For much of the past decade, science real estate has revolved around laboratories, talent and venture capital. Those factors remain critical, but power – the availability and quantum of – is increasingly influencing occupier decisions. While funding markets remain volatile, the UK remains one of Europe's leading science ecosystems. Advances in artificial intelligence are beginning to reshape the specification of science buildings at a pace that few anticipated. The most immediate challenge is that tenant power requirements are starting to outstrip the allowances many schemes were originally designed around. Occupiers are seeking significantly greater electrical capacity, resilience and cooling infrastructure than landlords anticipated only a few years ago. In some cases, newly completed buildings risk becoming underpowered before they are occupied.

One reason, amongst others, is the growing adoption of AI-enabled research and computational biology. Workloads previously expected to sit in remote cloud environments are increasingly being brought back into the building through dedicated server rooms, high-performance computing environments and localised data processing. This shift places additional demands on power supply, cooling systems and plant infrastructure, often beyond what a traditional laboratory building was designed to accommodate. The building services strategy is therefore becoming as important as the laboratory design itself.

These trends are also beginning to influence occupational demand patterns. Dry-lab and AI-focused occupiers are taking highly specified wet-lab space, sometimes using only part of the specialist laboratory infrastructure but securing access to the power capacity associated with it. As a result, flexible "TechBio" buildings that can support a combination of laboratory, computational and office uses are increasingly attractive.

At the same time, development economics are becoming more challenging. Although a substantial volume of space is under construction and consented across the Golden Triangle, rising construction costs and higher financing costs are slowing delivery. Many landlords are now reluctant to fund enhanced CAT A+ fit-outs without a committed occupier, which means space takes longer to procure. This can often mean occupiers misjudge how much space is needed and must return to the market for more.

The implication for investors and developers is that future science buildings must be designed around flexibility, power resilience, cooling capacity and structural loading from the outset. As AI becomes embedded across drug discovery, research and automation, the most successful assets will not necessarily be those with the most laboratory space, but those capable of adapting to rapidly evolving technical requirements. As with logistics and data centres, power is increasingly becoming a key location driver in the science sector too.

_{Read the articles within Savills Build: Perspective report below}