When Jeff Bezos predicted we’d be building data centers in space in twenty years while speaking at the Italian Tech Week last October, the claim made global headlines.

But for the companies working hard to deliver data centers on Earth, the challenge is far more grounded. Before we think about conquering new terrains, let’s focus on the challenges facing data centers on Earth – and the technologies it will take to overcome them.

Firstly, the scale and speed required for data center construction projects are unprecedented. The typical full-scale data center is approximately 100,000 sq ft (9,290 sqm) and houses roughly 100,000 servers. Yet, hyperscalers often expect these projects to be delivered within 12-14 months – around the same timeframe it takes to build a simple residential home.

The challenge holding back data center construction isn’t a lack of funding – far from it. Morgan Stanley forecasts $3 trillion will be invested into AI data centers worldwide by 2029, and half of this outlay will be on construction alone. The real barrier for development is a perfect storm of increasingly high demand, lengthy planning processes, an undersized workforce, and a siloed industry with varying degrees of technological adoption. The industry can’t solve these problems by simply throwing money at them.

Let’s explore how technology can overcome the common pitfalls facing data center construction and help the sector realise a multi-trillion dollar opportunity.

Overcoming workforce shortages

Despite being a fast growing, high value segment, data center construction isn’t immune to the workforce shortages the rest of the construction industry is facing.

In the US, Deloitte identified an average of 382,000 construction job openings each month in 2024 – and in the UK, more than 140,000 job vacancies threaten to stall data center development and other vital infrastructure projects. This challenge is set to worsen as the workforce ages: by 2030, the average craft worker will be 46 years old. To make matters worse, data center construction requires specialist experience in areas like mechanical, electrical, and plumbing (MEP) systems, as well as specialist skills like precision surveying.

The good news is that automation can plug gaps in overstretched teams and allow less experienced workers to take on highly specialised work. For example, much of the equipment used in a data center build – such as excavators, motor graders, dozers, and lifting equipment – typically requires highly skilled operators, who can be hard to recruit.

Machine control, a technology that uses GNSS (GPS), total stations, lasers, and 3D design models, can augment the operator or even guide heavy equipment automatically by syncing with a digital model to direct the machine’s movements. This technology enables less experienced operators to take on more responsibility, while minimising errors and reducing the need for constant manual surveying and staking.

Automation also reduces the need for specialised skills that are hard to find and decreases the size of the team required on a project. For example, one surveyor can oversee multiple machines simultaneously, rather than a crew of several people checking grades and alignment manually.

Detecting delays before they arise

Clashes are one of the most common pitfalls in data center design, and can cause anywhere from weeks to months of delays. For example, if MEP elements clash with the structural design and that conflict isn’t identified early, it can trigger a cascading set of changes that require significant rework.

Each modification has a knock-on effect, forcing architects and engineers to revisit earlier decisions and often pushing plans back through additional rounds of approval. These reworks quickly add up to wasted resources and avoidable delays.

What’s more, many data center developers rely on prefabrication to accelerate the build, manufacturing components off-site before assembling them on location. If a clash is discovered after prefabrication, these parts may need to be remanufactured to the updated specifications, which is a costly and time-consuming process.

Working within a common data environment helps eliminate these risks by connecting architectural, structural, and MEP data in a single model. When every team works from the same source of truth, design changes are instantly visible, and clashes can be resolved early, well before prefabrication begins. This reduces time and material waste and helps keep projects on schedule.

After construction is completed, the common data environment continues to add value throughout the data center’s lifecycle. By handing over an accurate, as-built digital model to the hyperscaler, contractors enable a smooth transition into operations and maintenance. The model becomes a living record of every system and asset, giving facility teams the context they need to manage the site effectively.

Beyond technology

Technology cannot remove every obstacle. Regulatory delays, infrastructure limitations, grid capacity, and water availability still hinge on geographical and political realities. But digital technologies can ease many of these challenges, for instance, providing more comprehensive evidence for planning approvals, smarter modeling for grid connections, and more efficient water use through advanced cooling systems. In combination, these technologies give developers the ability to manage infrastructural challenges, even if they cannot eliminate them.

Governments and regulators have an equally important role to play in modernising planning processes, investing in power networks, and creating clear frameworks for high efficiency builds.

Ultimately, meeting hyperscalers’ ambitious targets will depend on more advanced, consistent adoption of technology throughout the lifecycle of the data center, from planning to maintenance, supported by forward-looking regulation.