speed — demand is growing even faster than the most optimistic projections pre-2023. By the end of 2025, almost half of all the power data centres consume globally could be used to power AI systems.
"The data centre commissioning space we’re operating in today has transformed dramatically. On the construction and design side, huge changes, not just in how buildings are constructed, but in the technology inside those buildings, are reshaping how we commission them.
"The battle to capitalise on the GenAI boom is a battle to overcome three challenges: access to power, materials, and talent.
"GenAI requires an order of magnitude more power than traditional colocation or cloud workloads. As a result, there are serious concerns about power availability across Europe, especially in the UK. We can’t build the data centres we need to capitalise on the GenAI boom because there’s just not enough power. There are some encouraging signs that governments are taking this challenge seriously. For example, the UK government has responded by creating 'AI Growth Zones' to unlock investment in AI-enabled data centres by improving access to power and providing planning support in some areas of the country. The European Union’s AI Continent Plan also includes plans to build large-scale AI data and computing infrastructures, including at least 13 operational 'AI factories' by 2026 and up to five 'gigafactories' at some point after that.
"However, power constraints and baroque planning and approvals processes threaten to undermine these efforts. Multiple data centre markets are already facing pushback from local councils and communities against new infrastructure because of their effect on power grids and local water supplies. Dublin and Amsterdam already stymied new builds even before the GenAI boom. This comes with risk, because AI engines can be built anywhere. GDPR means data must be housed in-country, but if Europe and the UK don’t move faster, large US AI firms will resort to building their massive centres stateside and deploy the tech across the Atlantic later. Once an AI engine is trained, it can run on less demanding infrastructure. We risk stifling the AI industry in Europe and the UK if we don’t start building faster and making more power available today.
"The other key constraints are access to raw materials and components. Global supply chain challenges have spiked the cost of construction materials, and the lead times for data-centre-specific components like cooling equipment can be as much as six months, further complicating the process of building new infrastructure.
"Access to talent is another pain point that threatens to slow the industry at a time when it should be speeding up. Commissioning is a vital part of the data centre design, construction, and approvals process, and our sector is facing a generational talent crisis. There isn’t enough young talent coming into the sector. That has to change across the board—not just in commissioning, but for project managers, consultants, everyone, everywhere. The pain point is particularly acute in commissioning, however, because of the sector’s relatively niche pipeline and stringent requirements. You can’t just walk in off the street and become a commissioning engineer. The field demands a solid background in either electrical or mechanical engineering or through a trade. Right now, the pipelines to produce the next generation of data centre commissioning professionals just isn’t producing the numbers of new hires the industry needs.
"This obviously affects all data centre commissioning, not just AI. The scale of demand and speed at which the industry is moving means this risks becoming a serious pinch point not too far down the line.
"Looking at the next few years, it’s impossible to say exactly where we’re headed, but it’s clear that, unless Europe and the UK can secure access to reliable, affordable energy, as well as clear the way for data centre approvals to move quickly, pain points like the industry talent shortage and rising materials costs (not to mention lead times) threaten to leave the region behind in the race to capture, refine, and capitalise on the new oil: GenAI."