The Big Technology Takeaways from Data Centre LIVE 2026

Data Centre LIVE: The London Summit offered a window into the challenges that could define how AI infrastructure develops over the next decade.

Hundreds of industry experts gathered at Exhibition White City to discuss energy infrastructure, geopolitical risk, grid access and investment strategies.

The event revealed that the race to build AI-ready facilities extends well beyond purchasing computing hardware.

Power access, government intervention and public perception have become factors that could determine which projects succeed.

Power access delays infrastructure rollout

Vittorio Pierangeli of Rolls-Royce Power Systems opened the summit with an assessment of the sector's constraint.

"Power is the main bottleneck today for the realisation of data centre infrastructure," he said. "The typical construction timeframe for data centres is 18 to 24 months."

The issue is not data centre construction itself. It is the time required to secure access to power infrastructure and grid connectivity.

"We are seeing the lead times to get grid connection in several jurisdictions globally increasing to five to seven years," he added. For operators building AI-ready facilities, this could change deployment strategy, making energy resilience a become a business priority.

The gap between construction timelines and power availability could mean that site selection increasingly depends on existing grid capacity.

AI demand pushes facilities beyond campus scale

The Global Data Centre Strategies panel highlighted how AI demand is pushing facilities into gigawatt-scale territory.

Giampiero Frisio of ABB explains that the industry is moving from campuses measured in tens of megawatts to sites operating at "hundreds of megawatts, or even gigawatts".

Governments are becoming more active in attracting data centre investment. Nomin Chinbat, Minister of Digital Development, Innovation and Communications for Mongolia, described the country as an emerging investment destination.

She argued that its natural cooling capabilities, available land and political stability are advantages for future growth.

The shift toward gigawatt-scale sites could mean that countries with available land and stable power infrastructure gain opportunities they did not have during previous technology cycles.

Government involvement in site selection and infrastructure planning has become more visible. Policy decisions around power generation and land allocation could influence where major AI infrastructure projects locate over the next five years.

Carbon matching targets face practical limits

A panel on the future of the energy transition addressed issues like microgrids, back-up power, power purchase agreements and net-zero targets. One subject discussed was the concept of 24/7 carbon-free power matching.

This approach obliges data centre operators to match every kilowatt hour of electricity they consume with the same amount of renewable energy generation. Martin Reed of CBRE suggested that the approach is "difficult and uncertain".

His firm has managed to achieve a 92% to 93% level of matching at some sites, raising the question whether matching 100% of a data centre's energy was yet viable.

Dame Dawn Childs, Director of Pure Data Centres, argues that now is not the time to let perfectionism get in the way of progress: "I don't think the desire to have 24/7 matched power consumption should stop data centre development. AI is part of the challenge but is also part of the solution."

The 'Jensen Huang effect'

During her fireside conversation, Petrina Steele of Equinix discussed how quantum computing might converge with AI in the data centres of the future.

She acknowledged how important Jensen Huang has been to the quantum sector in the past 18 months.

She recalled how the market reacted when he expressed his scepticism about quantum computing early last year: "He wiped around US$8bn off the market."

But when he publicly changed his mind a few months later, the quantum sector was buoyant again. Petrina calls this the "Jensen Huang effect".

She encouraged companies and individuals to become more curious about the potential of quantum.

"If the King of AI, Jensen Huang, is looking at this," she explained, "then we probably should too."

Controlling the narrative: data centre myth busting 

The AI Data Centre Debate brought together four industry leaders to discuss trends, bust myths and address supply chain struggles, net zero, water consumption, modularity and how to satiate AI's appetite for energy.

Jean-François Berche of GreenScale discussed the misconceptions about how much water data centres use.

"A data centre uses less water than 14 golf courses – we don't talk about this," he says. He argued that the public perception of a data centre's resource consumption is overblown.

Jamie Allen of STACK agreed that a data centre's thirst for water is something of a myth: "It's a technical challenge that has long been overcome."

Jamie referred to how the majority of data centres now use closed-loop cooling systems that recycle water repeatedly.

This could mean that regulatory decisions and public opposition may be based on outdated assumptions about water use rather than current practice.